2022-01-02 21:38:21 +00:00
|
|
|
# Copyright 2022 Oliver Smith
|
2020-02-20 20:07:28 +00:00
|
|
|
# SPDX-License-Identifier: GPL-3.0-or-later
|
2018-02-19 22:04:01 +00:00
|
|
|
import datetime
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
import logging
|
2018-02-09 18:43:58 +00:00
|
|
|
import os
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
import pmb.build
|
|
|
|
import pmb.build.autodetect
|
|
|
|
import pmb.chroot
|
|
|
|
import pmb.chroot.apk
|
|
|
|
import pmb.chroot.distccd
|
2018-11-15 07:30:49 +00:00
|
|
|
import pmb.helpers.pmaports
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
import pmb.helpers.repo
|
|
|
|
import pmb.parse
|
|
|
|
import pmb.parse.arch
|
|
|
|
|
|
|
|
|
2021-11-09 11:54:07 +00:00
|
|
|
def skip_already_built(pkgname, arch):
|
2017-12-05 23:17:39 +00:00
|
|
|
"""
|
|
|
|
Check if the package was already built in this session, and add it
|
|
|
|
to the cache in case it was not built yet.
|
|
|
|
|
|
|
|
:returns: True when it can be skipped or False
|
|
|
|
"""
|
2021-10-30 12:20:57 +00:00
|
|
|
if arch not in pmb.helpers.other.cache["built"]:
|
|
|
|
pmb.helpers.other.cache["built"][arch] = []
|
|
|
|
if pkgname in pmb.helpers.other.cache["built"][arch]:
|
2017-12-05 23:17:39 +00:00
|
|
|
logging.verbose(pkgname + ": already checked this session,"
|
|
|
|
" no need to build it or its dependencies")
|
|
|
|
return True
|
2021-10-30 12:20:57 +00:00
|
|
|
pmb.helpers.other.cache["built"][arch].append(pkgname)
|
2017-12-05 23:17:39 +00:00
|
|
|
return False
|
|
|
|
|
|
|
|
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
def get_apkbuild(args, pkgname, arch):
|
|
|
|
"""
|
new action: 'pmbootstrap repo_missing'
Add a new action that lists all aports, for which no binary packages
exist. Only list packages that can be built for the relevant arch
(specified with --arch). This works recursively: when a package can be
built for a certain arch, but one of its dependencies
(or their depends) can not be built for that arch, then don't list it.
This action will be used for the new sr.ht based build infrastructure,
to figure out which packages need to be built ahead of time (so we can
trigger each of them as single build job). Determining the order of the
packages to be built is not determined with pmbootstrap, the serverside
code of build.postmarketos.org takes care of that.
For testing purposes, a single package can also be specified and the
action will list if it can be built for that arch with its
dependencies, and what needs to be built exactly.
Add pmb/helpers/package.py to hold functions that work on both pmaports
and (binary package) repos - in contrary to the existing
pmb/helpers/pmaports.py (see previous commit) and pmb/helpers/repo.py,
which only work with one of those.
Refactoring:
* pmb/helpers/pmaports.py: add a get_list() function, which lists all
aports and use it instead of writing the same glob loop over and over
* add pmb.helpers.pmaports.get(), which finds an APKBUILD and parses it
in one step.
* rename pmb.build._package.check_arch to ...check_arch_abort to
distinguish it from the other check_arch function
2018-11-15 07:36:39 +00:00
|
|
|
Parse the APKBUILD path for pkgname. When there is none, try to find it in
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
the binary package APKINDEX files or raise an exception.
|
|
|
|
|
|
|
|
:param pkgname: package name to be built, as specified in the APKBUILD
|
new action: 'pmbootstrap repo_missing'
Add a new action that lists all aports, for which no binary packages
exist. Only list packages that can be built for the relevant arch
(specified with --arch). This works recursively: when a package can be
built for a certain arch, but one of its dependencies
(or their depends) can not be built for that arch, then don't list it.
This action will be used for the new sr.ht based build infrastructure,
to figure out which packages need to be built ahead of time (so we can
trigger each of them as single build job). Determining the order of the
packages to be built is not determined with pmbootstrap, the serverside
code of build.postmarketos.org takes care of that.
For testing purposes, a single package can also be specified and the
action will list if it can be built for that arch with its
dependencies, and what needs to be built exactly.
Add pmb/helpers/package.py to hold functions that work on both pmaports
and (binary package) repos - in contrary to the existing
pmb/helpers/pmaports.py (see previous commit) and pmb/helpers/repo.py,
which only work with one of those.
Refactoring:
* pmb/helpers/pmaports.py: add a get_list() function, which lists all
aports and use it instead of writing the same glob loop over and over
* add pmb.helpers.pmaports.get(), which finds an APKBUILD and parses it
in one step.
* rename pmb.build._package.check_arch to ...check_arch_abort to
distinguish it from the other check_arch function
2018-11-15 07:36:39 +00:00
|
|
|
:returns: None or parsed APKBUILD
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
|
|
|
# Get existing binary package indexes
|
2018-03-04 13:44:27 +00:00
|
|
|
pmb.helpers.repo.update(args, arch)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
new action: 'pmbootstrap repo_missing'
Add a new action that lists all aports, for which no binary packages
exist. Only list packages that can be built for the relevant arch
(specified with --arch). This works recursively: when a package can be
built for a certain arch, but one of its dependencies
(or their depends) can not be built for that arch, then don't list it.
This action will be used for the new sr.ht based build infrastructure,
to figure out which packages need to be built ahead of time (so we can
trigger each of them as single build job). Determining the order of the
packages to be built is not determined with pmbootstrap, the serverside
code of build.postmarketos.org takes care of that.
For testing purposes, a single package can also be specified and the
action will list if it can be built for that arch with its
dependencies, and what needs to be built exactly.
Add pmb/helpers/package.py to hold functions that work on both pmaports
and (binary package) repos - in contrary to the existing
pmb/helpers/pmaports.py (see previous commit) and pmb/helpers/repo.py,
which only work with one of those.
Refactoring:
* pmb/helpers/pmaports.py: add a get_list() function, which lists all
aports and use it instead of writing the same glob loop over and over
* add pmb.helpers.pmaports.get(), which finds an APKBUILD and parses it
in one step.
* rename pmb.build._package.check_arch to ...check_arch_abort to
distinguish it from the other check_arch function
2018-11-15 07:36:39 +00:00
|
|
|
# Get pmaport, skip upstream only packages
|
|
|
|
pmaport = pmb.helpers.pmaports.get(args, pkgname, False)
|
|
|
|
if pmaport:
|
|
|
|
return pmaport
|
2018-02-20 19:52:28 +00:00
|
|
|
if pmb.parse.apkindex.providers(args, pkgname, arch, False):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
return None
|
|
|
|
raise RuntimeError("Package '" + pkgname + "': Could not find aport, and"
|
|
|
|
" could not find this package in any APKINDEX!")
|
|
|
|
|
|
|
|
|
2019-03-10 00:28:10 +00:00
|
|
|
def check_build_for_arch(args, pkgname, arch):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
2019-03-10 00:28:10 +00:00
|
|
|
Check if pmaport can be built or exists as binary for a specific arch.
|
|
|
|
:returns: * True when it can be built
|
|
|
|
* False when it can't be built, but exists in a binary repo
|
|
|
|
(e.g. temp/mesa can't be built for x86_64, but Alpine has it)
|
|
|
|
:raises: RuntimeError if the package can't be built for the given arch and
|
|
|
|
does not exist as binary package.
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
2019-03-10 00:28:10 +00:00
|
|
|
# Check for pmaport with arch
|
new action: 'pmbootstrap repo_missing'
Add a new action that lists all aports, for which no binary packages
exist. Only list packages that can be built for the relevant arch
(specified with --arch). This works recursively: when a package can be
built for a certain arch, but one of its dependencies
(or their depends) can not be built for that arch, then don't list it.
This action will be used for the new sr.ht based build infrastructure,
to figure out which packages need to be built ahead of time (so we can
trigger each of them as single build job). Determining the order of the
packages to be built is not determined with pmbootstrap, the serverside
code of build.postmarketos.org takes care of that.
For testing purposes, a single package can also be specified and the
action will list if it can be built for that arch with its
dependencies, and what needs to be built exactly.
Add pmb/helpers/package.py to hold functions that work on both pmaports
and (binary package) repos - in contrary to the existing
pmb/helpers/pmaports.py (see previous commit) and pmb/helpers/repo.py,
which only work with one of those.
Refactoring:
* pmb/helpers/pmaports.py: add a get_list() function, which lists all
aports and use it instead of writing the same glob loop over and over
* add pmb.helpers.pmaports.get(), which finds an APKBUILD and parses it
in one step.
* rename pmb.build._package.check_arch to ...check_arch_abort to
distinguish it from the other check_arch function
2018-11-15 07:36:39 +00:00
|
|
|
if pmb.helpers.package.check_arch(args, pkgname, arch, False):
|
2019-03-10 00:28:10 +00:00
|
|
|
return True
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
2019-03-10 00:28:10 +00:00
|
|
|
# Check for binary package
|
|
|
|
binary = pmb.parse.apkindex.package(args, pkgname, arch, False)
|
|
|
|
if binary:
|
|
|
|
pmaport = pmb.helpers.pmaports.get(args, pkgname)
|
|
|
|
pmaport_version = pmaport["pkgver"] + "-r" + pmaport["pkgrel"]
|
|
|
|
logging.debug(pkgname + ": found pmaport (" + pmaport_version + ") and"
|
|
|
|
" binary package (" + binary["version"] + ", from"
|
|
|
|
" postmarketOS or Alpine), but pmaport can't be built"
|
|
|
|
" for " + arch + " -> using binary package")
|
|
|
|
return False
|
|
|
|
|
|
|
|
# No binary package exists and can't build it
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
logging.info("NOTE: You can edit the 'arch=' line inside the APKBUILD")
|
|
|
|
if args.action == "build":
|
|
|
|
logging.info("NOTE: Alternatively, use --arch to build for another"
|
2018-02-01 21:02:04 +00:00
|
|
|
" architecture ('pmbootstrap build --arch=armhf " +
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
pkgname + "')")
|
|
|
|
raise RuntimeError("Can't build '" + pkgname + "' for architecture " +
|
|
|
|
arch)
|
|
|
|
|
|
|
|
|
|
|
|
def get_depends(args, apkbuild):
|
|
|
|
"""
|
|
|
|
Alpine's abuild always builds/installs the "depends" and "makedepends"
|
|
|
|
of a package before building it. We used to only care about "makedepends"
|
|
|
|
and it's still possible to ignore the depends with --ignore-depends.
|
|
|
|
|
|
|
|
:returns: list of dependency pkgnames (eg. ["sdl2", "sdl2_net"])
|
|
|
|
"""
|
2017-12-02 11:51:43 +00:00
|
|
|
# Read makedepends and depends
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
ret = list(apkbuild["makedepends"])
|
2018-06-04 19:07:52 +00:00
|
|
|
if "!check" not in apkbuild["options"]:
|
|
|
|
ret += apkbuild["checkdepends"]
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
if "ignore_depends" not in args or not args.ignore_depends:
|
|
|
|
ret += apkbuild["depends"]
|
2017-12-02 11:51:43 +00:00
|
|
|
ret = sorted(set(ret))
|
|
|
|
|
|
|
|
# Don't recurse forever when a package depends on itself (#948)
|
2021-04-27 02:56:42 +00:00
|
|
|
for pkgname in ([apkbuild["pkgname"]] +
|
|
|
|
list(apkbuild["subpackages"].keys())):
|
2017-12-02 11:51:43 +00:00
|
|
|
if pkgname in ret:
|
|
|
|
logging.verbose(apkbuild["pkgname"] + ": ignoring dependency on"
|
|
|
|
" itself: " + pkgname)
|
|
|
|
ret.remove(pkgname)
|
|
|
|
return ret
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
|
|
|
|
def build_depends(args, apkbuild, arch, strict):
|
|
|
|
"""
|
|
|
|
Get and build dependencies with verbose logging messages.
|
|
|
|
|
|
|
|
:returns: (depends, depends_built)
|
|
|
|
"""
|
|
|
|
# Get dependencies
|
|
|
|
pkgname = apkbuild["pkgname"]
|
|
|
|
depends = get_depends(args, apkbuild)
|
|
|
|
logging.verbose(pkgname + ": build/install dependencies: " +
|
|
|
|
", ".join(depends))
|
|
|
|
|
2019-03-15 08:54:27 +00:00
|
|
|
# --no-depends: check for binary packages
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
depends_built = []
|
2019-03-15 08:54:27 +00:00
|
|
|
if "no_depends" in args and args.no_depends:
|
2019-11-04 18:57:24 +00:00
|
|
|
pmb.helpers.repo.update(args, arch)
|
2019-03-15 08:54:27 +00:00
|
|
|
for depend in depends:
|
2021-12-06 19:30:58 +00:00
|
|
|
# Ignore conflicting dependencies
|
|
|
|
if depend.startswith("!"):
|
|
|
|
continue
|
2020-03-31 20:50:51 +00:00
|
|
|
# Check if binary package is missing
|
2019-03-15 08:54:27 +00:00
|
|
|
if not pmb.parse.apkindex.package(args, depend, arch, False):
|
|
|
|
raise RuntimeError("Missing binary package for dependency '" +
|
|
|
|
depend + "' of '" + pkgname + "', but"
|
|
|
|
" pmbootstrap won't build any depends since"
|
|
|
|
" it was started with --no-depends.")
|
2020-03-31 20:50:51 +00:00
|
|
|
# Check if binary package is outdated
|
|
|
|
apkbuild_dep = get_apkbuild(args, depend, arch)
|
2021-04-27 02:56:42 +00:00
|
|
|
if apkbuild_dep and \
|
|
|
|
pmb.build.is_necessary(args, arch, apkbuild_dep):
|
2020-03-31 20:50:51 +00:00
|
|
|
raise RuntimeError(f"Binary package for dependency '{depend}'"
|
|
|
|
f" of '{pkgname}' is outdated, but"
|
|
|
|
f" pmbootstrap won't build any depends"
|
|
|
|
f" since it was started with --no-depends.")
|
2019-03-15 08:54:27 +00:00
|
|
|
else:
|
|
|
|
# Build the dependencies
|
|
|
|
for depend in depends:
|
2021-12-06 19:30:58 +00:00
|
|
|
if depend.startswith("!"):
|
|
|
|
continue
|
2019-03-15 08:54:27 +00:00
|
|
|
if package(args, depend, arch, strict=strict):
|
|
|
|
depends_built += [depend]
|
|
|
|
logging.verbose(pkgname + ": build dependencies: done, built: " +
|
|
|
|
", ".join(depends_built))
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
return (depends, depends_built)
|
|
|
|
|
|
|
|
|
|
|
|
def is_necessary_warn_depends(args, apkbuild, arch, force, depends_built):
|
|
|
|
"""
|
|
|
|
Check if a build is necessary, and warn if it is not, but there were
|
|
|
|
dependencies built.
|
|
|
|
|
|
|
|
:returns: True or False
|
|
|
|
"""
|
|
|
|
pkgname = apkbuild["pkgname"]
|
2018-11-06 06:58:05 +00:00
|
|
|
|
|
|
|
# Check if necessary (this warns about binary version > aport version, so
|
|
|
|
# call it even in force mode)
|
|
|
|
ret = pmb.build.is_necessary(args, arch, apkbuild)
|
|
|
|
if force:
|
|
|
|
ret = True
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
if not ret and len(depends_built):
|
2021-04-11 16:07:26 +00:00
|
|
|
logging.verbose(f"{pkgname}: depends on rebuilt package(s): "
|
|
|
|
f" {', '.join(depends_built)}")
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
logging.verbose(pkgname + ": build necessary: " + str(ret))
|
|
|
|
return ret
|
|
|
|
|
|
|
|
|
|
|
|
def init_buildenv(args, apkbuild, arch, strict=False, force=False, cross=None,
|
2018-02-19 22:04:01 +00:00
|
|
|
suffix="native", skip_init_buildenv=False, src=None):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
|
|
|
Build all dependencies, check if we need to build at all (otherwise we've
|
|
|
|
just initialized the build environment for nothing) and then setup the
|
|
|
|
whole build environment (abuild, gcc, dependencies, cross-compiler).
|
|
|
|
|
2019-02-15 23:21:23 +00:00
|
|
|
:param cross: None, "native", "distcc", or "crossdirect"
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:param skip_init_buildenv: can be set to False to avoid initializing the
|
|
|
|
build environment. Use this when building
|
|
|
|
something during initialization of the build
|
|
|
|
environment (e.g. qemu aarch64 bug workaround)
|
2018-02-19 22:04:01 +00:00
|
|
|
:param src: override source used to build the package with a local folder
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:returns: True when the build is necessary (otherwise False)
|
|
|
|
"""
|
2018-11-30 08:31:38 +00:00
|
|
|
|
|
|
|
depends_arch = arch
|
|
|
|
if cross == "native":
|
2021-10-16 16:33:27 +00:00
|
|
|
depends_arch = pmb.config.arch_native
|
2018-11-30 08:31:38 +00:00
|
|
|
|
2019-03-15 08:51:41 +00:00
|
|
|
# Build dependencies
|
2018-11-30 08:31:38 +00:00
|
|
|
depends, built = build_depends(args, apkbuild, depends_arch, strict)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
2019-03-15 08:51:41 +00:00
|
|
|
# Check if build is necessary
|
2018-11-30 16:26:10 +00:00
|
|
|
if not is_necessary_warn_depends(args, apkbuild, arch, force, built):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
return False
|
|
|
|
|
2017-12-21 16:42:29 +00:00
|
|
|
# Install and configure abuild, ccache, gcc, dependencies
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
if not skip_init_buildenv:
|
|
|
|
pmb.build.init(args, suffix)
|
|
|
|
pmb.build.other.configure_abuild(args, suffix)
|
2017-12-21 16:42:29 +00:00
|
|
|
pmb.build.other.configure_ccache(args, suffix)
|
2019-03-27 21:31:49 +00:00
|
|
|
if not strict and "pmb:strict" not in apkbuild["options"] and len(depends):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
pmb.chroot.apk.install(args, depends, suffix)
|
2018-02-19 22:04:01 +00:00
|
|
|
if src:
|
|
|
|
pmb.chroot.apk.install(args, ["rsync"], suffix)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
# Cross-compiler init
|
|
|
|
if cross:
|
2021-03-10 07:29:25 +00:00
|
|
|
pmb.build.init_compiler(args, depends, cross, arch)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
if cross == "distcc":
|
|
|
|
pmb.chroot.distccd.start(args, arch)
|
2019-02-15 23:21:23 +00:00
|
|
|
if cross == "crossdirect":
|
|
|
|
pmb.chroot.mount_native_into_foreign(args, suffix)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
return True
|
|
|
|
|
|
|
|
|
2017-12-21 16:42:29 +00:00
|
|
|
def get_gcc_version(args, arch):
|
|
|
|
"""
|
|
|
|
Get the GCC version for a specific arch from parsing the right APKINDEX.
|
|
|
|
We feed this to ccache, so it knows the right GCC version, when
|
|
|
|
cross-compiling in a foreign arch chroot with distcc. See the "using
|
|
|
|
ccache with other compiler wrappers" section of their man page:
|
|
|
|
<https://linux.die.net/man/1/ccache>
|
|
|
|
:returns: a string like "6.4.0-r5"
|
|
|
|
"""
|
2018-03-04 13:44:27 +00:00
|
|
|
return pmb.parse.apkindex.package(args, "gcc-" + arch,
|
2021-10-16 16:33:27 +00:00
|
|
|
pmb.config.arch_native)["version"]
|
2017-12-21 16:42:29 +00:00
|
|
|
|
|
|
|
|
2018-02-19 22:04:01 +00:00
|
|
|
def get_pkgver(original_pkgver, original_source=False, now=None):
|
|
|
|
"""
|
|
|
|
Get the original pkgver when using the original source. Otherwise, get the
|
|
|
|
pkgver with an appended suffix of current date and time. For example:
|
|
|
|
_p20180218550502
|
|
|
|
When appending the suffix, an existing suffix (e.g. _git20171231) gets
|
|
|
|
replaced.
|
|
|
|
|
|
|
|
:param original_pkgver: unmodified pkgver from the package's APKBUILD.
|
|
|
|
:param original_source: the original source is used instead of overriding
|
|
|
|
it with --src.
|
|
|
|
:param now: use a specific date instead of current date (for test cases)
|
|
|
|
"""
|
|
|
|
if original_source:
|
|
|
|
return original_pkgver
|
|
|
|
|
|
|
|
# Append current date
|
|
|
|
no_suffix = original_pkgver.split("_", 1)[0]
|
|
|
|
now = now if now else datetime.datetime.now()
|
|
|
|
new_suffix = "_p" + now.strftime("%Y%m%d%H%M%S")
|
|
|
|
return no_suffix + new_suffix
|
|
|
|
|
|
|
|
|
|
|
|
def override_source(args, apkbuild, pkgver, src, suffix="native"):
|
|
|
|
"""
|
|
|
|
Mount local source inside chroot and append new functions (prepare() etc.)
|
|
|
|
to the APKBUILD to make it use the local source.
|
|
|
|
"""
|
|
|
|
if not src:
|
|
|
|
return
|
|
|
|
|
|
|
|
# Mount source in chroot
|
|
|
|
mount_path = "/mnt/pmbootstrap-source-override/"
|
|
|
|
mount_path_outside = args.work + "/chroot_" + suffix + mount_path
|
|
|
|
pmb.helpers.mount.bind(args, src, mount_path_outside, umount=True)
|
|
|
|
|
|
|
|
# Delete existing append file
|
|
|
|
append_path = "/tmp/APKBUILD.append"
|
|
|
|
append_path_outside = args.work + "/chroot_" + suffix + append_path
|
|
|
|
if os.path.exists(append_path_outside):
|
2018-03-05 21:04:28 +00:00
|
|
|
pmb.chroot.root(args, ["rm", append_path], suffix)
|
2018-02-19 22:04:01 +00:00
|
|
|
|
|
|
|
# Add src path to pkgdesc, cut it off after max length
|
|
|
|
pkgdesc = ("[" + src + "] " + apkbuild["pkgdesc"])[:127]
|
|
|
|
|
|
|
|
# Appended content
|
|
|
|
append = """
|
|
|
|
# ** Overrides below appended by pmbootstrap for --src **
|
|
|
|
|
|
|
|
pkgver=\"""" + pkgver + """\"
|
|
|
|
pkgdesc=\"""" + pkgdesc + """\"
|
|
|
|
_pmb_src_copy="/tmp/pmbootstrap-local-source-copy"
|
|
|
|
|
|
|
|
# Empty $source avoids patching in prepare()
|
|
|
|
_pmb_source_original="$source"
|
|
|
|
source=""
|
|
|
|
sha512sums=""
|
|
|
|
|
|
|
|
fetch() {
|
|
|
|
# Update source copy
|
|
|
|
msg "Copying source from host system: """ + src + """\"
|
|
|
|
rsync -a --exclude=".git/" --delete --ignore-errors --force \\
|
|
|
|
\"""" + mount_path + """\" "$_pmb_src_copy" || true
|
|
|
|
|
|
|
|
# Link local source files (e.g. kernel config)
|
|
|
|
mkdir "$srcdir"
|
|
|
|
local s
|
|
|
|
for s in $_pmb_source_original; do
|
|
|
|
is_remote "$s" || ln -sf "$startdir/$s" "$srcdir/"
|
|
|
|
done
|
|
|
|
}
|
|
|
|
|
|
|
|
unpack() {
|
|
|
|
ln -sv "$_pmb_src_copy" "$builddir"
|
|
|
|
}
|
|
|
|
"""
|
|
|
|
|
|
|
|
# Write and log append file
|
|
|
|
with open(append_path_outside, "w", encoding="utf-8") as handle:
|
|
|
|
for line in append.split("\n"):
|
|
|
|
handle.write(line[13:].replace(" " * 4, "\t") + "\n")
|
2018-03-05 21:04:28 +00:00
|
|
|
pmb.chroot.user(args, ["cat", append_path], suffix)
|
2018-02-19 22:04:01 +00:00
|
|
|
|
|
|
|
# Append it to the APKBUILD
|
|
|
|
apkbuild_path = "/home/pmos/build/APKBUILD"
|
|
|
|
shell_cmd = ("cat " + apkbuild_path + " " + append_path + " > " +
|
|
|
|
append_path + "_")
|
Properly escape commands in pmb.chroot.user() (#1316)
## Introduction
In #1302 we noticed that `pmb.chroot.user()` does not escape commands
properly: When passing one string with spaces, it would pass them as
two strings to the chroot. The use case is passing a description with
a space inside to `newapkbuild` with `pmboostrap newapkbuild`.
This is not a security issue, as we don't pass strings from untrusted
input to this function.
## Functions for running commands in pmbootstrap
To put the rest of the description in context: We have four high level
functions that run commands:
* `pmb.helpers.run.user()`
* `pmb.helpers.run.root()`
* `pmb.chroot.root()`
* `pmb.chroot.user()`
In addition, one low level function that the others invoke:
* `pmb.helpers.run.core()`
## Flawed test case
The issue described above did not get detected for so long, because we
have a test case in place since day one, which verifies that all of the
functions above escape everything properly:
* `test/test_shell_escape.py`
So the test case ran a given command through all these functions, and
compared the result each time. However, `pmb.chroot.root()`
modified the command variable (passed by reference) and did the
escaping already, which means `pmb.chroot.user()` running directly
afterwards only returns the right output when *not* doing any escaping.
Without questioning the accuracy of the test case, I've escaped
commands and environment variables with `shlex.quote()` *before*
passing them to `pmb.chroot.user()`. In retrospective this does not
make sense at all and is reverted with this commit.
## Environment variables
By coincidence, we have only passed custom environment variables to
`pmb.chroot.user()`, never to the other high level functions. This only
worked, because we did not do any escaping and the passed line gets
executed as shell command:
```
$ MYENV=test echo test2
test 2
```
If it was properly escaped as one shell command:
```
$ 'MYENV=test echo test2'
sh: MYENV=test echo test2: not found
```
So doing that clearly doesn't work anymore. I have added a new `env`
parameter to `pmb.chroot.user()` (and to all other high level functions
for consistency), where environment variables can be passed as a
dictionary. Then the function knows what to do and we end up with
properly escaped commands and environment variables.
## Details
* Add new `env` parameter to all high level command execution functions
* New `pmb.helpers.run.flat_cmd()` function, that takes a command as
list and environment variables as dict, and creates a properly escaped
flat string from the input.
* Use that function for proper escaping in all high level exec funcs
* Don't escape commands *before* passing them to `pmb.chroot.user()`
* Describe parameters of the command execution functions
* `pmbootstrap -v` writes the exact command to the log that was
executed (in addition to the simplified form we always write down for
readability)
* `test_shell_escape.py`: verify that the command passed by reference
has not been modified, add a new test for strings with spaces, add
tests for new function `pmb.helpers.run.flat_cmd()`
* Remove obsolete commend in `pmb.chroot.distccd` about environment
variables, because we don't use any there anymore
* Add `TERM=xterm` to default environment variables in the chroot,
so running ncurses applications like `menuconfig` and `nano` works out of
the box
2018-03-10 22:58:39 +00:00
|
|
|
pmb.chroot.user(args, ["sh", "-c", shell_cmd], suffix)
|
2018-03-05 21:04:28 +00:00
|
|
|
pmb.chroot.user(args, ["mv", append_path + "_", apkbuild_path], suffix)
|
2018-02-19 22:04:01 +00:00
|
|
|
|
|
|
|
|
pmbootstrap lint: avoid looping and copying files (MR 2100)
Before this commit, package folders were copied into the chroot one by
one in order to run apkbuild-lint on them. This logic is replaced by
mounting pmaports.git into the chroot and using a single apkbuild-lint
invocation to lint the supplied packages.
Both of these changes result in a performance improvement, especially
when linting multiple packages at once.
Before this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m5,261s
user 0m7,046s
sys 0m1,842s
Using the pmaports.git mount but calling apkbuild-lint in a loop:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m4,089s
user 0m6,418s
sys 0m1,219s
After this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m3,518s
user 0m5,968s
sys 0m0,959s
Additionally, running apkbuild-lint from the pmaports.git mount point
has the benefit that every printed violation contains a nice source
identifier à la "./cross/grub-x86/APKBUILD". This makes it possible to
differentiate between different packages even though only a single
apkbuild-lint invocation is used.
Relates: postmarketOS/pmaports#564
2021-08-30 07:24:46 +00:00
|
|
|
def mount_pmaports(args, destination, suffix="native"):
|
|
|
|
"""
|
|
|
|
Mount pmaports.git in chroot.
|
|
|
|
|
|
|
|
:param destination: mount point inside the chroot
|
|
|
|
"""
|
|
|
|
outside_destination = args.work + "/chroot_" + suffix + destination
|
|
|
|
pmb.helpers.mount.bind(args, args.aports, outside_destination, umount=True)
|
|
|
|
|
|
|
|
|
2019-11-10 10:20:00 +00:00
|
|
|
def link_to_git_dir(args, suffix):
|
|
|
|
"""
|
|
|
|
Make /home/pmos/build/.git point to the .git dir from pmaports.git, with a
|
|
|
|
symlink so abuild does not fail (#1841).
|
|
|
|
|
|
|
|
abuild expects the current working directory to be a subdirectory of a
|
|
|
|
cloned git repository (e.g. main/openrc from aports.git). If git is
|
|
|
|
installed, it will try to get the last git commit from that repository, and
|
|
|
|
place it in the resulting apk (.PKGINFO) as well as use the date from that
|
|
|
|
commit as SOURCE_DATE_EPOCH (for reproducible builds).
|
|
|
|
|
|
|
|
With that symlink, we actually make it use the last git commit from
|
|
|
|
pmaports.git for SOURCE_DATE_EPOCH and have that in the resulting apk's
|
|
|
|
.PKGINFO.
|
|
|
|
"""
|
|
|
|
# Mount pmaports.git in chroot, in case the user did not use pmbootstrap to
|
|
|
|
# clone it (e.g. how we build on sourcehut). Do this here and not at the
|
|
|
|
# initialization of the chroot, because the pmaports dir may not exist yet
|
|
|
|
# at that point. Use umount=True, so we don't have an old path mounted
|
|
|
|
# (some tests change the pmaports dir).
|
pmbootstrap lint: avoid looping and copying files (MR 2100)
Before this commit, package folders were copied into the chroot one by
one in order to run apkbuild-lint on them. This logic is replaced by
mounting pmaports.git into the chroot and using a single apkbuild-lint
invocation to lint the supplied packages.
Both of these changes result in a performance improvement, especially
when linting multiple packages at once.
Before this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m5,261s
user 0m7,046s
sys 0m1,842s
Using the pmaports.git mount but calling apkbuild-lint in a loop:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m4,089s
user 0m6,418s
sys 0m1,219s
After this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m3,518s
user 0m5,968s
sys 0m0,959s
Additionally, running apkbuild-lint from the pmaports.git mount point
has the benefit that every printed violation contains a nice source
identifier à la "./cross/grub-x86/APKBUILD". This makes it possible to
differentiate between different packages even though only a single
apkbuild-lint invocation is used.
Relates: postmarketOS/pmaports#564
2021-08-30 07:24:46 +00:00
|
|
|
destination = "/mnt/pmaports"
|
|
|
|
mount_pmaports(args, destination, suffix)
|
2019-11-10 10:20:00 +00:00
|
|
|
|
|
|
|
# Create .git symlink
|
|
|
|
pmb.chroot.user(args, ["mkdir", "-p", "/home/pmos/build"], suffix)
|
pmbootstrap lint: avoid looping and copying files (MR 2100)
Before this commit, package folders were copied into the chroot one by
one in order to run apkbuild-lint on them. This logic is replaced by
mounting pmaports.git into the chroot and using a single apkbuild-lint
invocation to lint the supplied packages.
Both of these changes result in a performance improvement, especially
when linting multiple packages at once.
Before this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m5,261s
user 0m7,046s
sys 0m1,842s
Using the pmaports.git mount but calling apkbuild-lint in a loop:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m4,089s
user 0m6,418s
sys 0m1,219s
After this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m3,518s
user 0m5,968s
sys 0m0,959s
Additionally, running apkbuild-lint from the pmaports.git mount point
has the benefit that every printed violation contains a nice source
identifier à la "./cross/grub-x86/APKBUILD". This makes it possible to
differentiate between different packages even though only a single
apkbuild-lint invocation is used.
Relates: postmarketOS/pmaports#564
2021-08-30 07:24:46 +00:00
|
|
|
pmb.chroot.user(args, ["ln", "-sf", destination + "/.git",
|
2019-11-10 10:20:00 +00:00
|
|
|
"/home/pmos/build/.git"], suffix)
|
|
|
|
|
|
|
|
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
def run_abuild(args, apkbuild, arch, strict=False, force=False, cross=None,
|
2018-02-19 22:04:01 +00:00
|
|
|
suffix="native", src=None):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
|
|
|
Set up all environment variables and construct the abuild command (all
|
|
|
|
depending on the cross-compiler method and target architecture), copy
|
|
|
|
the aport to the chroot and execute abuild.
|
|
|
|
|
2019-02-15 23:21:23 +00:00
|
|
|
:param cross: None, "native", "distcc", or "crossdirect"
|
2018-02-19 22:04:01 +00:00
|
|
|
:param src: override source used to build the package with a local folder
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:returns: (output, cmd, env), output is the destination apk path relative
|
|
|
|
to the package folder ("x86_64/hello-1-r2.apk"). cmd and env are
|
|
|
|
used by the test case, and they are the full abuild command and
|
|
|
|
the environment variables dict generated in this function.
|
|
|
|
"""
|
|
|
|
# Sanity check
|
|
|
|
if cross == "native" and "!tracedeps" not in apkbuild["options"]:
|
|
|
|
logging.info("WARNING: Option !tracedeps is not set, but we're"
|
|
|
|
" cross-compiling in the native chroot. This will"
|
|
|
|
" probably fail!")
|
|
|
|
|
|
|
|
# Pretty log message
|
2018-02-19 22:04:01 +00:00
|
|
|
pkgver = get_pkgver(apkbuild["pkgver"], src is None)
|
|
|
|
output = (arch + "/" + apkbuild["pkgname"] + "-" + pkgver +
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"-r" + apkbuild["pkgrel"] + ".apk")
|
2018-02-19 22:04:01 +00:00
|
|
|
message = "(" + suffix + ") build " + output
|
|
|
|
if src:
|
|
|
|
message += " (source: " + src + ")"
|
|
|
|
logging.info(message)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
# Environment variables
|
2018-02-09 18:43:58 +00:00
|
|
|
env = {"CARCH": arch,
|
|
|
|
"SUDO_APK": "abuild-apk --no-progress"}
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
if cross == "native":
|
|
|
|
hostspec = pmb.parse.arch.alpine_to_hostspec(arch)
|
|
|
|
env["CROSS_COMPILE"] = hostspec + "-"
|
|
|
|
env["CC"] = hostspec + "-gcc"
|
|
|
|
if cross == "distcc":
|
2017-12-21 16:42:29 +00:00
|
|
|
env["CCACHE_PREFIX"] = "distcc"
|
2021-04-27 02:56:42 +00:00
|
|
|
env["CCACHE_PATH"] = f"/usr/lib/arch-bin-masquerade/{arch}:/usr/bin"
|
2017-12-21 16:42:29 +00:00
|
|
|
env["CCACHE_COMPILERCHECK"] = "string:" + get_gcc_version(args, arch)
|
pmb: adjust to distcc 3.3 and wrap it with sshd
Overview:
Since Alpine updated to distcc 3.3 last week, pmbootstrap wasn't able to use
distcc for cross compilation anymore. It always falled back to running the
compiler in QEMU (which works, but is a lot slower). The reason for that is,
that distcc requires all compilers that are being used in a whitelist now.
This partially fixes CVE-2004-2687 in distccd, which allowed trivial remote
code execution by any process connecting to the distccd server. We only run
distccd on localhost, but still this can be used for privilege escalation of
sandboxed processes running on the host system (not part of pmbootstrap
chroots).
Because the CVE is only partially fixed (see the comment in
`pmb/chroot/distccd.py` for details), we make sure that only the building
chroots can talk to the distcc server by running distcc over ssh.
Details:
* Completely refactored `pmb/chroot/distccd.py` to run distcc over ssh
* Store the running distcc server's arguments as JSON now, not as INI
* Make debugging distcc issues easy:
* Set DISTCC_BACKOFF_PERIOD=0, so the distcc client will not ignore the
server after errors happened (this masks the original error!)
* New pmbootstrap parameters:
* `--distcc-nofallback`: avoids falling back to compiling with QEMU and not
throwing an error
* `--ccache-disable`: avoid ccache (when the compiler output is cached,
distcc does not get used)
* `--verbose` prints verbose output of the distcc too
* New test case, that uses the new pmbootstrap parameters to force
compilation through distcc, and shows the output of distcc and distccd in
verbose mode on error (as well as the log of sshd)
2018-07-25 19:09:45 +00:00
|
|
|
env["DISTCC_HOSTS"] = "@127.0.0.1:/home/pmos/.distcc-sshd/distccd"
|
|
|
|
env["DISTCC_SSH"] = ("ssh -o StrictHostKeyChecking=no -p" +
|
|
|
|
args.port_distccd)
|
|
|
|
env["DISTCC_BACKOFF_PERIOD"] = "0"
|
|
|
|
if not args.distcc_fallback:
|
|
|
|
env["DISTCC_FALLBACK"] = "0"
|
|
|
|
if args.verbose:
|
|
|
|
env["DISTCC_VERBOSE"] = "1"
|
2019-02-15 23:21:23 +00:00
|
|
|
if cross == "crossdirect":
|
2019-04-03 19:12:17 +00:00
|
|
|
env["PATH"] = ":".join(["/native/usr/lib/crossdirect/" + arch,
|
|
|
|
pmb.config.chroot_path])
|
pmb: adjust to distcc 3.3 and wrap it with sshd
Overview:
Since Alpine updated to distcc 3.3 last week, pmbootstrap wasn't able to use
distcc for cross compilation anymore. It always falled back to running the
compiler in QEMU (which works, but is a lot slower). The reason for that is,
that distcc requires all compilers that are being used in a whitelist now.
This partially fixes CVE-2004-2687 in distccd, which allowed trivial remote
code execution by any process connecting to the distccd server. We only run
distccd on localhost, but still this can be used for privilege escalation of
sandboxed processes running on the host system (not part of pmbootstrap
chroots).
Because the CVE is only partially fixed (see the comment in
`pmb/chroot/distccd.py` for details), we make sure that only the building
chroots can talk to the distcc server by running distcc over ssh.
Details:
* Completely refactored `pmb/chroot/distccd.py` to run distcc over ssh
* Store the running distcc server's arguments as JSON now, not as INI
* Make debugging distcc issues easy:
* Set DISTCC_BACKOFF_PERIOD=0, so the distcc client will not ignore the
server after errors happened (this masks the original error!)
* New pmbootstrap parameters:
* `--distcc-nofallback`: avoids falling back to compiling with QEMU and not
throwing an error
* `--ccache-disable`: avoid ccache (when the compiler output is cached,
distcc does not get used)
* `--verbose` prints verbose output of the distcc too
* New test case, that uses the new pmbootstrap parameters to force
compilation through distcc, and shows the output of distcc and distccd in
verbose mode on error (as well as the log of sshd)
2018-07-25 19:09:45 +00:00
|
|
|
if not args.ccache:
|
|
|
|
env["CCACHE_DISABLE"] = "1"
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
# Build the abuild command
|
2018-04-19 00:14:52 +00:00
|
|
|
cmd = ["abuild", "-D", "postmarketOS"]
|
2019-03-27 21:31:49 +00:00
|
|
|
if strict or "pmb:strict" in apkbuild["options"]:
|
|
|
|
if not strict:
|
|
|
|
logging.debug(apkbuild["pkgname"] + ": 'pmb:strict' found in"
|
|
|
|
" options, building in strict mode")
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
cmd += ["-r"] # install depends with abuild
|
|
|
|
else:
|
|
|
|
cmd += ["-d"] # do not install depends with abuild
|
|
|
|
if force:
|
|
|
|
cmd += ["-f"]
|
|
|
|
|
|
|
|
# Copy the aport to the chroot and build it
|
|
|
|
pmb.build.copy_to_buildpath(args, apkbuild["pkgname"], suffix)
|
2018-02-19 22:04:01 +00:00
|
|
|
override_source(args, apkbuild, pkgver, src, suffix)
|
2019-11-10 10:20:00 +00:00
|
|
|
link_to_git_dir(args, suffix)
|
Properly escape commands in pmb.chroot.user() (#1316)
## Introduction
In #1302 we noticed that `pmb.chroot.user()` does not escape commands
properly: When passing one string with spaces, it would pass them as
two strings to the chroot. The use case is passing a description with
a space inside to `newapkbuild` with `pmboostrap newapkbuild`.
This is not a security issue, as we don't pass strings from untrusted
input to this function.
## Functions for running commands in pmbootstrap
To put the rest of the description in context: We have four high level
functions that run commands:
* `pmb.helpers.run.user()`
* `pmb.helpers.run.root()`
* `pmb.chroot.root()`
* `pmb.chroot.user()`
In addition, one low level function that the others invoke:
* `pmb.helpers.run.core()`
## Flawed test case
The issue described above did not get detected for so long, because we
have a test case in place since day one, which verifies that all of the
functions above escape everything properly:
* `test/test_shell_escape.py`
So the test case ran a given command through all these functions, and
compared the result each time. However, `pmb.chroot.root()`
modified the command variable (passed by reference) and did the
escaping already, which means `pmb.chroot.user()` running directly
afterwards only returns the right output when *not* doing any escaping.
Without questioning the accuracy of the test case, I've escaped
commands and environment variables with `shlex.quote()` *before*
passing them to `pmb.chroot.user()`. In retrospective this does not
make sense at all and is reverted with this commit.
## Environment variables
By coincidence, we have only passed custom environment variables to
`pmb.chroot.user()`, never to the other high level functions. This only
worked, because we did not do any escaping and the passed line gets
executed as shell command:
```
$ MYENV=test echo test2
test 2
```
If it was properly escaped as one shell command:
```
$ 'MYENV=test echo test2'
sh: MYENV=test echo test2: not found
```
So doing that clearly doesn't work anymore. I have added a new `env`
parameter to `pmb.chroot.user()` (and to all other high level functions
for consistency), where environment variables can be passed as a
dictionary. Then the function knows what to do and we end up with
properly escaped commands and environment variables.
## Details
* Add new `env` parameter to all high level command execution functions
* New `pmb.helpers.run.flat_cmd()` function, that takes a command as
list and environment variables as dict, and creates a properly escaped
flat string from the input.
* Use that function for proper escaping in all high level exec funcs
* Don't escape commands *before* passing them to `pmb.chroot.user()`
* Describe parameters of the command execution functions
* `pmbootstrap -v` writes the exact command to the log that was
executed (in addition to the simplified form we always write down for
readability)
* `test_shell_escape.py`: verify that the command passed by reference
has not been modified, add a new test for strings with spaces, add
tests for new function `pmb.helpers.run.flat_cmd()`
* Remove obsolete commend in `pmb.chroot.distccd` about environment
variables, because we don't use any there anymore
* Add `TERM=xterm` to default environment variables in the chroot,
so running ncurses applications like `menuconfig` and `nano` works out of
the box
2018-03-10 22:58:39 +00:00
|
|
|
pmb.chroot.user(args, cmd, suffix, "/home/pmos/build", env=env)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
return (output, cmd, env)
|
|
|
|
|
|
|
|
|
2018-02-01 22:03:21 +00:00
|
|
|
def finish(args, apkbuild, arch, output, strict=False, suffix="native"):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
|
|
|
Various finishing tasks that need to be done after a build.
|
|
|
|
"""
|
|
|
|
# Verify output file
|
2020-04-10 13:21:23 +00:00
|
|
|
channel = pmb.config.pmaports.read_config(args)["channel"]
|
|
|
|
path = f"{args.work}/packages/{channel}/{output}"
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
if not os.path.exists(path):
|
|
|
|
raise RuntimeError("Package not found after build: " + path)
|
|
|
|
|
|
|
|
# Clear APKINDEX cache (we only parse APKINDEX files once per session and
|
|
|
|
# cache the result for faster dependency resolving, but after we built a
|
|
|
|
# package we need to parse it again)
|
2021-11-09 11:54:07 +00:00
|
|
|
pmb.parse.apkindex.clear_cache(f"{args.work}/packages/{channel}"
|
|
|
|
f"/{arch}/APKINDEX.tar.gz")
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
# Uninstall build dependencies (strict mode)
|
2019-03-27 21:31:49 +00:00
|
|
|
if strict or "pmb:strict" in apkbuild["options"]:
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
logging.info("(" + suffix + ") uninstall build dependencies")
|
Properly escape commands in pmb.chroot.user() (#1316)
## Introduction
In #1302 we noticed that `pmb.chroot.user()` does not escape commands
properly: When passing one string with spaces, it would pass them as
two strings to the chroot. The use case is passing a description with
a space inside to `newapkbuild` with `pmboostrap newapkbuild`.
This is not a security issue, as we don't pass strings from untrusted
input to this function.
## Functions for running commands in pmbootstrap
To put the rest of the description in context: We have four high level
functions that run commands:
* `pmb.helpers.run.user()`
* `pmb.helpers.run.root()`
* `pmb.chroot.root()`
* `pmb.chroot.user()`
In addition, one low level function that the others invoke:
* `pmb.helpers.run.core()`
## Flawed test case
The issue described above did not get detected for so long, because we
have a test case in place since day one, which verifies that all of the
functions above escape everything properly:
* `test/test_shell_escape.py`
So the test case ran a given command through all these functions, and
compared the result each time. However, `pmb.chroot.root()`
modified the command variable (passed by reference) and did the
escaping already, which means `pmb.chroot.user()` running directly
afterwards only returns the right output when *not* doing any escaping.
Without questioning the accuracy of the test case, I've escaped
commands and environment variables with `shlex.quote()` *before*
passing them to `pmb.chroot.user()`. In retrospective this does not
make sense at all and is reverted with this commit.
## Environment variables
By coincidence, we have only passed custom environment variables to
`pmb.chroot.user()`, never to the other high level functions. This only
worked, because we did not do any escaping and the passed line gets
executed as shell command:
```
$ MYENV=test echo test2
test 2
```
If it was properly escaped as one shell command:
```
$ 'MYENV=test echo test2'
sh: MYENV=test echo test2: not found
```
So doing that clearly doesn't work anymore. I have added a new `env`
parameter to `pmb.chroot.user()` (and to all other high level functions
for consistency), where environment variables can be passed as a
dictionary. Then the function knows what to do and we end up with
properly escaped commands and environment variables.
## Details
* Add new `env` parameter to all high level command execution functions
* New `pmb.helpers.run.flat_cmd()` function, that takes a command as
list and environment variables as dict, and creates a properly escaped
flat string from the input.
* Use that function for proper escaping in all high level exec funcs
* Don't escape commands *before* passing them to `pmb.chroot.user()`
* Describe parameters of the command execution functions
* `pmbootstrap -v` writes the exact command to the log that was
executed (in addition to the simplified form we always write down for
readability)
* `test_shell_escape.py`: verify that the command passed by reference
has not been modified, add a new test for strings with spaces, add
tests for new function `pmb.helpers.run.flat_cmd()`
* Remove obsolete commend in `pmb.chroot.distccd` about environment
variables, because we don't use any there anymore
* Add `TERM=xterm` to default environment variables in the chroot,
so running ncurses applications like `menuconfig` and `nano` works out of
the box
2018-03-10 22:58:39 +00:00
|
|
|
pmb.chroot.user(args, ["abuild", "undeps"], suffix, "/home/pmos/build",
|
|
|
|
env={"SUDO_APK": "abuild-apk --no-progress"})
|
2021-09-09 21:32:14 +00:00
|
|
|
# If the build depends contain postmarketos-keys or postmarketos-base,
|
|
|
|
# abuild will have removed the postmarketOS repository key (pma#1230)
|
|
|
|
pmb.chroot.init_keys(args)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
|
|
|
|
|
2018-02-01 22:03:21 +00:00
|
|
|
def package(args, pkgname, arch=None, force=False, strict=False,
|
2018-02-19 22:04:01 +00:00
|
|
|
skip_init_buildenv=False, src=None):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
"""
|
|
|
|
Build a package and its dependencies with Alpine Linux' abuild.
|
|
|
|
|
2020-12-09 15:16:02 +00:00
|
|
|
If this function is called multiple times on the same pkgname but first
|
|
|
|
with force=False and then force=True the force argument will be ignored due
|
|
|
|
to the package cache.
|
|
|
|
See the skip_already_built() call below.
|
|
|
|
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:param pkgname: package name to be built, as specified in the APKBUILD
|
|
|
|
:param arch: architecture we're building for (default: native)
|
2020-12-09 15:16:02 +00:00
|
|
|
:param force: always build, even if not necessary
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:param strict: avoid building with irrelevant dependencies installed by
|
|
|
|
letting abuild install and uninstall all dependencies.
|
|
|
|
:param skip_init_buildenv: can be set to False to avoid initializing the
|
|
|
|
build environment. Use this when building
|
|
|
|
something during initialization of the build
|
|
|
|
environment (e.g. qemu aarch64 bug workaround)
|
2018-02-19 22:04:01 +00:00
|
|
|
:param src: override source used to build the package with a local folder
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
:returns: None if the build was not necessary
|
|
|
|
output path relative to the packages folder ("armhf/ab-1-r2.apk")
|
|
|
|
"""
|
2017-12-05 23:17:39 +00:00
|
|
|
# Once per session is enough
|
2021-10-16 16:33:27 +00:00
|
|
|
arch = arch or pmb.config.arch_native
|
2021-11-09 11:54:07 +00:00
|
|
|
if skip_already_built(pkgname, arch):
|
2017-12-05 23:17:39 +00:00
|
|
|
return
|
|
|
|
|
|
|
|
# Only build when APKBUILD exists
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
apkbuild = get_apkbuild(args, pkgname, arch)
|
|
|
|
if not apkbuild:
|
|
|
|
return
|
|
|
|
|
|
|
|
# Detect the build environment (skip unnecessary builds)
|
2019-03-10 00:28:10 +00:00
|
|
|
if not check_build_for_arch(args, pkgname, arch):
|
|
|
|
return
|
2021-10-16 16:33:27 +00:00
|
|
|
suffix = pmb.build.autodetect.suffix(apkbuild, arch)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
cross = pmb.build.autodetect.crosscompile(args, apkbuild, arch, suffix)
|
|
|
|
if not init_buildenv(args, apkbuild, arch, strict, force, cross, suffix,
|
2018-02-19 22:04:01 +00:00
|
|
|
skip_init_buildenv, src):
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
return
|
|
|
|
|
|
|
|
# Build and finish up
|
|
|
|
(output, cmd, env) = run_abuild(args, apkbuild, arch, strict, force, cross,
|
2018-02-19 22:04:01 +00:00
|
|
|
suffix, src)
|
2018-02-01 22:03:21 +00:00
|
|
|
finish(args, apkbuild, arch, output, strict, suffix)
|
Fix #824: Refactor pmb/build/package.py (make depends work like in abuild) (#935)
* Rename pmb/build/package.py to pmb/build/_package.py, so we can
access the functions it contains in testcases, and still use
pmb.build.package()
* Refactor the entire file. Instead of one big function that does
too many things, we have many small ones now, that are tested
in the testsuite and easier to modify
* Whenever building a package, pmbootstrap does not only build and
install the "makedepends" (like we did before), now it does the
same for the "depends". That's required to be compatible with
abuild. The old behavior can still be used with 'pmbootstrap
build --ignore-depends'.
* Because of that change, noarch packages can no longer be built in
the native chroot if we need them for a foreign chroot. A device-
package depending on a kernel would pull in the same kernel for
the native architecture otherwise.
* Running 'pmbootstrap build device-...' without '--ignore-depends'
and without a matching '--arch' displays a note that explains
this change to the user and tells how to use it instead.
* Noarch packages no longer get symlinked. That was only
implemented for packages built in the native chroot, and now that
is not always the case anymore. Symlinking these packages creates
packages with broken dependencies anyway (e.g.
device-samsung-i9100 can't be installed in x86_64, because
linux-samsung-i9100 is armhf only).
* Rename "carch" to "arch" wherever used. Naming it "carch"
sometimes is confusing with no benefit.
* Add a testcase for the aarch64 qemu workaround (because it failed
first and I needed to know for sure if it is working again).
* Improved some verbose logging, which helped with development of
this feature.
* Removed the old "build" test case (which was disabled in
testcases_fast.sh) as the new "build_package" test case covers its
functionallity.
* Only build indexes if the packages folder exists for that arch (Travis
couldn't run a test case otherwise)
2017-11-26 14:32:02 +00:00
|
|
|
return output
|