Browse source

Changelog

Query
From:
To:
Branch:
Path:
Message:
BugID:
User: +
Format:
: 3 checkins (+130/-135) by 2 people : 4 checkins (+75/-19) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 2 checkins (+7/-1) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 1 checkin (+25/-16) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 3 checkins (+15/-9) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 14 checkins (+246/-54) by 3 people : 35 checkins (+247/-203) by 2 people : 16 checkins (+387/-218) by 2 people : 13 checkins (+50/-34) by 2 people : 8 checkins (+56/-17) by 2 people : 3 checkins (+7/-1) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 3 checkins (+42/-14) by 2 people : 5 checkins (+418/-23) by 3 people : 2 checkins (+57/-42) by 2 people : 15 checkins (+1500/-1314) by 2 people : 2 checkins (+33/-28) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 1 checkin (+191/-0) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 3 checkins (+9/-4) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 3 checkins (+23/-16) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 6 checkins (+88/-0) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 21 checkins (+534/-304) by 2 people : 24 checkins (+1124/-846) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 1 checkin (+7/-6) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 8 checkins (+76/-64) by 2 people : 125 checkins (+1214/-1097) by 2 people : 1 checkin (+6/-4) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 3 checkins (+341/-4) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 58 checkins (+389/-292) by 2 people : 180 checkins (+2218/-1999) by 4 people : 20 checkins (+188/-10) by 2 people : 514 checkins (+772/-33765) by 4 people : 489 checkins (+142/-139) by 2 people : 9 checkins (+101/-45) by 2 people : 6 checkins (+460/-58) by 3 people : 2 checkins (+14/-12) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 1 checkin (+8/-2) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 1 checkin (+64/-10) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 23 checkins (+84/-53) by 4 people : 51 checkins (+1006/-631) by 3 people : 9 checkins (+44/-13) by 2 people Lines added/removed recently

Yesterday; Wednesday 02 September 2015

2015-09-02 22:16:22 (6 hours ago) by Martin Nilsson <nilsson@fastmail.com>

Disable extended master secret extension by default. It isn't standardized yet, and currently doesn't interopt with Chrome.

2015-09-02 22:08:06 (6 hours ago) by Martin Nilsson <nilsson@fastmail.com>

Truncated HMAC may be a security issues, and isn't really supported by anyone else. Disable it by default.

2015-09-02 21:58:53 (6 hours ago) by Martin Nilsson <nilsson@fastmail.com>

Documentation updates.

2015-09-02 17:44:22 (11 hours ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Threads: this_thread() needs to survive cleanup on exit.

There apparently are some lfun::destroy() that call this_thread()
at a time when Pike_interpreter.thread_state->thread_obj is NULL.

2015-09-02 17:40:42 (11 hours ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Threads: this_thread() needs to survive cleanup on exit.

There apparently are some lfun::destroy() that call this_thread()
at a time when Pike_interpreter.thread_state->thread_obj is NULL.

2015-09-02 01:33:33 (1 day ago) by Martin Nilsson <nilsson@fastmail.com>

Better types (no more object).

Tuesday 01 September 2015

2015-09-01 18:29:35 (1 day ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Standards.JSON]: UNDEFINED is encoded as null.

FIXME: This change is likely to have compatibility issues.
Add a compat module!

2015-09-01 18:24:34 (1 day ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [ADT.TreeScheduler]: Added some basic tests.

2015-09-01 18:21:32 (1 day ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.https: Support setting the PORT via -D.

Feature backported from Pike 8.0.

2015-09-01 17:41:08 (1 day ago) by Per Hedbor <ph@opera.com>

Optimized store_prog_string and store_constant somewhat.

They are now O(n) instead of O(n^2) for common cases.

For strings, when the number of references to the string is 1, and for
constants when equal is not true (using == instead of equal) and
refs==1 and the whatever is not an object but has references.
Optimized store_prog_string and store_constant somewhat.

They are now O(n) instead of O(n^2) for common cases.

For strings, when the number of references to the string is 1, and for
constants when equal is not true (using == instead of equal) and
refs==1 and the whatever is not an object but has references.

2015-09-01 14:52:40 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Updated to tzdata2015f.
Updated to tzdata2015e.

2015-09-01 14:50:34 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Updated to tzdata2015d.

2015-09-01 14:47:13 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Sql.odbc: Survive old Odbc module.

The Pike patches for Roxen typically don't patch the pike binary, so
make sure that the module compiles even without Odbc.odbc.typed_result.

2015-09-01 13:53:57 (2 days ago) by Per Hedbor <ph@opera.com>

Merge branch '8.1' into per/substrings
[substrings] Fixed sscanf speed

Monday 31 August 2015

2015-08-31 17:51:34 (2 days ago) by Martin Nilsson <nilsson@fastmail.com>

We don't need to test for __builtin_expect twice.

2015-08-31 17:46:02 (2 days ago) by Martin Nilsson <nilsson@fastmail.com>

Clean up some seeks.
Fixed constant leak.

2015-08-31 16:56:39 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation: Added some more RFC markup.

2015-08-31 14:36:57 (3 days ago) by Martin Nilsson <nilsson@fastmail.com>

Without PIKE_DEBUG rb_fatal is just Pike_fatal and takes the error message as first argument.

Conflicts:
src/rbtree.c

2015-08-31 14:33:59 (3 days ago) by Martin Nilsson <nilsson@fastmail.com>

Without PIKE_DEBUG rb_fatal is just Pike_fatal and takes the error message as first argument.

2015-08-31 12:17:37 (3 days ago) by Michał Gawron <mgawron@opera.com>

MT-1208: Fixes in pike8.

* Prevent adding binary ints to output stream when header values are ints.
* Don't try to call tell() on a parameters mapping. Seek in m->file instead.
MT-1211: Follow-up fix after MT-1208 (' ' interpreted as "32").

2015-08-31 12:17:37 (3 days ago) by Per Hedbor <ph@opera.com>

Encode UNDEFINED as null

It makes more sense than 0

2015-08-31 12:15:27 (3 days ago) by Per Hedbor <ph@opera.com>

Encode UNDEFINED as null

It makes more sense than 0

2015-08-31 12:15:26 (3 days ago) by Michał Gawron <mgawron@opera.com>

MT-1211: Follow-up fix after MT-1208 (' ' interpreted as "32").
MT-1208: Fixes in pike8.

* Prevent adding binary ints to output stream when header values are ints.
* Don't try to call tell() on a parameters mapping. Seek in m->file instead.

2015-08-31 04:55:23 (3 days ago) by Martin Nilsson <nilsson@fastmail.com>

We don't use URCU anymore.

2015-08-31 04:50:27 (3 days ago) by Martin Nilsson <nilsson@fastmail.com>

Removed unused headers.
Fixed tests.

2015-08-31 04:20:14 (3 days ago) by Martin Nilsson <nilsson@fastmail.com>

No need to check for the same headers more than once.
Running SSL tests in parallel is currently broken.

Sunday 30 August 2015

2015-08-30 08:17:09 (4 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [Standards.HPack]: Added quite a bit of AutoDoc markup.

Saturday 29 August 2015

2015-08-29 08:24:15 (5 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

AutoDoc: Updated documentation of @rfc{ @} markup.

Friday 28 August 2015

2015-08-28 16:54:56 (5 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Constants: Adjusted some symbol names.

Make sure that the names of the KRB5 export suites actually contain "export".

2015-08-28 16:50:24 (5 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Constants: Adjusted some symbol names.

Make sure that the names of the KRB5 export suites actually contain "export".

Thursday 27 August 2015

2015-08-27 15:49:26 (7 days ago) by Martin Nilsson <nilsson@fastmail.com>

Rewrite for clarity.

2015-08-27 15:32:29 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Stdio.FakePipe]: socktest now also tests Stdio.FakePipe.

2015-08-27 15:31:39 (7 days ago) by Martin Nilsson <nilsson@fastmail.com>

Fixed return type.

2015-08-27 15:26:55 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Stdio]: Support testing of emulated sockets in socktest.pike.

Avoid use of dup() and assign() as they are hard to emulate, and
not used by any normal code.

2015-08-27 15:22:45 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Stdio: Initial implementation of Stdio.FakePipe.

This is a class that attempts to emulate a bi-directional pipe/socket,
but without using any fds.

The intended use is to simplify implementation of APIs that require
emulation of sockets (like eg SSL.File).

2015-08-27 14:45:42 (7 days ago) by Per Hedbor <ph@opera.com>

Backported locked check fix for Stdio.Buffer from 8.1

Wednesday 26 August 2015

2015-08-26 20:39:19 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Parse EXIF as soon as that chunk is discovered. Ignore all tags except Orientation.
Allow for only specific tags to be parsed and the rest ignored.

2015-08-26 19:58:50 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Comment update

2015-08-26 13:50:24 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

GTK2: GdkPixbufSimpleAnim {s,g}et_loop() require GTK 2.18 or later.

2015-08-26 13:48:10 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

GTK2: GdkPixbufSimpleAnim {s,g}et_loop() require GTK 2.18 or later.

2015-08-26 10:55:36 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.101 by export.pike
release number bumped to 8.0.100 by export.pike

2015-08-26 04:11:28 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Small simplification.
Return data in the correct type.

Tuesday 25 August 2015

2015-08-25 18:50:15 (1 week ago) by Per Hedbor <ph@opera.com>

Merge remote-tracking branch 'origin/8.1' into per/substrings

2015-08-25 17:34:53 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [Protocols.LMTP]: Moved some documentation.

2015-08-25 17:32:44 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [HPack]: Stricter type in documentation.

Monday 24 August 2015

2015-08-24 19:56:32 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Added PNM P4 test.
GTK1 removed.

2015-08-24 19:42:42 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

GTK1 is deprecated

2015-08-24 16:43:50 (1 week ago) by Per Hedbor <ph@opera.com>

Merge remote-tracking branch 'origin/8.1' into per/substrings
Support the optimized substring generation in sscanf.

2015-08-24 15:46:37 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Reuse the tag parsing loop.

2015-08-24 15:34:20 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Confine the EXIF parsing in its JFIF segment.

2015-08-24 15:26:59 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Converted if-train to switch.

2015-08-24 15:11:35 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Abort if the tag type can't be parsed.
Object Orient harder. Keep offset and byteorder internal to the file object.

2015-08-24 15:01:18 (1 week ago) by Chris Angelico <rosuav@gmail.com>

Force GTK2 strings to be null-terminated

This prevents odd errors when connecting to some GTK signals

2015-08-24 14:29:22 (1 week ago) by Chris Angelico <rosuav@gmail.com>

De-staticify some GTK2 strings.

Partial reversion of 5e9fc01ab846ab05ae5b83a9b9cc6b535aa3824a

This prevents strange errors when connecting some GTK signals.

2015-08-24 14:24:18 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [HPack]: Added some RFC markup.

2015-08-24 14:22:08 (1 week ago) by Chris Angelico <rosuav@gmail.com>

De-staticify some GTK2 strings.

Partial reversion of 5e9fc01ab846ab05ae5b83a9b9cc6b535aa3824a

This prevents strange errors when connecting some GTK signals.

2015-08-24 14:03:32 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Tools.Standalone.autodoc_to_html: Take advantage of the RFC redirect.

Reduce knowledge about the site structure at http://pike.lysator.liu.se/
by taking advantage of the automatic RFC redirect at the site.

2015-08-24 03:23:13 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Properly parse JPEG headers instead of just brute force scanning first 100 bytes. This avoids false positives or parsing EXIF data in other than APP1 segments, and more importantly finds EXIF data that is further than 100 bytes from the start of the file (e.g. after ICC profile).

2015-08-24 01:57:11 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Ignore more files.

Sunday 23 August 2015

2015-08-23 22:25:41 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Support #[#] style string quoting in Pike tokenizer.
Support #[#] style string quoting in Pike tokenizer.
Test some oddly placed white spaces in preprocessor string quoting.
Support #[#] style string quoting in Pike tokenizer.

2015-08-23 22:14:41 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Don't crash in images with zero color range.

2015-08-23 22:11:54 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Don't crash in images with zero color range.

2015-08-23 20:40:22 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Added some PNM examples.

2015-08-23 20:24:45 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Support #[#] style string quoting in Pike tokenizer.

2015-08-23 20:21:20 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Support #[#] style string quoting in Pike tokenizer.
Support #[#] style string quoting in Pike tokenizer.
Test some oddly placed white spaces in preprocessor string quoting.

2015-08-23 11:38:39 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build: Fixed some warnings.

The first argument to low_quick_add_function() can't be const, since
it will have references added...
Strings: Fixed some issues --with-debug.

Saturday 22 August 2015

2015-08-22 22:58:38 (2 weeks ago) by Per Hedbor <ph@opera.com>

Initial substring support.

This adds support for substrings, and conversion of said strings to
static strings.

Left to do: gc conversion of substrings to malloced/short strings.

This requires two new stringtypes, at least.

The reason is that the substring string type is bigger than the normal
strings. I will most likely do it by adding a 2 or 1 to the 'parent'
pointer.

2015-08-22 21:44:39 (2 weeks ago) by Per Hedbor <ph@opera.com>

Merge branch 'per/string_split_merge' into 8.1

Conflicts:
src/builtin.cmod
src/stralloc.c

2015-08-22 20:54:15 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Support linking direct to RFC anchor.

2015-08-22 20:44:07 (2 weeks ago) by Per Hedbor <ph@opera.com>

Fixed accidentally included changes to hash table lookups.

(cherry picked from commit 28d7bb4dafe33d30585deb4599d8a06c933e296d)
Fixed accidentally included changes to hash table lookups.

2015-08-22 20:35:10 (2 weeks ago) by Per Hedbor <ph@opera.com>

Fixed possible leak-on-oom in begin_*shared_string

Unexported some functions, use xrealloc + xalloc instead of malloc + check.
Static:ified a lot of strings.

This is more or less all instances of push_text(static data here).

Left to do: make_shared_X_string(static data here). :)
un-exported some of the unused functions from stralloc.h

2015-08-22 20:35:09 (2 weeks ago) by Arne Goedeke <el@laramies.com>

Strings: print string types in memory_usage()
Strings: generate more static strings

quick_add_function and make_shared_string will now create
static strings when the argument is a compile time constant
Strings: correctly count bytes in strings
Strings: added and used push_static_text

Use the new static string allocation support where possible. This is the
case all code which is not dynamically loaded.

2015-08-22 20:32:00 (2 weeks ago) by Arne Goedeke <el@laramies.com>

strings: split string header and contents

All string structs are now allocated by the block allocator while the
actual string content is kept in a different area. Short strings with
byte length smaller than sizeof(struct pike_string) have their content
allocated from the block allocator, aswell. In most cases short string
contents are allocated on creation and therefore the data will end up
directly after the header in memory.
Strings: global const strings

2015-08-22 20:32:00 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Strings: Survive xalloc() throwing errors in realloc.

Short string contents would be passed to free(3C) if the
call to xalloc() failed.

Same for truncating strings if ba_alloc() ever starts throwing
errors instead of fatals.

2015-08-22 20:32:00 (2 weeks ago) by Arne Goedeke <el@laramies.com>

Strings: Readded handling for short string in add_string_status()

2015-08-22 20:24:17 (2 weeks ago) by Per Hedbor <ph@opera.com>

Fixed possible leak-on-oom in begin_*shared_string

Unexported some functions, use xrealloc + xalloc instead of malloc + check.
un-exported some of the unused functions from stralloc.h

2015-08-22 12:07:10 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [Crypto.Password]: Added some RFC markup.

2015-08-22 02:24:42 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Autodoc fixes.

2015-08-22 02:22:11 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Autodoc fixes.

2015-08-22 02:15:23 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Use @rfc{@} autodoc syntax.

2015-08-22 01:56:32 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Use @rfc{@} autodoc syntax.

Friday 21 August 2015

2015-08-21 16:53:51 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [MIME]: Use the new RFC reference markup.
Documentation [MIME]: Use the new RFC reference markup.

2015-08-21 15:54:45 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [MIME]: Use the new RFC reference markup.

2015-08-21 15:52:40 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [MIME]: Use the new RFC reference markup.

2015-08-21 11:35:02 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Tools.Autodoc: Added markup for references to RFCs.
Tools.Autodoc: Added markup for references to RFCs.

Thursday 20 August 2015

2015-08-20 13:59:37 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.TreeScheduler: Added scheduler from RFC 7540 section 5.3.

2015-08-20 10:15:47 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Thread.Thread [Documentation]: Adjusted return type of thread function.

The return value of the thread function is relevant (as it is returned
by wait())...

Fixes [LysLysKOM 21410113] reported by Per Cederqvist. Thanks.

Also adjusted the declared type of create() somewhat so that it requires
a thread function (ie the compiler will now complain if it is called
without arguments).

2015-08-20 10:09:16 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Thread.Thread [Documentation]: Adjusted return type of thread function.

The return value of the thread function is relevant (as it is returned
by wait())...

Fixes [LysLysKOM 21410113] reported by Per Cederqvist. Thanks.

Also adjusted the declared type of create() somewhat so that it requires
a thread function (ie the compiler will now complain if it is called
without arguments).

Wednesday 19 August 2015

2015-08-19 15:13:21 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.ServerConnection: Fixed session resumption.

The SMACKTest changes broke session resumption (by sending
double CCS on the server side for resumed sessions).

Tuesday 18 August 2015

2015-08-18 20:07:49 (2 weeks ago) by Per Hedbor <ph@opera.com>

Static:ified a lot of strings.

This is more or less all instances of push_text(static data here).

Left to do: make_shared_X_string(static data here). :)

2015-08-18 18:55:08 (2 weeks ago) by Per Hedbor <ph@opera.com>

Merge branch 'arne/string_alloc' into 8.1

Conflicts:
src/backend.cmod
src/builtin.cmod
src/builtin_functions.c
src/stralloc.c

2015-08-18 15:09:11 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: Use ADT.Heap()->remove().

There's no need to perform the remove()-operation by hand anymore.

Monday 17 August 2015

2015-08-17 17:44:55 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.99 by export.pike
release number bumped to 8.0.98 by export.pike

2015-08-17 16:09:30 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.ServerConnection: Delay sending of CCS.

Delay sending of the Change Cipher Spec message until we have
received and verified the Finished message from the peer.

This makes SMACKTest (http://smacktest.com/) happy.

2015-08-17 15:26:18 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Context: Default to the FFDHE2048 group.

Changes the default DHE group from MODP group 24 to FFDHE2048.

This makes SMACKTest (http://smacktest.com/) happy, and reduces
the risk of precalculated attacks against the MODP group.

2015-08-17 12:48:13 (2 weeks ago) by Per Hedbor <ph@opera.com>

Image.ANY: Reduced code duplication and added WebP.

2015-08-17 12:38:43 (2 weeks ago) by Per Hedbor <ph@opera.com>

Fixed decoding of WebP images with alpha

2015-08-17 12:35:04 (2 weeks ago) by Per Hedbor <ph@opera.com>

Added basic WebP support to 8.0 Image.ANY.decode

2015-08-17 12:11:56 (2 weeks ago) by Per Hedbor <ph@opera.com>

Fixed decoding of WebP images with alpha

Sunday 16 August 2015

2015-08-16 13:47:25 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Priority_queue: Support floating point priorities.

Saturday 15 August 2015

2015-08-15 11:48:22 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Updated to tzdata2015f.
Updated to tzdata2015f.

Friday 14 August 2015

2015-08-14 18:06:15 (3 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Stdio.Buffer gives the buffer object as second argument, while Error.Generic wants an array.

2015-08-14 16:44:31 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Updated to tzdata2015e.
Updated to tzdata2015e.

2015-08-14 12:22:01 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

GTK2: Fixed invalid shift in encode_truecolor_generic().

Make sure to widen the value before shifting it.

Fixes [CID 1294562].

2015-08-14 12:13:07 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

GTK2: Fixed invalid shift in encode_truecolor_generic().

Make sure to widen the value before shifting it.

Fixes [CID 1294562].

Thursday 13 August 2015

2015-08-13 11:19:14 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Heap: Shrink the value array on remove().
ADT.Heap: Shrink the value array on remove().

2015-08-13 11:15:18 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [ADT.Heap]: Added some cross-references.
ADT.Heap: Added remove().

2015-08-13 11:13:10 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [ADT.Heap]: Added some cross-references.
ADT.Heap: Added remove().

Wednesday 12 August 2015

2015-08-12 17:50:48 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: Keep Consumer()->offset up to date.

This simplifies code that needs to check the priority of
potentially inactive consumers.

2015-08-12 17:47:36 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: Keep the priority quanta in a variable.

This is to support caching of more complicated quanta calculations.

2015-08-12 17:43:49 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: Increased the priority normalization theshold to 256.

This is to support weights normalized to the range [0,1].

Tuesday 11 August 2015

2015-08-11 16:28:17 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Postgres: Removed some erroneous frees. Fixes [CID 1315417].

2015-08-11 16:19:59 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: LFUNs should typically be protected.
ADT.Scheduler: Bugfix.

Monday 10 August 2015

2015-08-10 15:09:35 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Scheduler: First version.

This is a data type intended to be used to distribute quanta of
a resource fairly among a set of prioritized consumers.

Sunday 09 August 2015

2015-08-09 17:13:42 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Svalues: Added PIKE_-prefix to @{,SUB}TYPEOF().

It seems Freetype 2.6 has a TYPEOF() in one of its headerfiles.

2015-08-09 13:07:18 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

ADT.Priority_queue: Simplify overloading.

adjust_pri() now uses elem()->set_pri() instead of accessing
the priority variable directly.

Saturday 08 August 2015

2015-08-08 23:08:30 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Fixed WebP extended header.

2015-08-08 22:55:52 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Support for WebP

2015-08-08 21:27:52 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Some refactoring.

2015-08-08 20:51:33 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Move the compat fallback for Dims.get into the compat module. Doesn't actually work currently, as the resolver in the image module doesn't know about compat resolvers.

2015-08-08 09:57:27 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

sybase: Converted module to cmod.

2015-08-08 04:58:35 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Optimize the calling of some of the most used image decoders by avoiding decoding all meta data. In addition this reduces the risk of crashing in the metadata code.

2015-08-08 04:40:19 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Test Image.ANY a bit. Made JPEG test fuzzier.

Friday 07 August 2015

2015-08-07 23:20:00 (4 weeks ago) by Per Hedbor <ph@opera.com>

Run SSL client/server tests in parallel.

2015-08-07 16:31:07 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler: LFUN prototypes are not useful...

Reenable the code that hid prototypes from the LFUN lookup table.

Thursday 06 August 2015

2015-08-06 23:23:39 (4 weeks ago) by Per Hedbor <ph@opera.com>

Todays micro-optimization: avoid re-doing overflowing multiplications.

Use the full 128-bit result from imul. Not really all that much of an
improvement in most cases, presumably.

The same thing can however also be done for + and -, at a minimum.

In general more mpz operations could be inlined where it makes sense
(the operators, mainly).

2015-08-06 14:08:57 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Protocols.HTTP2: Added the constants from RFC 7540.

2015-08-06 03:22:40 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

More things to do.

2015-08-06 00:53:56 (4 weeks ago) by Per Hedbor <ph@opera.com>

Use the set_unaligned16/32 function
Added direct-access versions of get/set_unalignedX.

Curently enabled on x86_64 only.

Wednesday 05 August 2015

2015-08-05 14:31:45 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Connection: Added shutdown().

Tuesday 04 August 2015

2015-08-04 18:49:16 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Postgres: Removed shadowing declaration. Fixes [CID 1315196].

2015-08-04 11:27:15 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

GTK2: GdkPixbufSimpleAnim requires GTK 2.8 or later.
GTK2: GdkPixbufSimpleAnim requires GTK 2.8 or later.

2015-08-04 01:06:48 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Test a few more signature DER identifiers. The ECDSA equivalent should probably move from Crypto.ECC to the PKCS.ECDSA module.

Monday 03 August 2015

2015-08-03 16:19:27 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Stdio.Buffer: Avoid unaligned accesses.

Fixes SIGBUS when adding integers to unaligned buffers on non-x86/amd64.

2015-08-03 14:34:11 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Process]: Added workaround for Linux kernel bug.

wait(2) et al in Linux 3.15 - 3.18 are not thread safe, and a blocking
wait() can hang when a different thread reaps the same process (eg via
a signal handler).

Fixes intermittent testsuite failures on Linux 3.15 - 3.18.

2015-08-03 14:31:51 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Inotify]: Fixed the filesystem cleanup some more.
Testsuite: Survive CRNL-mode in new-style literal strings.
Testsuite [Stdio.Buffer]: Updated testsuite to new buffer strategy.

2015-08-03 14:19:47 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.File: More race-condition fixes.

Potential fix for "Indexing the NULL value with "state"." in direct_write().

2015-08-03 05:52:54 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Call rdrand according to the documentation, mixing the last random value with the next. Shouldn't make a practical difference.

Sunday 02 August 2015

2015-08-02 14:51:17 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Postgres: Fixed potential memory leak on out of memory.

2015-08-02 14:46:44 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Postgres: Fixed type for big_query().

Fixes compilation issue with Sql.postgres.

2015-08-02 14:36:21 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build: Fixed some warnings.

The inline function ptr_hashfun() is often not used.

2015-08-02 13:12:34 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build: The declaration of my_rand() has moved to stuff.h.

Fixes quite a few warnings.

2015-08-02 04:14:03 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Fixed an error message.
Don't throw if EXIF parsing fails. Don't send image data over 1000 bytes as file name. Don't throw if opening said image data as a file name fails.

2015-08-02 00:07:21 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Moved my_rand and friends from port (wrong place) to stuff (less wrong place).

Saturday 01 August 2015

2015-08-01 23:54:53 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Whitespace changes.
__builtin_clzl has been in GCC since 4.3.0

2015-08-01 21:46:40 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Work for all random_string lengths.

2015-08-01 20:08:54 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Make random_string twice as fast when using rdrnd.

2015-08-01 19:52:31 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Somewhat simpler handling of rdrnd. 3% faster.

2015-08-01 19:44:16 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Assume that builtins never worked and fix them by removing one level of quoting.

2015-08-01 17:24:36 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Use Identifier instead of object.

2015-08-01 12:28:51 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Added convenience variant of decode().

2015-08-01 11:43:28 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Standards.HPack]: Added the testvectors from RFC 7541.

Context()->encode() and Context()->decode() are now tested.

2015-08-01 05:01:04 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Compilation fixes.

Friday 31 July 2015

2015-07-31 18:10:25 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Process]: Added workaround for Linux kernel bug.

wait(2) et al in Linux 3.15 - 3.18 are not thread safe, and a blocking
wait() can hang when a different thread reaps the same process (eg via
a signal handler).

Fixes intermittent testsuite failures on Linux 3.15 - 3.18.

2015-07-31 17:29:20 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Tools.Standalone.benchmark: Adjusted table field sizes.

The field size for test names has been extended from 39 to 41
characters since there's one benchmark test that has a 40
character name ("Adding element to array (private global)").

This fixes the missalinged result for that test.

Also adds a missing '-' at the last line.

2015-07-31 16:56:16 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Use this instead of this_object().

2015-07-31 11:22:49 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.File [Threads]: Reduce race-condition window.

If the SSL.File was closed by one thread while another was performing a
blocking write(), the second thread would sometimes index the destructed
Connection object with either "to_write" or "state".

Also fixes a similar problem in is_open().

This should fix intermittent SSL testsuite failures on Solaris.

2015-07-31 02:10:32 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Mark read and write as deprecated. Added documentation.
Use put instead of read on ADT.Queue.
Removed a few unneeded object.

2015-07-31 01:11:12 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Stricter types.

2015-07-31 00:27:14 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Hash can work on Stdio.Buffer objects directly.

2015-07-31 00:12:40 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Revert "Process: Reduce code compexity in the USE_SIGCHILD case."

This reverts commit b961f4d4e42cf355d952e0d3a9a13bb20056d61e.

Thursday 30 July 2015

2015-07-30 18:33:49 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Pike.identify_cycle: Fixed global variable overwrite.

Pike.identify_cycle() uses a mapping internally, and insert into
mappings can call Pike code (via LFUN:__hash()), which could
have a different thread also start messing with the same variables.

This fixes the multi-thread case, but there's still a problem if it
starts to recurse (ie LFUN::__hash() calling Pike.identify_cycle()
or Pike_count_memory()).
Build [no threads]: Added missing dummy macro.

2015-07-30 18:25:48 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Pike.identify_cycle: Fixed global variable overwrite.

Pike.identify_cycle() uses a mapping internally, and insert into
mappings can call Pike code (via LFUN:__hash()), which could
have a different thread also start messing with the same variables.

This fixes the multi-thread case, but there's still a problem if it
starts to recurse (ie LFUN::__hash() calling Pike.identify_cycle()
or Pike_count_memory()).

2015-07-30 18:23:23 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build [no threads]: Added missing dummy macro.

2015-07-30 16:29:59 (1 month ago) by Build system <distmaker@roxen.com>

release number bumped to 931 by export.pike
release number bumped to 930 by export.pike

2015-07-30 12:10:43 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [SSL]: Potential fix for intermittent failures.

SSL.Context()->get_suites() depends on {min,max}_version. If an earlier
test failed with server_ctx->max_version < PROTOCOL_TLS_1_2, this would
cause the later ECC tests to fail due to the new server_ctx lacking AEAD
suites. The new server_ctx should now contain the full set of suites.

Wednesday 29 July 2015

2015-07-29 13:17:13 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Setting the static max size now actually works.

2015-07-29 13:06:23 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Parser.HPack: Fixed encoder support for unindexed headers.

2015-07-29 12:54:16 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Fixed bug in add_header().

New headers are added last...

Tuesday 28 July 2015

2015-07-28 14:57:55 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Optimized huffman_decode() on LP64.

The internal bitbuffer in huffman_decode() is now an INT_TYPE
(and thus 64-bits on relevant architectures). This reduces the
number of special cases.

Monday 27 July 2015

2015-07-27 20:02:23 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Stdio.Buffer: Modifying read-only buffers is NOT a good idea...

Fixes SSL testsuite failure.

2015-07-27 13:00:57 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Stdio.Buffer: Added some locking to _sprintf().

The buffer could get trimmed during _sprintf() before it was rewound.

Sunday 26 July 2015

2015-07-26 13:49:43 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [Stdio.Buffer]: Updated testsuite to new buffer strategy.

2015-07-26 13:30:59 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite: Survive CRNL-mode in new-style literal strings.

2015-07-26 11:40:12 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Added encoding flags.

encode_header() now accepts a flag argument, and decode_header() exposes
the same when it is semantically significant.

Adds HEADER_INDEXED (default), HEADER_NOT_INDEXED (used for headers
that will only be used once) and HEADER_NEVER_INDEXED (used to reduce
exposure of passwords and sensitive cookies to on-line attackers).

2015-07-26 11:35:58 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Strings: Fixed hashtable breakage in stralloc_rehash().

The string hash table could get filled with random pointers.

Saturday 25 July 2015

2015-07-25 12:07:11 (1 month ago) by Per Hedbor <ph@opera.com>

Flat hash, now with similar speeds.

Still somewhat slower, 5% for existing strings in the benchmark on my
laptop.

I am somewhat uncertain why, since the average run length in that test
is 1.09.

Anyway, it saves 8 bytes per string. Worth it or not? Needs more
testing. :)

2015-07-25 09:47:31 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.HPack: Improved memory use in evict_dynamic_headers().

evict_dynamic_headers() now truncates the dynamic headers array
when more than 3/4:s of it are unused.

Also adds a special case for when dynamic_max_size is zero.

[permalink]

Bugs mentioned

  7540NEWcimg does not honor settings from static resources
Bugs? Suggestions?