]>
gitweb.factorcode.org Git - factor.git/log
Björn Lindqvist [Sun, 19 Oct 2014 12:17:54 +0000 (14:17 +0200)]
compiler.codegen.gc-maps: update docs so that they match how factors gc-maps currently work
Björn Lindqvist [Sun, 19 Oct 2014 12:05:06 +0000 (14:05 +0200)]
Docs: fixing smaller help-lint nitpicks
Björn Lindqvist [Sun, 19 Oct 2014 11:45:42 +0000 (13:45 +0200)]
compiler.cfg.*: more stub docs for cfg construction words
Björn Lindqvist [Sun, 19 Oct 2014 11:41:45 +0000 (13:41 +0200)]
compiler.codegen.*: help for resolve-label misplaced -- move it
Björn Lindqvist [Sun, 19 Oct 2014 11:40:41 +0000 (13:40 +0200)]
cpu.*: docs and tests for words in the assembler
Björn Lindqvist [Tue, 14 Oct 2014 15:37:46 +0000 (17:37 +0200)]
compiler,cpu: more docs
Björn Lindqvist [Tue, 14 Oct 2014 11:41:14 +0000 (13:41 +0200)]
compiler.cfg.*: more docs, tests and a small refactoring of the
admissible-registers word
Björn Lindqvist [Mon, 13 Oct 2014 06:33:27 +0000 (08:33 +0200)]
stack-checker,compiler: docs for stack-checker and compiler words
Björn Lindqvist [Mon, 25 Aug 2014 13:26:35 +0000 (15:26 +0200)]
cpu.architecture: %save-context example
Björn Lindqvist [Wed, 20 Aug 2014 19:15:00 +0000 (21:15 +0200)]
compiler.codegen.gc-maps: initial docs
Björn Lindqvist [Wed, 20 Aug 2014 19:14:28 +0000 (21:14 +0200)]
compiler.*: doc improvements
Doug Coleman [Thu, 16 Oct 2014 23:40:49 +0000 (16:40 -0700)]
io.launcher.windows: Make unit tests show that process fails rather than
reporting bogus error. Fixes #1152 (it wasn't really a bug, just bad
tests).
John Benediktsson [Thu, 16 Oct 2014 01:22:57 +0000 (18:22 -0700)]
magic.ffi: look for "magic" (works better on linux).
John Benediktsson [Wed, 15 Oct 2014 20:46:47 +0000 (13:46 -0700)]
magic: some minor fixes.
John Benediktsson [Wed, 15 Oct 2014 03:50:38 +0000 (20:50 -0700)]
magic: bindings for libmagic.
Björn Lindqvist [Fri, 10 Oct 2014 17:45:13 +0000 (19:45 +0200)]
fuel.eval: wrap fuel-pprint in a without-limits so fuel output isn't truncated
Doug Coleman [Fri, 10 Oct 2014 18:54:35 +0000 (11:54 -0700)]
io.streams.peek: Add file position indicator api calls.
Björn Lindqvist [Fri, 10 Oct 2014 09:37:54 +0000 (11:37 +0200)]
images.loader.tests: restructure the tests because the gtk linux backend doesn't handle some formats that the windows backend does
Björn Lindqvist [Fri, 10 Oct 2014 09:14:57 +0000 (11:14 +0200)]
images.loader.gtk: fix fall-through in case problem in image>GdkPixbuf
Doug Coleman [Fri, 10 Oct 2014 01:50:38 +0000 (18:50 -0700)]
prettyprint: @mrjbq7's funny word for . that doesn't limit the length of things.
Doug Coleman [Thu, 9 Oct 2014 22:40:41 +0000 (15:40 -0700)]
imap-tests: Reset test folder.
Björn Lindqvist [Wed, 8 Oct 2014 14:38:51 +0000 (16:38 +0200)]
images.loader.gdiplus: implementation of the image>stream word for GDI+
for Windows
Björn Lindqvist [Tue, 7 Oct 2014 15:33:30 +0000 (17:33 +0200)]
images: move rowstride to the images vocab because it will be used by
gdi too
Doug Coleman [Thu, 9 Oct 2014 19:22:04 +0000 (12:22 -0700)]
tools.deploy-tests: inc size for linux32
Doug Coleman [Thu, 9 Oct 2014 08:01:27 +0000 (01:01 -0700)]
stack-checker.alien: We just want to track callbacks to see if we leak them, but not actually dispose of them yet. Fixes #1151.
John Benediktsson [Wed, 8 Oct 2014 22:36:08 +0000 (15:36 -0700)]
calendar.format: change rfc3339 to always print microseconds.
Doug Coleman [Wed, 8 Oct 2014 18:41:34 +0000 (11:41 -0700)]
tools.deploy: Bump the build size. This is a comparatively sizeable bump
compared to usual, but I really want clean builds.
Doug Coleman [Wed, 8 Oct 2014 18:33:40 +0000 (11:33 -0700)]
smtp: Make the test pass on Windows.
Doug Coleman [Wed, 8 Oct 2014 17:12:16 +0000 (10:12 -0700)]
db.postgresql: This is not actually the case on my Mac.
Doug Coleman [Wed, 8 Oct 2014 17:11:04 +0000 (10:11 -0700)]
db.tester: postgresql template1 database name should be "template1"
Doug Coleman [Wed, 8 Oct 2014 06:56:12 +0000 (23:56 -0700)]
tools.profiler.sampling: Kick it down a notch. Makes win32 tests finish!
Björn Lindqvist [Tue, 7 Oct 2014 22:03:36 +0000 (00:03 +0200)]
imap: fix parse-store-mail parsing because it breaks on gmail
gmail responses include lines with the order of UID and FLAGS reversed
to indicate the previous flags for a mail. Just ignore those lines
because they aren't useful and non-standard.
Doug Coleman [Tue, 7 Oct 2014 20:51:13 +0000 (13:51 -0700)]
io.launcher.windows.test/env.factor: Fix script for new . behavior.
Doug Coleman [Tue, 7 Oct 2014 20:13:58 +0000 (13:13 -0700)]
tools.gc-decode: Just make sure that the gc-decode tests work. Any answer
is ok.
Doug Coleman [Sat, 4 Oct 2014 05:56:21 +0000 (22:56 -0700)]
compression.snappy: Add support for snappy compression by Google. Could possibly be optimized?
Björn Lindqvist [Fri, 3 Oct 2014 11:40:39 +0000 (13:40 +0200)]
graphviz.render.windows: need to return with the whole path in case
graphivz directory is not in PATH
Björn Lindqvist [Fri, 3 Oct 2014 10:53:16 +0000 (12:53 +0200)]
webapps.mason.backend.tests: change db filename to mason-test.db to
avoid interference from other tests
Björn Lindqvist [Wed, 24 Sep 2014 20:14:46 +0000 (22:14 +0200)]
lists.lazy.tests: plugs file descriptor leak
Björn Lindqvist [Tue, 23 Sep 2014 22:47:26 +0000 (00:47 +0200)]
db.tester: with-disposal needed to release the db connections in the pool
Björn Lindqvist [Tue, 23 Sep 2014 22:46:47 +0000 (00:46 +0200)]
smtp.tests: send-mail should be called to get the mock-smtp-server to terminate
Björn Lindqvist [Tue, 23 Sep 2014 22:45:46 +0000 (00:45 +0200)]
openssl.libcrypto.tests: BIO_free should be called to release connections
Björn Lindqvist [Thu, 2 Oct 2014 15:19:05 +0000 (17:19 +0200)]
python: docs for python-error
Björn Lindqvist [Thu, 2 Oct 2014 13:15:50 +0000 (15:15 +0200)]
python: tests and plug a reference leak in the exception throwing code
Björn Lindqvist [Fri, 26 Sep 2014 18:04:33 +0000 (20:04 +0200)]
python.throwing: new module to support capturing python tracebacks
Doug Coleman [Tue, 30 Sep 2014 19:42:25 +0000 (12:42 -0700)]
build-support: test failed on mac somehow, add a timeout so it fails faster.
John Benediktsson [Tue, 30 Sep 2014 15:23:32 +0000 (08:23 -0700)]
prettyprint.backend: use , not % in unparse-ch.
Björn Lindqvist [Tue, 30 Sep 2014 12:53:22 +0000 (14:53 +0200)]
db.postgresql: ensure that 2 line error messages are handled
Björn Lindqvist [Tue, 30 Sep 2014 12:51:07 +0000 (14:51 +0200)]
db.tester: the postgresql-db variable shouldn't be mutated
Björn Lindqvist [Sun, 17 Aug 2014 20:07:56 +0000 (22:07 +0200)]
db.postgresql: configurable test db setting
adds a \ postgresql-db global (analoguous to the \ imap-settings one)
which holds settings to a postgres db so that the tests have something
to test against.
Björn Lindqvist [Mon, 29 Sep 2014 12:26:08 +0000 (14:26 +0200)]
compiler.tests.alien: wrap all callback tests in special
unit-test-with-destructor blocks, that way they dont leak memory in the
callback heap
Björn Lindqvist [Sun, 21 Sep 2014 22:02:57 +0000 (00:02 +0200)]
alien,stack-checker.alien: move callback-destructor to the alien vocab
It can't be placed in stack-checker.alien because that vocab isn't
included in deployed images which lead to weird errors when the
destructors trigger.
Björn Lindqvist [Wed, 17 Sep 2014 10:26:00 +0000 (12:26 +0200)]
stack-checker.alien: use free-callback as a destructor when creating
callbacks
+ tests proving it works. now in case you create temporary callbacks you
can enclose them in with-destructors and it will just work(tm).
Björn Lindqvist [Tue, 16 Sep 2014 15:14:39 +0000 (17:14 +0200)]
alien: a test and docs for free-callback
Björn Lindqvist [Mon, 15 Sep 2014 16:20:22 +0000 (18:20 +0200)]
VM: new primitive free-callback which is able to free a callback
previously allocated using <callback>
Björn Lindqvist [Mon, 15 Sep 2014 10:35:55 +0000 (12:35 +0200)]
stack-checker.known-words: stack effect for <callback> is switched, it
should be ( word integer -- alien )
Björn Lindqvist [Wed, 10 Sep 2014 15:01:48 +0000 (17:01 +0200)]
VM: new primitive (callback-room) for querying the VM about memory usage
The word works exactly like (code-room) except it looks at the memory
usage in the callback heap instead of the code heap.
Björn Lindqvist [Tue, 9 Sep 2014 15:56:56 +0000 (17:56 +0200)]
VM: use a free_list_allocator for the callbacks, that way they can
individually be freed
Björn Lindqvist [Tue, 9 Sep 2014 12:45:01 +0000 (14:45 +0200)]
VM: "formalize" the callback_heaps object allocation using a
bump_allocator<code_block>
The advantage of doing it this way is that you can reuse the existing
each_object method to iterate all callbacks
Björn Lindqvist [Tue, 9 Sep 2014 10:44:01 +0000 (12:44 +0200)]
VM: try and recover with a kernel error from the callback heap running out
Björn Lindqvist [Mon, 22 Sep 2014 21:47:48 +0000 (23:47 +0200)]
furnace.db: db-persistence needs a dispose word that delegates to its pool slot
tests in http.tests refactored using "with-words" so that they always
dispose any db-persistence instances they create.
John Benediktsson [Mon, 29 Sep 2014 03:31:39 +0000 (20:31 -0700)]
mason.report: dos2unix.
Björn Lindqvist [Fri, 26 Sep 2014 15:17:14 +0000 (17:17 +0200)]
benchmark.reverse-complement.tests,mason.report.tests: fix for the test
cases so they write to temp files instead of in resource:
John Benediktsson [Mon, 29 Sep 2014 00:21:04 +0000 (17:21 -0700)]
lists: cleanup usings.
Iskander Sitdikov [Sun, 28 Sep 2014 22:20:50 +0000 (02:20 +0400)]
Refactor >list to be a generic word
Björn Lindqvist [Sun, 28 Sep 2014 20:53:56 +0000 (22:53 +0200)]
tools.deploy.backend: with-variables need to be on the whole block
Björn Lindqvist [Sun, 28 Sep 2014 19:16:39 +0000 (21:16 +0200)]
tools.deploy.backend: rearranging the order of parameters in the deploy process, -run needs to be last
Doug Coleman [Fri, 26 Sep 2014 20:33:04 +0000 (13:33 -0700)]
math.statistics: Get smart about the geometric mean. Calculate it in logspace so we can use addition of small logarithms and multiplication instead of the nth root of a giant product.
John Benediktsson [Fri, 26 Sep 2014 04:30:36 +0000 (21:30 -0700)]
help.html: simplify (and separate) index generation.
John Benediktsson [Thu, 25 Sep 2014 19:37:46 +0000 (12:37 -0700)]
help.html: make docs use same completions algorithm as listener.
John Benediktsson [Thu, 25 Sep 2014 19:02:23 +0000 (12:02 -0700)]
command-line: make script and -run= param behave similarly.
Björn Lindqvist [Thu, 25 Sep 2014 15:06:01 +0000 (17:06 +0200)]
csv: need to handle \r because windows lines ends with \r\n
Doug Coleman [Thu, 25 Sep 2014 00:20:57 +0000 (17:20 -0700)]
tools.gc-decode: linux32 is a slow vm and i'm committing from mac and things passed on mac last time. ugh.
Doug Coleman [Thu, 25 Sep 2014 00:15:18 +0000 (17:15 -0700)]
tools.gc-decode: typo..
Doug Coleman [Thu, 25 Sep 2014 00:08:44 +0000 (17:08 -0700)]
tools.deploy: bump sizes for linux32. so not hardcore.
Doug Coleman [Thu, 25 Sep 2014 00:07:11 +0000 (17:07 -0700)]
tools.gc-decode: Fix test for linux32.
Björn Lindqvist [Tue, 23 Sep 2014 15:05:21 +0000 (17:05 +0200)]
io.ports.tests: better to use with-disposal than calling the dispose word explicitly
Björn Lindqvist [Tue, 23 Sep 2014 15:04:14 +0000 (17:04 +0200)]
io.streams.limited.tests: use with-stream in the tests to ensure that the pipes are disposed
John Benediktsson [Wed, 24 Sep 2014 14:42:45 +0000 (07:42 -0700)]
Revert "calendar.format: change to use formatting instead of math.parser.private."
This reverts commit
3cfac7916a4d3708f231875ade0cd5e586ec43e6 .
Björn Lindqvist [Mon, 22 Sep 2014 18:47:25 +0000 (20:47 +0200)]
imap.tests: created connections needs to be disposed
Björn Lindqvist [Mon, 22 Sep 2014 18:32:26 +0000 (20:32 +0200)]
db.postgresql.lib: if PQsetdbLogin fails, PQfinish must be called
Björn Lindqvist [Wed, 17 Sep 2014 17:46:05 +0000 (19:46 +0200)]
VM: GC_BIGNUM must be called in the same scope as the variable declaration
Doug Coleman [Fri, 12 Sep 2014 19:08:05 +0000 (12:08 -0700)]
tools.depoy: Bump test size for mac32.
Doug Coleman [Fri, 12 Sep 2014 19:07:26 +0000 (12:07 -0700)]
tools.gc-decode: Fix unit tests for 32bit.
Iskander Sitdikov [Wed, 10 Sep 2014 09:07:32 +0000 (13:07 +0400)]
Fix the Atom feed generator
Doug Coleman [Tue, 9 Sep 2014 00:48:46 +0000 (17:48 -0700)]
tools.gc-decode: Fix unit test for compiler fix.
Doug Coleman [Tue, 9 Sep 2014 00:48:22 +0000 (17:48 -0700)]
compiler.cfg.stacks.vacant: Fix docs for word rename.
Björn Lindqvist [Mon, 1 Sep 2014 14:59:47 +0000 (16:59 +0200)]
VM: dont scrub and check when visiting a callstack object
callstack objects doesn't have stacks so it doesn't make sense to scrub
and check. also added the methods check_stack and scrub_stack to deal
with code duplication
Björn Lindqvist [Sun, 31 Aug 2014 04:14:20 +0000 (06:14 +0200)]
compiler.cfg.stacks.vacant: ##call instructions clear the set of overinitialized locations
Björn Lindqvist [Sat, 30 Aug 2014 04:52:53 +0000 (06:52 +0200)]
VM: silly error: should be index * check_r_count, not index + check_r_count
Björn Lindqvist [Thu, 28 Aug 2014 15:32:36 +0000 (17:32 +0200)]
VM: move the stack scrubbing logic to call_frame_slot_visitor to avoid visiting the callstack twice
primitive_minor_gc() iterates the stack twice, first to scrub stack
locations, then to trace overinitialized ones and gc roots. By running
visit_callstack() before visit_stack_elements() you only need to do it
once.
Björn Lindqvist [Wed, 27 Aug 2014 02:22:20 +0000 (04:22 +0200)]
VM: use the check_d and check_r to trace the overinitialized stack locations
Björn Lindqvist [Tue, 26 Aug 2014 18:53:58 +0000 (20:53 +0200)]
tools.gc-decode: update vocab to match gc-map changes
Björn Lindqvist [Tue, 26 Aug 2014 18:40:01 +0000 (20:40 +0200)]
compiler.codegen.gc-maps: the gc-map-needed? word can be called with f
Björn Lindqvist [Tue, 26 Aug 2014 02:44:03 +0000 (04:44 +0200)]
compiler.codegen.gc-maps: check-d>> and check-r>> now used in the code generator
the approach looks sound and now the gc can be fixed to take advantage
of the extra info.
Björn Lindqvist [Tue, 26 Aug 2014 02:28:52 +0000 (04:28 +0200)]
compiler.cfg.stacks.vacant: invert meaning of 0/1 for overinit:ed locs to make some logic simpler
Björn Lindqvist [Mon, 25 Aug 2014 20:33:54 +0000 (22:33 +0200)]
compiler.cfg.instructions: added check-d and check-r slots to gc-map
the data is output from compiler.cfg.stacks.vacant in a reasonable
format but not yet used for code generation.
Björn Lindqvist [Mon, 25 Aug 2014 16:55:27 +0000 (18:55 +0200)]
compiler.cfg.stacks.vacant: more docs and removed usage of byte-arrays
Björn Lindqvist [Thu, 14 Aug 2014 20:47:06 +0000 (22:47 +0200)]
compiler.cfg.stacks.vacant: sub docs seeded with a stolen comment from uninitialized.factor
Björn Lindqvist [Wed, 13 Aug 2014 22:11:49 +0000 (00:11 +0200)]
compiler.cfg.stacks.vacant: fix usings and remove commented out code
Björn Lindqvist [Wed, 13 Aug 2014 11:06:34 +0000 (13:06 +0200)]
compiler.cfg.stacks.vacant: compiler pass replacing uninitialized for more accurate stack maps
This pass uses a better algorithm for keeping track of all stack
manipulations which means that the stack maps it assigns are much more
precise.