e03c8267dd
Update federation to match latest paper-journal work: proper propagation of Add/Del
2019-05-20 21:49:19 +01:00
8aaf1d1902
Improve distributed debug output
2019-05-20 21:45:40 +01:00
43a0578f19
Buffer uplink communications to give a chance for stateful entities at each end to initialise themselves before sending messages.
...
An alternative approach would be to reintroduce flow-control, which was
removed as part of commit 94e27e1
2019-05-16 23:14:39 +01:00
8daff63e8e
Fix (observe (server-envelope A B)) => (server-proposal A (observe B))
2019-05-16 23:13:56 +01:00
94e27e1cbc
Change federation protocol to be transport-neutral, carried via Syndicate itself
2019-05-16 22:28:42 +01:00
84361dcbaf
Move away from use of "broker"
2019-05-16 20:14:19 +01:00
6952aec992
Remove federation-prototype
2019-05-16 15:48:51 +01:00
cd2d7e428b
Close input/output ports in web.rkt once connection is logically over
2019-05-16 14:58:33 +01:00
025f101d09
Disconnect websocket service on error
2019-05-16 13:12:21 +01:00
c9b62c9a6d
Drain ws-ch in an additional, previously-overlooked case
2019-05-16 12:39:16 +01:00
22a517f1b5
Handle "Ping" requests from peers in federated mode
2019-05-16 12:38:26 +01:00
7ec9752f41
Accept and discard "Pong" replies from clients
2019-05-16 12:38:06 +01:00
af40134c53
Correct server websocket http-request pattern
2019-05-16 12:37:39 +01:00
ae97fb1d1e
Flow-controlled TCP (except in netstack); flow-controlled server/federation messages; fixes federation startup glitching
2019-05-12 13:07:38 +01:00
12650e2058
Repair test compilation
2019-05-09 11:43:41 +01:00
e9b1983408
Move command-line scripts into submodule main, so tests don't get stuck
2019-05-09 11:43:34 +01:00
7f7a6d03d8
Command-line programs for starting various kinds of server
2019-05-09 11:39:49 +01:00
386a66747e
Tweak logging levels
2019-05-09 11:39:49 +01:00
2b882cc10a
Integrate federation/peering with client/server code; almost but not quite working yet
2019-05-09 11:39:49 +01:00
0dca5bfbe3
Allow specification of host/port/scope in server-chat-client.rkt
2019-05-09 11:16:29 +01:00
88ec80487d
Remove unused definition
2019-05-07 14:06:06 +01:00
39906803b2
Move prototype federation code out of the way
2019-05-07 13:45:52 +01:00
cf92ae14c5
Select client/server scope on connect; stub out federation/peering; protocol error and disconnection support
2019-05-07 12:56:40 +01:00
0ac2bb768e
New "test" illustrating aspects of the turn-based approach taken
2019-05-07 12:07:30 +01:00
db523a8974
The Great Renaming broker -> server, part 2
2019-05-05 16:51:23 +01:00
f2b29a798a
The Great Renaming broker -> server, part 1
2019-05-05 16:37:03 +01:00
a08574a403
Remove unnecessary require
2019-05-05 15:58:01 +01:00
d5ff1dd007
Use definitions from wire-protocol.rkt in federation
2019-05-05 15:57:51 +01:00
aad3fb6866
Move federation protocol a step closer to client/server protocol
2019-05-05 15:54:28 +01:00
924512f7de
Make internal broker isolation protocol asymmetric, to support the needs of federation
2019-05-05 12:55:16 +01:00
a3a229532a
add-observer-endpoint!, add-raw-observer-endpoint!
2019-05-04 22:58:45 +01:00
dabc74976a
Refine output of tf.rkt
2019-05-04 22:48:37 +01:00
a86478b7b9
Remove unneeded #:capture-projection argument to term->skeleton-interest
2019-05-03 20:29:40 +01:00
40ee3b4ca7
Test and fix for observation visibility in broker
2019-05-03 17:53:24 +01:00
83c81fa293
Rename federation2.rkt -> federation.rkt
2019-05-03 13:45:05 +01:00
508e37171d
Bounce relay
2019-05-03 11:20:22 +01:00
ed7cf375ad
Only remove an entry from specs when no holders remain
2019-05-03 11:20:12 +01:00
736c9de3dc
Note on semantic difference
2019-05-03 10:32:16 +01:00
f847f3c049
Fix self- and peer-signalling of observation; generalize broker protocol to permit multiple uses of leaf2 on different nodes; printf -> log-info
2019-05-03 10:25:20 +01:00
38e3894a3d
Clean up conn-subs on Unsubscribe
2019-05-03 10:23:08 +01:00
1c2bd11373
Connect router to dataspace
2019-05-03 00:10:15 +01:00
79e8d54a51
Avoid silly use of macro
2019-05-02 23:12:02 +01:00
17d5b88784
Federation kernel
2019-05-02 15:59:25 +01:00
4fd2c9895c
Remove old, dead code
2019-04-25 12:58:48 +01:00
419bd093d1
Support transformations (e.g. rotation) in the scene prelude
2019-04-25 12:58:30 +01:00
759bbdf1c3
Add visibility-restriction, making the test pass
2019-03-25 16:32:09 +00:00
9e923e1c63
Failing test case: need visibility-restriction in broker clients
2019-03-25 12:16:30 +00:00
eb33fdbd59
Steps toward activation in test cases
2019-03-25 12:15:09 +00:00
a39bd458d9
Restructure broker protocol adapters; add loopback
2019-03-25 12:14:35 +00:00
25970d9f16
Split out server-connection.rkt
2019-03-24 17:17:26 +00:00
8a5d91a175
strong-gensym
2019-03-24 17:16:57 +00:00
025289500a
Configurable broker tcp/http ports
2019-03-22 12:51:59 +00:00
4732650041
module+ main in reload.rkt
2019-03-20 22:36:13 +00:00
63ea6a9d3a
extend-ground-boot!
2019-03-20 22:36:05 +00:00
3ff50a6f65
broker: client; ping for keepalive; client example
2019-03-18 23:29:12 +00:00
4e33429b19
tcp-rejected; reassert-on
2019-03-18 23:27:59 +00:00
fc43d5c4a0
Initial broker implementation
2019-03-18 15:34:14 +00:00
f4f3b67cdf
NOTICE support
2019-02-09 10:36:03 +00:00
94c66094e6
Cope with libpurple weirdness
2019-02-05 13:37:18 +00:00
21131dc02c
Force membership of channel on login
2019-02-05 13:37:04 +00:00
c5848e4c0b
PASS; better NICK collision avoidance
2019-02-05 13:30:18 +00:00
823754c8b5
Nested reloaders; simple greeter
2019-01-30 16:40:52 +00:00
28d7bd6653
Make reloaders in turn reloadable
2019-01-30 16:40:12 +00:00
8b516b0190
Filesystem driver: add background-activity signal
2019-01-30 16:39:13 +00:00
636055cbea
Scriptlet for starting a client
2019-01-29 23:08:04 +00:00
1e3e978e36
/LIST
2019-01-29 20:47:17 +00:00
7e2284ac11
User count; define/query-count
2019-01-29 20:47:06 +00:00
9c5d9768b4
Pinned channels
2019-01-29 20:46:38 +00:00
ed0307d4e5
Parity with the other version
2019-01-28 13:54:18 +00:00
bd5f2d8a1b
Ported ircd example to imperative-syndicate
2019-01-28 13:50:49 +00:00
60412bd6f2
config driver
2019-01-28 13:47:29 +00:00
832adc94c0
Enable immediate-query
2019-01-28 13:46:54 +00:00
e72e22f250
tcp-connection-peer
2019-01-28 01:14:58 +00:00
23580010bb
Supervision and reloading
2019-01-28 01:14:33 +00:00
636ca9acf7
Note in HOWITWORKS about `opaque-placeholder`.
2018-11-20 13:22:44 +00:00
937bb7a7c4
Identify, expose, and repair bug (using the new `opaque-placeholder`)
2018-11-20 13:20:31 +00:00
6c1e3b033d
Support binary/text websocket payloads
2018-11-15 06:49:31 +00:00
93371894a0
Repair longstanding, subtle bug in both old- and new-syndicate.
...
during/spawn used not to add linkage assertions to its
initial-assertion set. In addition, if a spawned actor died in its
initial boot procedure, its initial assertions would never be visible.
These two problems interlocked to cause a space leak in during/spawn,
where monitoring facets would never be cleaned up.
This change does two things:
- adds linkage assertions to the initial-assertion set in during/spawn
- properly briefly signals initial-assertions even when a new actor
immediately crashes.
Together, these repair the space leak in during/spawn with a crashy
child startup procedure.
2018-11-04 13:43:17 +00:00
45f638ebae
Silly mistake
2018-11-02 15:09:20 +00:00
9571125f3b
More notes
2018-11-02 12:18:44 +00:00
fb195eb9b7
Improve (?) test ergonomics.
2018-11-02 12:15:21 +00:00
7f14dd900c
Detect and repair error in error-handling and -recovery.
2018-11-02 12:15:04 +00:00
bc4fb5ef94
Ensure patched assertions are added before being removed, to avoid glitching
2018-11-02 00:38:40 +00:00
22b4f965b9
Separate out narration
2018-10-24 16:14:54 +01:00
84298d5571
The problem specifies 10 elves
2018-10-24 14:29:10 +01:00
838bde3740
Supply #f as cleanup function in multicast udp-dataspace, instead of void; notes on cleanup functions and visibility-restrictions
2018-10-24 13:21:20 +01:00
6ba3734d9d
Santa Claus Problem, in Syndicate/rkt
2018-10-24 11:40:16 +01:00
4ddcdfb6a9
Present-to-absent check is necessary in assertion removal handler
2018-10-21 13:32:06 +01:00
02181354dd
Tweaks
2018-10-21 01:04:05 +01:00
db563fc309
Minor corrections and tweaks
2018-10-21 00:59:38 +01:00
836bae0e27
First JavaScript steps, based on HOWITWORKS.md
2018-10-21 00:58:40 +01:00
a178ec5ddf
Update to explain the "dummy" outermost wrapper constructor
2018-10-20 21:54:58 +01:00
035eb06a9d
Nope, unsafe-struct-ref is still not measurably faster than the safe variants
2018-10-20 21:47:47 +01:00
a5d147e277
Avoid a few gratuitous conversions during `skeleton-modify!`.
2018-10-20 21:42:38 +01:00
b511e640bd
Minor tweak
2018-10-20 19:52:37 +01:00
cbf5224599
Remove superfluous update-path helper
2018-10-20 19:13:07 +01:00
6e06c4d502
Fix bug exposed by previous commit
2018-10-20 19:09:25 +01:00
8ff6dc5ff8
Eliminate one set of dummy wrappers
2018-10-20 18:27:15 +01:00
0c701809e1
Describe how the efficient index structure works
2018-10-14 21:25:13 +01:00
d497d9e6d1
Split out preserves into its own repository
2018-09-29 17:22:34 +01:00
2c03883b00
SOH-prefixed embedded format considered harmful
2018-09-29 00:07:49 +01:00
268afe8a40
Notes on embedded binary values
2018-09-29 00:04:56 +01:00
348cb3e2a2
Notes on NaNs
2018-09-28 11:48:58 +01:00
b0de674117
Grammar
2018-09-28 11:12:58 +01:00
9c38509973
Typo
2018-09-28 11:12:53 +01:00
4187ccb77d
Clarification
2018-09-28 11:12:35 +01:00
4f65284ef9
More tests and fixes
2018-09-28 11:00:50 +01:00
d27f9cb7cc
Disallow whitespace between a label and its open-parenthesis
2018-09-28 11:00:40 +01:00
d267f65b71
More tests, coverage and fixes
2018-09-27 23:14:43 +01:00
5f9f2175f0
More tests and fixes
2018-09-27 22:13:46 +01:00
47519680d5
Bug fixes to text reader, and more tests
2018-09-27 21:35:03 +01:00
b807b38a44
More TODOs in the text; initial textual reader in Racket
2018-09-27 19:25:28 +01:00
f9497d64c5
Handle a couple of TODOs
2018-09-27 13:34:32 +01:00
6fa0dde8f4
WIP from the early hours of this morning, adding textual syntax
2018-09-27 11:42:55 +01:00
906f8a01b6
Python preserves
2018-09-25 15:53:56 +01:00
b6a3c480b3
Small fix and new question
2018-09-25 15:53:42 +01:00
d0ff78e406
Remove version; improve tests
2018-09-25 15:53:35 +01:00
240cee9d52
Bring preserve.rkt up to spec
2018-09-25 11:49:32 +01:00
780bc6fec2
Python SignedInteger rep needs long as well as int
2018-09-25 10:20:35 +01:00
6e98ba0097
Link to Racket docs for prefab struct labels
2018-09-25 10:08:22 +01:00
0444a2a096
Streamed binaries always use ByteString chunks
2018-09-24 23:15:36 +01:00
a9445b2ca7
Squeak Smalltalk mapping
2018-09-24 19:54:59 +01:00
e25b7dacbc
Improve (?) Erlang mapping
2018-09-24 19:54:52 +01:00
4237200210
Remove page-break override
2018-09-24 19:54:43 +01:00
354776199f
Improve formatting now that section numbers exist
2018-09-24 19:54:23 +01:00
3c7215397d
RFC7159 -> RFC8259
2018-09-24 19:12:43 +01:00
35726436c9
Specially number appendices
2018-09-24 19:12:29 +01:00
e438085e74
Tweaks; python mapping
2018-09-24 18:34:07 +01:00
3b12a290a3
Single-colon pseudoselectors
2018-09-24 18:11:34 +01:00
1275aaded1
Fix the jolly section numbering
2018-09-24 18:11:19 +01:00
96c46c19b4
Split out CSS; attempt to get section numbering (grrr!)
2018-09-24 16:45:39 +01:00
4716b4f4e9
Minor print layout tweaks, and minor content fixes
2018-09-24 16:08:48 +01:00
b656df8632
Tighten
2018-09-24 15:33:19 +01:00
a22ded2f16
Literal small integers
2018-09-24 14:09:26 +01:00
b4d4092b90
Trim and improve
2018-09-24 12:59:22 +01:00
b2eb53e664
Tweak print stylesheet
2018-09-24 10:36:10 +01:00
7f71620136
Print makefile
2018-09-23 22:56:41 +01:00
f2c2e4d39e
Tweaks for Chrome and for print
2018-09-23 22:56:36 +01:00
ff5cb65f5f
Fixes
2018-09-23 22:44:43 +01:00
00a69ae012
Progress
2018-09-23 22:35:00 +01:00
f2f57385ce
Many improvements
2018-09-23 18:14:58 +01:00
2996970cbe
move codec.md together with preserve.md
2018-09-23 14:39:46 +01:00
738fad360a
codec.md draft
2018-09-23 14:39:02 +01:00
c23285781c
preserve.md based on codec.md which I'm about to check in
2018-09-23 14:37:20 +01:00
f6ab8320c5
sqlite driver
2018-08-29 15:03:12 +01:00
939396b9b7
Rename codec --> preserve
2018-08-27 10:24:11 +01:00
542d0a886a
Explore struct inheritance
2018-08-21 10:54:17 +01:00
7dacb5ba6e
Better codec
2018-08-19 22:13:42 +01:00
77b0addcd0
Improve multicast protocol
2018-08-19 17:54:32 +01:00
c125564f5f
Improve example
2018-08-15 10:11:07 +01:00
061765041b
Initial quasi-port of Syndicate/rkt web driver
2018-08-14 17:58:36 +01:00
76674c77b3
Multicast-UDP-based dataspace sketch, from last night
2018-08-14 12:35:56 +01:00
e9457af8c2
Clear no-longer-interesting timers
2018-08-14 12:33:50 +01:00
979d057f3f
TODO
2018-08-13 21:45:44 +01:00
f5efa77d91
term-intersect
2018-08-13 21:32:51 +01:00
86412b273d
UDP and UDP multicast
2018-08-13 20:07:27 +01:00
ea16d24be8
filesystem driver
2018-05-11 09:58:25 +01:00
43f1eab1f3
Port netstack to imperative-syndicate.
...
If you change `racket-bitsyntax` to use `typed/racket/base/no-check`
for its `bitstring.rkt` module, this runs about 15x faster than the
`syndicate` version of the stack. Otherwise, it runs about 3x faster
than the `syndicate` version of the stack.
2018-05-06 15:28:41 +01:00
35726f3831
Notes on tcp2 protocol
2018-05-06 11:24:28 +01:00
96d24fbb4b
New test case
2018-05-06 11:06:36 +01:00
9fc621c829
Add missing assertion for outbound connections.
2018-05-06 11:03:39 +01:00
61cd0613ca
Repair bogus test. I should have caught this earlier!
2018-05-06 10:55:21 +01:00
be1bc6c220
Improve the kinds of tests we can do for expected output slightly.
2018-05-06 10:55:02 +01:00
daba892a07
pattern-expander.rkt
2018-05-04 23:15:09 +01:00
d053edc101
GUI example
2018-05-04 16:57:22 +01:00
46ddb47050
Cosmetic - extra logging (when uncommented)
2018-05-04 16:56:56 +01:00
d3f27488ed
Repair long-standing error in skeleton-walking.
2018-05-04 16:56:45 +01:00
000fa3c008
Extremely limited support for quasiquoting and quoting in patterns
2018-05-04 16:56:23 +01:00
1868b10617
Clock face example
2018-05-04 16:09:12 +01:00
bc4c080641
Quasi-useful debug output
2018-05-04 16:04:42 +01:00
ab64f71766
Restriction-paths: right idea (?), wrong implementation. This time maybe.
...
Instead of having restriction-paths as an adjunct to a change, they're
more propertly a part of each assertion itself. The new `skeleton.rkt`
keeps an optional restriction-path with each assertion, treating it as
distinct from its underlying assertion. The idea of not signalling
changes in assertions that have a restriction-path mismatch stays.
2018-05-04 15:55:53 +01:00
ef4584a60f
Skip apply-patch! make-work if patch is empty
2018-05-03 22:56:40 +01:00
69a8893da3
Add the concept of "restriction-path", which limits assertion visibility.
...
This is used when relaying: because we don't have access to the full
term, but only the projection results, we are inserting various
`(discard)`s. This is the cause of the failure in
`test/core/nesting-confusion.rkt`.
By adding `restriction-path`, we allow the inner dataspace to avoid
showing a reconstructed term to endpoints that might be able to
observe the reconstructed parts.
An alternative implementation approach is to generalize patterns in
the inner relay actor, translating all `(discard)`s into captures,
which would give us all the relevant terms that we need. The way I've
chosen to go (or at least, to try out) allows us to potentially keep
the "efficient" idea of just transmitting pattern-bound values across
some network link connecting dataspaces. The alternative would require
transmission of the full assertions, eliding no irrelevant detail.
2018-05-03 22:09:33 +01:00
8b36c52455
for/bag and for/bag/count
2018-05-03 22:09:13 +01:00
ee0033b6be
Make test-implementation language expose all of `racket` rather than just `racket/base`, for convenience.
2018-05-03 22:08:52 +01:00
f01b17dea4
Reuse apply-patch for initial-assertions too
2018-05-03 20:09:34 +01:00
00d50e6700
Factor out assertion structure definitions
2018-05-03 16:14:30 +01:00
fecd2f9787
Notes on install and test
2018-05-03 15:21:33 +01:00
47a113026c
README.md
2018-05-03 15:06:53 +01:00
fa3505e3a6
Add second, non-dataspace, test case to illustrate correct behaviour
2018-05-02 23:36:03 +01:00
852eff8d0b
New (failing) test case re: the approach to relaying
2018-05-02 18:23:02 +01:00
bc5b28c403
Correct printing of test procedures
2018-05-02 18:20:33 +01:00
e4a1a6e075
Add `bag-empty?`
2018-05-02 18:20:24 +01:00
17cee51342
Ensure (??) that relays' assertions can't outlive them
2018-05-02 18:10:52 +01:00
dd123f4a27
Notes on eventual improvements to dataspace nesting
2018-05-02 13:19:18 +01:00
74cb70bdf3
Actually-useful instantaneous frame rate measurement
2018-05-01 22:40:12 +01:00
eb208f3a10
Handy logging for diagnosing problems
2018-05-01 22:39:50 +01:00
367b39ce39
Eliminate a (mostly harmless) source of #f current-facet.
2018-05-01 22:39:26 +01:00
669122766d
Allow multiple "root" facets, by creating a dummy root
2018-05-01 22:39:02 +01:00
c85f05ced1
A few names for unnamed actors make it clear which actor is faulting
2018-05-01 21:19:19 +01:00
bbcb8a3b80
Adapt platformer to imperative-syndicate. Mostly working!
...
It's also nicely fast compared to the old-Syndicate version :-)
There are still some problems with parameters in cross-ds relaying;
the symptom is #f for (current-facet) at some point, leading to some
sprites that don't get retracted (!).
2018-05-01 21:15:22 +01:00
869a495392
First sketch of `quit-dataspace!`.
2018-05-01 21:12:54 +01:00
896cfe2498
Multi-assert!/retract! for adhoc assertions
2018-05-01 20:58:26 +01:00
e43054a6bd
More thoughtless uncommenting of query-* forms
2018-05-01 20:58:02 +01:00
9c22868bd2
Correct error from earlier thoughtless uncommenting
2018-05-01 20:57:42 +01:00
e4e0b5f9d4
Support vector in patterns, like we already support list
2018-05-01 20:57:22 +01:00
2b2d12075a
gl-2d driver and examples
2018-05-01 17:39:17 +01:00
38c6bfd20e
Drain multiple external ground events at once. (Good or bad idea?)
2018-05-01 17:34:51 +01:00
65b2c94161
Easy access to profiling
2018-05-01 17:34:34 +01:00
32d217f3e0
Care less about exact set ordering in test case
2018-05-01 17:32:03 +01:00
475e6527f3
Uncomment query-value and friends. May need tweaks.
2018-04-30 22:48:50 +01:00
ec0933ffdd
Clamp `adhoc-retract!` to not drop below zero.
2018-04-30 22:48:27 +01:00
2575b71db4
In case there are no activation-forms, add a gratuitous `(void)` at the end.
2018-04-30 22:47:53 +01:00
f3ca2b6a15
Repair incorrect interleaving of actions and scripts
2018-04-30 22:47:25 +01:00
3c4076287e
Switch another in-hash to hash-for-each
2018-04-30 14:46:43 +01:00
671e9742c7
Speedup from avoiding uselessly enqueueing an empty work queue
2018-04-30 11:19:03 +01:00
0e37037b6e
Eliminate another parameter: now only `current-facet` remains
2018-04-30 11:18:49 +01:00
14ee4f70ef
Avoid manipulating `in-script?` all the time
2018-04-30 10:44:35 +01:00
c2cb624e42
Another small speed bump from using hash-for-each over in-hash-keys
2018-04-30 10:04:33 +01:00
00121d9710
Avoid sets on the hot path (use hashes instead); good speed improvement
2018-04-30 09:32:08 +01:00
eb3a93e32b
Add in-* annotations to for loops
2018-04-30 09:14:00 +01:00
a6811f2ba5
Add timer/timestate support; this prompted a change to endpoint registration.
...
The `add-endpoint!` call is changed in two ways:
- the old `assertion-fn` has become `update-fn`, yielding both
an assertion *and* an optional handler, because if the handler
depends on a field which changes, previously the handler wasn't
being updated
- a new parameter, `dynamic?`, can be set to #f (it's usually #t)
to ensure that the assertion and skeleton-interest are calculated
only once ever, and are not connected to the dataflow machinery.
The first change makes it possible for the `(later-than (deadline))`
pattern, where `deadline` is a field, to work; the second change makes
`during` and `during/spawn` work correctly in the face of field
updates.
2018-04-29 22:27:55 +01:00
91596b066f
External-event and TCP drivers; chat server and stdin echo programs
2018-04-29 18:43:39 +01:00
f45323a097
Extract raw dataspace test to separate file
2018-04-29 16:08:52 +01:00
2652147141
Commented-out debug-printing in relay.rkt
2018-04-29 16:08:35 +01:00
8ba3779eda
Add `test` make target
2018-04-29 16:08:20 +01:00
5b435ee8d8
bag-key-count
2018-04-29 16:08:01 +01:00
b3d6b74dab
Repair scheduling of inner dataspaces, thus allowing e.g. double nesting
2018-04-29 16:07:49 +01:00
e69473a115
Makefile
2018-04-29 14:54:28 +01:00
1fe134d920
Tweak .gitignore
2018-04-29 14:54:22 +01:00
dd2cddb6a7
Nested dataspaces
2018-04-29 14:54:14 +01:00
204197c3eb
First stab at #lang infrastructure
2018-04-29 12:22:12 +01:00
f40e7c15cc
Refactor tests
2018-04-29 11:55:32 +01:00
afad4cd0be
Repair rotten pattern tests
2018-04-29 11:30:26 +01:00
3a857db4b9
Box-and-client speed test
2018-04-27 23:59:04 +01:00
290170f450
Simple test harness
2018-04-27 23:53:31 +01:00
d6c0241c57
Cosmetic notes
2018-04-27 18:23:34 +01:00
a4abea9d87
Remove current-dataspace parameter
2018-04-27 18:05:33 +01:00
0cbb5b061c
Cosmetic notes
2018-04-27 17:36:43 +01:00
edb8e719f7
More robust approach to cleanup of assertions on actor termination.
...
We now explicitly track *committed* assertions of each actor in a new
field, `actor-cleanup-changes`. Each time a patch action is
*performed*, `actor-cleanup-changes` is updated. When an actor quits,
it enqueues a special new kind of action, a `quit` action.
When a `quit` action is performed, any remaining contents of
`actor-cleanup-changes` are processed in order to fully remove any
leftover assertions. (Leftover assertions will only arise in
exceptional cases: when some stop-script or facet boot-script raises
an uncaught exception.)
As part of this commit, we undo the effect of commit 8624047
2018-04-27 17:30:26 +01:00
65a221ae68
New failing test case. This, plus recent commits, calls for a
...
different strategy for tracking the retractions necessary at actor
termination.
2018-04-27 10:55:21 +01:00
0d5b7e9c02
Cosmetic
2018-04-27 10:54:58 +01:00
a247d8fd2b
Switch adhoc-assertions to a bag (from a set).
...
This repairs a bug regarding crashes in a new actor's boot-proc.
Previously, if boot-proc raised an exception, the initial assertions
would stick around forever. By changing adhoc-assertions to a bag
rather than a set, and putting the initial assertions in the bag, we
put them somewhere they are guaranteed to be processed during actor
termination, even when an exception is signalled during boot.
This is an API change wrt the previous Syndicate implementation:
assert!/retract! now have bag semantics, not set semantics. We can add
set-semantics APIs if we end up needing them, of course, layered on
top of the bag implementation.
2018-04-27 09:59:03 +01:00
ff4799b282
Immutable bags
2018-04-27 09:55:19 +01:00
8624047ecd
The failing test now passes. The reasons for this are subtle:
...
The patch here removes a terminated facet from its parent's
`facet-children` set only in a script, and only after all other
scripts enqueued as part of facet termination have executed without an
uncaught exception.
This means that, if (say) a stop script raises an uncaught exception,
it might have happened after some *but not all* scripts resulting from
calls to `retract-facet-assertions-and-subscriptions!` have already
executed. So some endpoints' assertions and subscriptions will have
been removed.
When the uncaught exception is caught by the handler in
`with-current-facet`, a call to `abandon-queued-work!` is made, which
discards queued scripts, including the remaining assertion-cleanup
scripts as well as the scripts for removing dead facets from their
parents' `facet-children` sets. It also (crucially) discards queued
patch actions, including those resulting from already-executed
assertion-cleanup scripts.
At this point, we have a facet tree with some dead facets still in it,
and no queued outbound patches. The assertions for the still-present
dead facets are still logically asserted.
Then, a call to `terminate-actor!` happens, which traverses the whole
tree enqueueing assertion-cleanup scripts. No user code is enqueued,
so (in principle) no exceptions can be signalled.
Once these `terminate-actor!`-enqueued scripts execute, a pending
patch exists that will remove all remaining endpoint assertions.
The remaining sticky point is the calls to `dataspace-unsubscribe!`.
Happily, these are idempotent because of the implementation in
`skeleton.rkt`.
Prior to this patch, terminating facets were removed early from their
parents' `facet-children` sets, meaning there was no way to find them
again to clean up if a failure occurred during a stop script.
Ideally, it'd be easy to see that the code is correct in this respect.
We're not there yet.
2018-04-25 20:43:25 +01:00
928e48113f
Failing test: an exception in a stop script messes up cleanup
2018-04-25 20:13:54 +01:00
ce6cd58930
Run stop-scripts in one scheduled execution
2018-04-25 19:54:58 +01:00
dfe8a3e96e
Be specific in for-loops
2018-04-25 19:47:18 +01:00
c06d5ba3ce
Guard against double-adhoc-assertion
2018-04-25 19:46:08 +01:00
9726376c82
Remove redundant `facet-live?` test
2018-04-25 19:40:53 +01:00
bc9b931312
Omit "name" to ctor for dataspace struct - name is/will be a property of the downward relay connection
2018-04-25 19:30:45 +01:00
e6d6f67021
during/spawn
2018-04-22 21:07:35 +01:00
1523ef17cf
Correct typo
2018-04-22 21:06:18 +01:00
efa7f517c4
Repair: it is incorrect to invoke retraction-handlers when retracting an interest
2018-04-22 21:04:22 +01:00
f58c2fb8dc
Repair error: stop-facet scripts must run in parent facet's context!
2018-04-22 21:03:21 +01:00
a3dfb2c5aa
Notice attempts to install more than one root facet in an actor
2018-04-22 21:03:01 +01:00
4be935a160
Introduce action-groups
2018-04-22 21:02:40 +01:00
634b114afc
Improve debug-printing of facets
2018-04-22 21:00:35 +01:00
b50d18cd46
Remove unused dataspace-actors field
2018-04-22 20:59:57 +01:00
55f4b55784
Implement during and ad-hoc assertions
2018-04-19 17:55:52 +01:00
5377a486c9
Better Isolated Turn Principle
2018-04-11 12:28:09 +01:00
533d6f9a24
Improve syntax location tracking
2018-04-09 10:29:14 +01:00
3452f1fe6e
Script suspend and resume
2018-04-09 10:23:22 +01:00
bd9dcb61ca
Simple messages-per-second test (~195kHz at present)
2018-04-08 12:06:37 +01:00
eb24d566e9
Progress on syntax veneer
2018-04-08 11:44:32 +01:00
e84f2e2e39
Eliminate (interesting uses of) facet IDs by storing references directly
2018-04-08 08:52:37 +01:00
fe6b895d8a
More structured FID representation.
2018-04-08 07:58:17 +01:00
3c70496688
Steps toward API usable by syntax layer; beginning of syntax layer, adapted from previous implementation
2018-04-08 07:39:39 +01:00
7f65d9e452
Fine-tune skeleton-accumulator representations for persistency/reuse and for potential speed benefit
2018-04-08 07:01:39 +01:00
c122add136
Correct typo
2018-04-07 08:42:38 +01:00
5d6e0f3904
printf to log-info; clean up demo output
2018-04-06 12:07:09 +01:00
4c4afc6b6e
Eliminate a few useless parameters
2018-04-06 11:58:49 +01:00
2e67feee6d
POC imperative dataspace implementation, with set-box example
2018-04-06 11:37:59 +01:00
25997cc778
Switch to skeleton-stx with struct-type rather than predicate
2018-03-27 22:21:49 +13:00
1d67cbe65e
Rearrange
2018-03-27 22:21:49 +13:00
883fcb9b75
Start translation of pattern syntax to various structures
2018-03-27 22:21:49 +13:00
86930702e0
Initial sketch of new routing-table idea
2018-03-27 22:21:49 +13:00
Tony Garnock-Jones