2017-05-26 20:08:45 +00:00
|
|
|
"""
|
|
|
|
Copyright 2017 Oliver Smith
|
|
|
|
|
|
|
|
This file is part of pmbootstrap.
|
|
|
|
|
|
|
|
pmbootstrap is free software: you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
the Free Software Foundation, either version 3 of the License, or
|
|
|
|
(at your option) any later version.
|
|
|
|
|
|
|
|
pmbootstrap is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with pmbootstrap. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
"""
|
|
|
|
import distutils.version
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
import logging
|
2017-05-26 20:08:45 +00:00
|
|
|
import os
|
|
|
|
import tarfile
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
import pmb.chroot.apk
|
|
|
|
import pmb.helpers.repo
|
2017-05-26 20:08:45 +00:00
|
|
|
|
|
|
|
|
|
|
|
def compare_version(a_str, b_str):
|
|
|
|
"""
|
|
|
|
http://stackoverflow.com/a/11887885
|
|
|
|
LooseVersion behaves just like apk's version check, at least
|
|
|
|
for all package versions, that have "-r".
|
|
|
|
|
|
|
|
:returns:
|
|
|
|
(a < b): -1
|
|
|
|
(a == b): 0
|
|
|
|
(a > b): 1
|
|
|
|
"""
|
|
|
|
if a_str is None:
|
|
|
|
a_str = "0"
|
|
|
|
if b_str is None:
|
|
|
|
b_str = "0"
|
|
|
|
a = distutils.version.LooseVersion(a_str)
|
|
|
|
b = distutils.version.LooseVersion(b_str)
|
|
|
|
if a < b:
|
|
|
|
return -1
|
|
|
|
if a == b:
|
|
|
|
return 0
|
|
|
|
return 1
|
|
|
|
|
|
|
|
|
2017-06-17 22:46:14 +00:00
|
|
|
def parse_next_block(args, path, lines, start):
|
|
|
|
"""
|
|
|
|
Parse the next block in an APKINDEX.
|
|
|
|
|
|
|
|
:param path: to the APKINDEX.tar.gz
|
|
|
|
:param start: current index in lines, gets increased in this
|
|
|
|
function. Wrapped into a list, so it can be modified
|
|
|
|
"by reference". Example: [5]
|
|
|
|
:param lines: all lines from the "APKINDEX" file inside the archive
|
|
|
|
:returns: a dictionary with the following structure:
|
|
|
|
{ "pkgname": "postmarketos-mkinitfs",
|
|
|
|
"version": "0.0.4-r10",
|
|
|
|
"depends": ["busybox-extras", "lddtree", ... ],
|
|
|
|
"provides": ["mkinitfs=0.0.1"],
|
|
|
|
}
|
|
|
|
:returns: None, when there are no more blocks
|
|
|
|
"""
|
|
|
|
|
|
|
|
# Parse until we hit an empty line or end of file
|
|
|
|
ret = {}
|
|
|
|
mapping = {
|
|
|
|
"P": "pkgname",
|
|
|
|
"V": "version",
|
|
|
|
"D": "depends",
|
2017-06-20 18:13:05 +00:00
|
|
|
"p": "provides",
|
|
|
|
"t": "timestamp"
|
2017-06-17 22:46:14 +00:00
|
|
|
}
|
|
|
|
end_of_block_found = False
|
|
|
|
for i in range(start[0], len(lines)):
|
|
|
|
# Check for empty line
|
|
|
|
start[0] = i + 1
|
2017-06-27 15:56:15 +00:00
|
|
|
line = lines[i]
|
|
|
|
if not isinstance(line, str):
|
|
|
|
line = line.decode()
|
2017-06-17 22:46:14 +00:00
|
|
|
if line == "\n":
|
|
|
|
end_of_block_found = True
|
|
|
|
break
|
|
|
|
|
|
|
|
# Parse keys from the mapping
|
|
|
|
for letter, key in mapping.items():
|
|
|
|
if line.startswith(letter + ":"):
|
|
|
|
if key in ret:
|
|
|
|
raise RuntimeError(
|
|
|
|
"Key " + key + " (" + letter + ":) specified twice"
|
|
|
|
" in block: " + str(ret) + ", file: " + path)
|
|
|
|
ret[key] = line[2:-1]
|
|
|
|
|
|
|
|
# Format and return the block
|
|
|
|
if end_of_block_found:
|
|
|
|
# Check for required keys
|
2017-06-20 18:13:05 +00:00
|
|
|
for key in ["pkgname", "version", "timestamp"]:
|
2017-06-17 22:46:14 +00:00
|
|
|
if key not in ret:
|
|
|
|
raise RuntimeError("Missing required key '" + key +
|
|
|
|
"' in block " + str(ret) + ", file: " + path)
|
|
|
|
|
|
|
|
# Format optional lists
|
|
|
|
for key in ["provides", "depends"]:
|
|
|
|
if key in ret and ret[key] != "":
|
2017-07-22 09:54:49 +00:00
|
|
|
# Ignore all operators for now
|
|
|
|
values = ret[key].split(" ")
|
|
|
|
ret[key] = []
|
|
|
|
for value in values:
|
|
|
|
if value.startswith("!"):
|
|
|
|
continue
|
|
|
|
for operator in [">", "=", "<"]:
|
|
|
|
if operator in value:
|
|
|
|
value = value.split(operator)[0]
|
|
|
|
break
|
|
|
|
ret[key].append(value)
|
2017-06-17 22:46:14 +00:00
|
|
|
else:
|
|
|
|
ret[key] = []
|
|
|
|
return ret
|
|
|
|
|
|
|
|
# No more blocks
|
|
|
|
elif ret != {}:
|
|
|
|
raise RuntimeError("Last block in " + path + " does not end"
|
|
|
|
" with a new line! Delete the file and"
|
|
|
|
" try again. Last block: " + str(ret))
|
|
|
|
return None
|
|
|
|
|
|
|
|
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
def parse_add_block(path, strict, ret, block, pkgname=None):
|
2017-06-17 22:46:14 +00:00
|
|
|
"""
|
|
|
|
Add one block to the return dictionary of parse().
|
|
|
|
|
|
|
|
:param path: to the APKINDEX.tar.gz
|
|
|
|
:param strict: When set to True, only allow one entry per pkgname.
|
|
|
|
In case there are two, raise an exception.
|
|
|
|
When set to False, and there are multiple entries
|
|
|
|
for one pkgname, it uses the latest one.
|
|
|
|
:param ret: dictionary of all packages in the APKINDEX, that is
|
|
|
|
getting built right now. This function will extend it.
|
|
|
|
:param block: return value from parse_next_block().
|
|
|
|
:param pkgname: defaults to the real pkgname, could be an alias
|
|
|
|
from the "provides" list.
|
|
|
|
:param version: defaults to the real version, could be a value
|
|
|
|
from the "provides" list.
|
|
|
|
"""
|
|
|
|
|
|
|
|
# Defaults
|
|
|
|
if not pkgname:
|
|
|
|
pkgname = block["pkgname"]
|
|
|
|
|
|
|
|
# Handle duplicate entries
|
|
|
|
if pkgname in ret:
|
|
|
|
if strict:
|
2017-06-18 20:15:36 +00:00
|
|
|
raise RuntimeError("Multiple blocks for " + pkgname +
|
|
|
|
" in " + path)
|
|
|
|
# Ignore the block, if the block we already have has a higher
|
|
|
|
# version
|
|
|
|
version_old = ret[pkgname]["version"]
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
version_new = block["version"]
|
2017-06-18 20:15:36 +00:00
|
|
|
if compare_version(version_old, version_new) == 1:
|
2017-06-17 22:46:14 +00:00
|
|
|
return
|
|
|
|
|
|
|
|
# Add it to the result set
|
|
|
|
ret[pkgname] = block
|
|
|
|
|
|
|
|
|
|
|
|
def parse(args, path, strict=False):
|
|
|
|
"""
|
|
|
|
Parse an APKINDEX.tar.gz file, and return its content as dictionary.
|
|
|
|
|
|
|
|
:param strict: When set to True, only allow one entry per pkgname.
|
|
|
|
In case there are two, raise an exception.
|
|
|
|
When set to False, and there are multiple entries
|
|
|
|
for one pkgname, it uses the latest one.
|
|
|
|
:returns: a dictionary with the following structure:
|
|
|
|
{ "postmarketos-mkinitfs":
|
|
|
|
{
|
|
|
|
"pkgname": "postmarketos-mkinitfs"
|
|
|
|
"version": "0.0.4-r10",
|
|
|
|
"depends": ["busybox-extras", "lddtree", ...],
|
|
|
|
"provides": ["mkinitfs=0.0.1"]
|
|
|
|
}, ...
|
|
|
|
}
|
|
|
|
"""
|
2017-06-17 23:09:21 +00:00
|
|
|
|
|
|
|
# Try to get a cached result first
|
|
|
|
lastmod = os.path.getmtime(path)
|
|
|
|
if path in args.cache["apkindex"]:
|
|
|
|
cache = args.cache["apkindex"][path]
|
|
|
|
if cache["lastmod"] == lastmod:
|
|
|
|
return cache["ret"]
|
|
|
|
|
2017-06-27 15:56:15 +00:00
|
|
|
# Read all lines
|
|
|
|
if tarfile.is_tarfile(path):
|
|
|
|
with tarfile.open(path, "r:gz") as tar:
|
|
|
|
with tar.extractfile(tar.getmember("APKINDEX")) as handle:
|
|
|
|
lines = handle.readlines()
|
|
|
|
else:
|
|
|
|
with open(path, "r", encoding="utf-8") as handle:
|
|
|
|
lines = handle.readlines()
|
|
|
|
|
|
|
|
# Parse the whole APKINDEX file
|
2017-06-17 22:46:14 +00:00
|
|
|
ret = {}
|
|
|
|
start = [0]
|
2017-06-27 15:56:15 +00:00
|
|
|
while True:
|
|
|
|
block = parse_next_block(args, path, lines, start)
|
|
|
|
if not block:
|
|
|
|
break
|
|
|
|
|
|
|
|
# Add the next package and all aliases
|
|
|
|
parse_add_block(path, strict, ret, block)
|
|
|
|
if "provides" in block:
|
|
|
|
for alias in block["provides"]:
|
2017-07-22 09:54:49 +00:00
|
|
|
parse_add_block(path, strict, ret, block, alias)
|
2017-06-27 15:56:15 +00:00
|
|
|
|
2017-06-17 23:09:21 +00:00
|
|
|
# Update the cache
|
|
|
|
args.cache["apkindex"][path] = {"lastmod": lastmod, "ret": ret}
|
|
|
|
|
2017-06-17 22:46:14 +00:00
|
|
|
return ret
|
|
|
|
|
|
|
|
|
2017-05-26 20:08:45 +00:00
|
|
|
def read(args, package, path, must_exist=True):
|
|
|
|
"""
|
2017-06-17 22:46:14 +00:00
|
|
|
Get information about a single package from an APKINDEX.tar.gz file.
|
|
|
|
|
2017-05-26 20:08:45 +00:00
|
|
|
:param path: Path to APKINDEX.tar.gz, defaults to $WORK/APKINDEX.tar.gz
|
|
|
|
:param package: The package of which you want to read the properties.
|
|
|
|
:param must_exist: When set to true, raise an exception when the package is
|
|
|
|
missing in the index, or the index file was not found.
|
|
|
|
:returns: {"pkgname": ..., "version": ..., "depends": [...]}
|
|
|
|
When the package appears multiple times in the APKINDEX, this
|
|
|
|
function returns the attributes of the latest version.
|
|
|
|
"""
|
|
|
|
# Verify APKINDEX path
|
|
|
|
if not os.path.exists(path):
|
|
|
|
if not must_exist:
|
|
|
|
return None
|
|
|
|
raise RuntimeError("File not found: " + path)
|
|
|
|
|
2017-06-17 22:46:14 +00:00
|
|
|
# Parse the APKINDEX
|
|
|
|
apkindex = parse(args, path)
|
|
|
|
if package not in apkindex:
|
|
|
|
if must_exist:
|
|
|
|
raise RuntimeError("Package '" + package +
|
|
|
|
"' not found in " + path)
|
|
|
|
else:
|
|
|
|
return None
|
|
|
|
return apkindex[package]
|
2017-05-26 20:08:45 +00:00
|
|
|
|
|
|
|
|
|
|
|
def read_any_index(args, package, arch=None):
|
|
|
|
"""
|
2017-06-17 22:46:14 +00:00
|
|
|
Get information about a single package from any APKINDEX.tar.gz.
|
2017-05-26 20:08:45 +00:00
|
|
|
|
|
|
|
:param arch: defaults to native architecture
|
2017-06-17 22:46:14 +00:00
|
|
|
:returns: the same format as read()
|
2017-05-26 20:08:45 +00:00
|
|
|
"""
|
|
|
|
if not arch:
|
|
|
|
arch = args.arch_native
|
|
|
|
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
# Return first match
|
|
|
|
for index in pmb.helpers.repo.apkindex_files(args, arch):
|
2017-05-26 20:08:45 +00:00
|
|
|
index_data = read(args, package, index, False)
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
logging.verbose("Search for " + package + " in " + index +
|
|
|
|
" - result: " + str(index_data))
|
2017-05-26 20:08:45 +00:00
|
|
|
if index_data:
|
|
|
|
return index_data
|
Properly rebuild/install packages when something changed (Fix #120, #108, #131) (#129)
TLDR: Always rebuild/install packages when something changed when executing "pmbootstrap install/initfs/flash", more speed in dependency resolution.
---
pmbootstrap has already gotten some support for "timestamp based rebuilds", which modifies the logic for when packages should be rebuilt. It doesn't only consider packages outdated with old pkgver/pkgrel combinations, but also packages, where a source file has a newer timestamp, than the built package has.
I've found out, that this can lead to more rebuilds than expected. For example, when you check out the pmbootstrap git repository again into another folder, although you have already built packages. Then all files have the timestamp of the checkout, and the packages will appear to be outdated. While this is not largely a concern now, this will become a problem once we have a binary package repository, because then the packages from the binary repo will always seem to be outdated, if you just freshly checked out the repository.
To combat this, git gets asked if the files from the aport we're looking at are in sync with upstream, or not. Only when the files are not in sync with upstream and the timestamps of the sources are newer, a rebuild gets triggered from now on.
In case this logic should fail, I've added an option during "pmbootstrap init" where you can enable or disable the "timestamp based rebuilds" option.
In addition to that, this commit also works on fixing #120: packages do not get updated in "pmbootstrap install" after they have been rebuilt. For this to work, we specify all packages explicitly for abuild, instead of letting abuild do the resolving. This feature will also work with the "timestamp based rebuilds".
This commit also fixes the working_dir argument in pmb.helpers.run.user, which was simply ignored before.
Finally, the performance of the dependency resolution is faster again (when compared to the current version in master), because the parsed apkbuilds and finding the aport by pkgname gets cached during one pmbootstrap call (in args.cache, which also makes it easy to put fake data there in testcases).
The new dependency resolution code can output lots of verbose messages for debugging by specifying the `-v` parameter. The meaning of that changed, it used to output the file names where log messages come from, but no one seemed to use that anyway.
2017-07-10 15:23:43 +00:00
|
|
|
|
|
|
|
logging.verbose("No match found in any APKINDEX.tar.gz!")
|
2017-05-26 20:08:45 +00:00
|
|
|
return None
|