Browse source

Changelog

Query
From:
To:
Branch:
Path:
Message:
BugID:
User: +
Format:
: 1 checkin (+6/-0) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 5 checkins (+84/-57) by 3 people : 5 checkins (+28/-8) by 2 people : 6 checkins (+95/-70) by 2 people : 56 checkins (+3819/-473) by 3 people : 14 checkins (+279/-184) by 4 people : 4 checkins (+82/-22) by 2 people : 3 checkins (+79/-24) by 2 people : 2 checkins (+16/-1) by 2 people : 49 checkins (+3858/-381) by 2 people : 2 checkins (+7/-1) by 2 people : 11 checkins (+116/-45) by 2 people : 6 checkins (+187/-108) by 2 people : 3 checkins (+55/-9) by 2 people : 7 checkins (+311/-158) by 2 people : 7 checkins (+122/-113) by 2 people : 3 checkins (+51/-55) by 2 people : 8 checkins (+148/-233) by 2 people : 3 checkins (+166/-160) by 2 people : 10 checkins (+132/-71) by 3 people : 5 checkins (+19/-13) by 3 people : 5 checkins (+62/-4) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 10 checkins (+71/-30) by 4 people : 5 checkins (+14/-7) by 3 people : 2 checkins (+5/-3) by 2 people : 17 checkins (+449/-84) by 3 people : 5 checkins (+14/-9) by 3 people : 7 checkins (+782/-1040) by 2 people : 6 checkins (+28/-2) by 2 people : 10 checkins (+224/-12) by 4 people : 3 checkins (+49/-0) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 4 checkins (+10/-7) by 3 people : 10 checkins (+29/-10) by 2 people : 6 checkins (+68/-24) by Henrik Grubbström (Grubba) <grubba@grubba.org> : 19 checkins (+1542/-82) by 5 people : 10 checkins (+1049/-299) by 2 people : 22 checkins (+702/-476) by 2 people : 21 checkins (+1303/-815) by 3 people : 5 checkins (+77/-22) by 2 people : 11 checkins (+325/-215) by 4 people Lines added/removed recently

Today; Monday 26 September 2016

2016-09-26 01:16:39 (5 hours ago) by Martin Nilsson <nilsson@fastmail.com>

copy_mapping defined twice.

Yesterday; Sunday 25 September 2016

2016-09-25 21:40:47 (8 hours ago) by Tobias S. Josefowitz <tobij@tobij.de>

ARM32: complete fast check threads

Slow path now only generated once per program.

2016-09-25 21:38:48 (8 hours ago) by Arne Goedeke <el@laramies.com>

ARM32: some debugging checks

2016-09-25 21:38:48 (8 hours ago) by Tobias S. Josefowitz <tobij@tobij.de>

ARM32: preliminary check_threads_etc

Uses dedicated register for the fast_check_threads_etc counter but
is generated in too many places for now.

2016-09-25 21:38:48 (8 hours ago) by Arne Goedeke <el@laramies.com>

Compiler: make whole program executable

Some machine code backends generate commonly used at the beginning of a
program and jump there from other opcodes. To make constant evaluation
work we need to flush instruction caches for the full program.
ARM32: check if MOVW/MOVH are supported
ARM32: removed unused code
ARM32: de-inlined F_RETURN
ARM32: de-inline free_svalue

To reduce code size, we generate a simple free_svalue version
at the beginning of every program. When freeing an svalue, we
jump to that de-inlined version for every type which is reference
counted. This reduces the size of the generated machine code by around
10%.
ARM32: use Pike_fatal for assert

2016-09-25 20:38:43 (9 hours ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: More code cleanup.

Adds string_builder_append_cpp_quoted_pcharp().

Saturday 24 September 2016

2016-09-24 14:38:20 (2 days ago) by Martin Nilsson <nilsson@fastmail.com>

Fix for missing SECP curves.

2016-09-24 14:32:55 (2 days ago) by Martin Nilsson <nilsson@fastmail.com>

Fix for missing SECP curves.

2016-09-24 13:24:23 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Removed some redundant tests.

Friday 23 September 2016

2016-09-23 20:09:24 (2 days ago) by Stephen R. van den Berg <srb@cuci.nl>

EngineIO/SocketIO: Improve documentation.
EngineIO: Close race during destruction.
EngineIO: Close race during destruction.
EngineIO/SocketIO: Improve documentation.

2016-09-23 18:21:36 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Nettle: Probe for the SECP curves.

Some versions of RedHat have censored the weaker curves
(secp_192r1 abd secp_224r1), so we need to survive this.

Thanks to Martin Bähr <mbaehr+pike@realss.com> for the report.

Fixes [LysLysKOM 21863021].

2016-09-23 18:10:05 (2 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Nettle: Probe for the SECP curves.

Some versions of RedHat have censored the weaker curves
(secp_192r1 abd secp_224r1), so we need to survive this.

Thanks to Martin Bähr <mbaehr+pike@realss.com> for the report.

Fixes [LysLysKOM 21863021].

2016-09-23 16:02:47 (3 days ago) by Stephen R. van den Berg <srb@cuci.nl>

EngineIO: Prevent race if destruction and upgrade cross each other.
EngineIO: Prevent race if destruction and upgrade cross each other.

2016-09-23 15:06:20 (3 days ago) by Stephen R. van den Berg <srb@cuci.nl>

EngineIO/SocketIO:

Rename API to more closely follow the upstream naming conventions.
Add events per namespace: on(), off().
EngineIO/SocketIO:

Rename API to more closely follow the upstream naming conventions.
Add events per namespace: on(), off().

2016-09-23 10:18:15 (3 days ago) by Tobias S. Josefowitz <tobij@tobij.de>

Process.create_process: Be more stern if read() fails with EINTR

Turns out you can get interrupted in read() more than 16 times under
cerrtain workloads.

2016-09-23 10:15:03 (3 days ago) by Tobias S. Josefowitz <tobij@tobij.de>

Process.create_process: Be more stern if read() fails with EINTR

Turns out you can get interrupted in read() more than 16 times under
cerrtain workloads.
Process.create_process: Be more stern if read() fails with EINTR

Turns out you can get interrupted in read() more than 16 times under
cerrtain workloads.

Thursday 22 September 2016

2016-09-22 19:18:59 (3 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.315 by export.pike
release number bumped to 8.0.314 by export.pike

2016-09-22 19:05:17 (3 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

mariadb-config: Added new tool.

2016-09-22 18:57:49 (3 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

rntcl: Ignore -fPIC and -fpic.

Some Makefiles (eg poppler) assume that all C++ compilers are GCC
and/or that they understand the option -fPIC.

2016-09-22 18:03:56 (3 days ago) by Stephen R. van den Berg <srb@cuci.nl>

SocketIO/EngineIO:

Define PDT() debugging macro.
Define DUSERERROR() macro.
Throw proper errors with backtraces.
Link statically as far as possible.
Provide sid and lastrequest members on SocketIO.Server.
Pass up ERROR messages from the protocol.

2016-09-22 18:01:35 (3 days ago) by Stephen R. van den Berg <srb@cuci.nl>

SocketIO/EngineIO:

Define PDT() debugging macro.
Define DUSERERROR() macro.
Throw proper errors with backtraces.
Link statically as far as possible.
Provide sid and lastrequest members on SocketIO.Server.
Pass up ERROR messages from the protocol.

2016-09-22 15:57:47 (4 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.312 by export.pike
release number bumped to 8.0.313 by export.pike

2016-09-22 15:35:40 (4 days ago) by Stephen R. van den Berg <srb@cuci.nl>

EngineIO: Make it threadsafe(r), and fix a protocol issue with mixed content.
SocketIO: Extra debugging, fix ack callback.
SocketIO: Make it thread safe, and simplify the API.

2016-09-22 15:26:24 (4 days ago) by Stephen R. van den Berg <srb@cuci.nl>

SocketIO: Make it thread safe, and simplify the API.
EngineIO: Make it threadsafe(r), and fix a protocol issue with mixed content.
SocketIO: Extra debugging, fix ack callback.

2016-09-22 12:35:57 (4 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.311 by export.pike
release number bumped to 8.0.310 by export.pike

2016-09-22 11:14:10 (4 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler: Object variables declared as mixed may contain this...

Don't reference count direct cyclic references via mixed variables.

Fixes [LysLysKOM 21860701]/[Pike mailinglist 14570].
Compiler: Object variables declared as mixed may contain this...

Don't reference count direct cyclic references via mixed variables.

Fixes [LysLysKOM 21860701]/[Pike mailinglist 14570].

Wednesday 21 September 2016

2016-09-21 20:41:26 (4 days ago) by Stephen R. van den Berg <srb@cuci.nl>

SocketIO: Turn off debugging.
SocketIO: Turn off debugging.

2016-09-21 20:39:08 (4 days ago) by Stephen R. van den Berg <srb@cuci.nl>

Clean-room basic Socket.IO implementation.

Namespaces are supported.
Events belong in the next layer.
Rooms belong even one layer further up.
EngineIO:

_sprintf() includes protocol revision.
Add lastrequest member.
Consolidate setting callbacks to a single function.
Binary frames have their type in binary sometimes too.
Run the close_cb just prior to shutting down the write channel.
Use the faster sprintf in favour of Buffers for %c%s type situations.
Add accessible close() method.
Avoid circular references in id.
Close Transport without destructing.
Tie up dangling refences when closing is forced.
Clean-room basic Socket.IO implementation.

Namespaces are supported.
Events belong in the next layer.
Rooms belong even one layer further up.
EngineIO:

_sprintf() includes protocol revision.
Add lastrequest member.
Consolidate setting callbacks to a single function.
Binary frames have their type in binary sometimes too.
Run the close_cb just prior to shutting down the write channel.
Use the faster sprintf in favour of Buffers for %c%s type situations.
Add accessible close() method.
Avoid circular references in id.
Close Transport without destructing.
Tie up dangling refences when closing is forced.

2016-09-21 16:18:25 (5 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.309 by export.pike
release number bumped to 8.0.308 by export.pike

2016-09-21 12:23:01 (5 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Some code clean up.

Adds internal function apply_define().

Tuesday 20 September 2016

2016-09-20 22:19:11 (5 days ago) by Martin Karlgren <marty@roxen.com>

Stdio.FakeFile: Improved sendfile compatibility.
Stdio.FakeFile: Improved sendfile compatibility.

2016-09-20 18:30:20 (5 days ago) by Chris Angelico <rosuav@gmail.com>

GTK2: Document GdkPixbufLoader stubbily so the magic happens

(Do the magic! Do the magic!)

2016-09-20 15:53:55 (6 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build [Postgres]: Detect Postgres 10.x and later.

2016-09-20 11:44:43 (6 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.307 by export.pike
release number bumped to 8.0.306 by export.pike

2016-09-20 10:55:08 (6 days ago) by Arne Goedeke <el@laramies.com>

Protocols.WebSocket: some client fixes

* do not send host header twice
* mask all frames
* connect to the correct port
* pass full query string in HTTP request
Protocols.WebSocket: close connection manually

Calling destruct is more reliable than waiting for the file to go out
of references.
Protocols.WebSocket: added defragment()

defragment() can be used to automatically recombine fragmented frames.
Protocols.WebSocket: handle more protocol errors

The following protocol errors are now handled:

* invalid close status codes
* invalid utf8 in TEXT frames
* long control frames
* fragmented control frames
* check that the reserved bits are zero

Some of these checks are only active if WEBSOCKET_DEBUG is turned on,
especially those which might break future use of WebSocket extensions.

2016-09-20 10:34:43 (6 days ago) by Arne Goedeke <el@laramies.com>

Protocols.WebSocket: added API for continuation frames

2016-09-20 10:08:36 (6 days ago) by Chris Angelico <rosuav@gmail.com>

Silence gcc warning by clarifying that the if guards just one statement

Another instance of the same change as c8622d

2016-09-20 09:58:08 (6 days ago) by Stephen R. van den Berg <srb@cuci.nl>

Clean-room Engine.IO implementation.
WebSocket: Call either http_cb or ws_cb, not both, at the proper moment.

2016-09-20 09:50:12 (6 days ago) by Stephen R. van den Berg <srb@cuci.nl>

WebSocket: Call either http_cb or ws_cb, not both, at the proper moment.
Clean-room Engine.IO implementation.

Monday 19 September 2016

2016-09-19 15:42:02 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.305 by export.pike
release number bumped to 8.0.304 by export.pike

2016-09-19 15:19:54 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [master]: Improved thread-safety.

Cast to program and cast to object should now be thread-safe.

Fixes some more of [bug 7783].

2016-09-19 15:17:39 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [master]: Improved thread-safety.

Cast to program and cast to object should now be thread-safe.

Fixes some more of [bug 7783].

2016-09-19 12:29:00 (7 days ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

release number bumped to 8.0.303 by export.pike
release number bumped to 8.0.302 by export.pike

2016-09-19 00:44:11 (1 week ago) by Chris Angelico <rosuav@gmail.com>

GTK2: Fix return value of get_gravity
GTK2: Fix return value of get_gravity

Sunday 18 September 2016

2016-09-18 21:33:51 (1 week ago) by Chris Angelico <rosuav@gmail.com>

Reference correct function in docs
Reference correct function in docs

2016-09-18 19:00:05 (1 week ago) by Chris Angelico <rosuav@gmail.com>

GTK2: Suppress compilation warning

2016-09-18 10:25:18 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [BSON]: Fix test on 32-bit platforms.

Saturday 17 September 2016

2016-09-17 19:54:35 (1 week ago) by Chris Angelico <rosuav@gmail.com>

GTK2: Use a proper is_a check so subclasses are correctly handled

2016-09-17 14:03:55 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Constants: Fixed typo in name of constant.
SSL.Constants: Fixed typo in name of constant.

2016-09-17 10:34:33 (1 week ago) by Martin Nilsson <nilsson@fastmail.com>

Increase default compression level to 9.

Friday 16 September 2016

2016-09-16 11:34:28 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [master]: Improved thread-safety.

Protect against the same file being compiled concurrently in multiple
threads.

Fixes some more of [bug 7783].
Compiler: Added CompilerEnvironment()->lock().

This is a class that exposes the compiler internal lock.

Fixes some of [bug 7783].

2016-09-16 11:20:14 (1 week ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [master]: Improved thread-safety.

Protect against the same file being compiled concurrently in multiple
threads.

Fixes some more of [bug 7783].

Thursday 15 September 2016

2016-09-15 23:07:45 (1 week ago) by Stephen R. van den Berg <srb@cuci.nl>

WebSocket: Parse query variables.
WebSocket: Parse query variables.

2016-09-15 16:51:21 (2 weeks ago) by Chris Angelico <rosuav@gmail.com>

Add a 'level' parameter to setsockopt(), and add more constants
Add set_nodelay() function to Stdio.File objects to control Nagling
Document the link between set_keepalive and setsockopt

2016-09-15 16:45:40 (2 weeks ago) by Chris Angelico <rosuav@gmail.com>

Implement Stdio.File()->setsockopt()

2016-09-15 16:41:52 (2 weeks ago) by Arne Goedeke <el@laramies.com>

Stdio: added getprotobyname()

2016-09-15 13:31:49 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler: Added CompilerEnvironment()->lock().

This is a class that exposes the compiler internal lock.

Fixes some of [bug 7783].

Wednesday 14 September 2016

2016-09-14 19:28:56 (2 weeks ago) by unknown <unknown user@unknown>

release number bumped to 8.0.301 by export.pike
release number bumped to 8.0.300 by export.pike

2016-09-14 11:12:01 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler: Fixed broken range optimization.

The range operator was lost in the expression

A = (A + B)[low..high];

due to a fallthrough.

Fixes [bug 7780].
Compiler: Fixed broken range optimization.

The range operator was lost in the expression

A = (A + B)[low..high];

due to a fallthrough.

Fixes [bug 7780].

Tuesday 13 September 2016

2016-09-13 20:41:47 (2 weeks ago) by Arne Goedeke <el@laramies.com>

ARM: merge pike register handling

Most of the register loading and storing is now shared between 32 and 64 bit.

2016-09-13 20:24:08 (2 weeks ago) by Arne Goedeke <el@laramies.com>

ARM: first step at merging arm 32 and 64 bit code base

moved label handling and register allocation into one source file.

2016-09-13 16:11:24 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

RandomSystem: Improved robustness against EINTR.

Monday 12 September 2016

2016-09-12 21:04:47 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Micro-optimization: Don't use a string of only one character as needle for has_value.

2016-09-12 20:13:04 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Revert use of String.range.

2016-09-12 16:11:56 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [cpp]: Fixed typo.

2016-09-12 14:34:36 (2 weeks ago) by Anders Johansson <anders@roxen.com>

release number bumped to 8.0.299 by export.pike
release number bumped to 8.0.298 by export.pike

Sunday 11 September 2016

2016-09-11 22:38:29 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM32 ARM64: Update PC before calling builtins

2016-09-11 22:20:53 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Encode array elements sequentially.

2016-09-11 22:16:31 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Added some tests from libbson
Added _sprintf

2016-09-11 22:03:35 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Sort defines.
Use String.range

2016-09-11 21:57:30 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Let's lower the paranoia and assume f_zero_type and f_search push integers.

2016-09-11 21:40:30 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Correct encoding of arrays. Order must be sequential.

2016-09-11 17:35:11 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: implement F_PRIVATE_GLOBAL(_AND_POP)

2016-09-11 17:19:06 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: optimize handling of globals

* keep Pike_fp->current_storage in a register
* keep track of whether or not the current object
could have been destructed
* actually start using F_PRIVATE_GLOBAL

2016-09-11 16:56:23 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: implement F_PRIVATE_GLOBAL

2016-09-11 16:35:42 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: added the F_CALL_BUILTIN opcodes

It would be useful to know at assembler time, if a given efun has
a return value or not.

2016-09-11 15:39:21 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: added missing argument to inc/dec opcodes

2016-09-11 15:18:08 (2 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: added INC/DEC opcodes

2016-09-11 13:04:34 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

String: Added QUOTE_NORMALIZE_WS.

Sequences of whitespace and control characters can now be consolidated
to a single SP by string_builder_quote_string().

Saturday 10 September 2016

2016-09-10 12:12:14 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [cpp]: Fixed warning generated by example.

2016-09-10 03:11:25 (2 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Refer to the HeaderParser as Protocols.HTTP.Server.HeaderParser instead.

Friday 09 September 2016

2016-09-09 18:14:40 (2 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Fixed refcount bug in simple_add_define().

2016-09-09 11:20:22 (2 weeks ago) by Chris Angelico <rosuav@gmail.com>

Silence gcc warning by clarifying that the if guards just one statement
Fix typo in docs

2016-09-09 10:54:32 (2 weeks ago) by Anders Johansson <anders@roxen.com>

release number bumped to 8.0.296 by export.pike
release number bumped to 8.0.297 by export.pike

Thursday 08 September 2016

2016-09-08 15:05:55 (3 weeks ago) by Arne Goedeke <el@laramies.com>

CHANGES: some more reordering

2016-09-08 14:53:26 (3 weeks ago) by Per Cederqvist <cederp@opera.com>

release number bumped to 8.1.6 by export.pike
release number bumped to 8.1.5 by export.pike

2016-09-08 14:52:01 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Renamed field "inside" to "flags".

Also adds CPP_MACRO_* for the valid values.

2016-09-08 14:44:07 (3 weeks ago) by Per Cederqvist <cederp@opera.com>

LysKOM: add support for 121=local-to-global-reverse

2016-09-08 14:10:42 (3 weeks ago) by Peter Bortas <bortas@gmail.com>

release number bumped to 8.0.294 by export.pike
release number bumped to 8.0.295 by export.pike

2016-09-08 13:59:54 (3 weeks ago) by Peter Bortas <bortas@gmail.com>

Reordered the naewest changes a bit.

2016-09-08 13:23:35 (3 weeks ago) by Per Cederqvist <cederp@opera.com>

Stdio.UDP()->query_fd: fix prototype

The function takes no arguments and returns int, but was
declared to take an int and return an object.

Wednesday 07 September 2016

2016-09-07 15:31:09 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

CHANGES: Added some more entries about recent changes.

2016-09-07 12:43:50 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Fixed stringify operator on strings with quotes.

2016-09-07 12:36:58 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Fixed stringify operator on strings with quotes.

Tuesday 06 September 2016

2016-09-06 22:09:59 (3 weeks ago) by Peter Bortas <bortas@gmail.com>

release number bumped to 8.0.293 by export.pike
release number bumped to 8.0.292 by export.pike

2016-09-06 17:06:35 (3 weeks ago) by Arne Goedeke <el@laramies.com>

Build: activate ARM32 machine code in Thumb mode

After some testing, it seems like interworks just works as expected.

2016-09-06 12:13:12 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Changed API for alloc_empty_defile() some more.

It now no longer steals a reference to the macro name.

2016-09-06 00:57:17 (3 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Removed some redundancy in adding defines.

2016-09-06 00:36:07 (3 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

hashtable.[ch] is no more.

Monday 05 September 2016

2016-09-05 21:51:33 (3 weeks ago) by Arne Goedeke <el@laramies.com>

Interpreter: use Pike_fp->current_storage
AMD64: use Pike_fp->current_storage

2016-09-05 21:29:00 (3 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: implement F_PRIVATE_GLOBAL(_AND_POP)
ARM32: optimize handling of globals

* keep Pike_fp->current_storage in a register
* keep track of whether or not the current object
could have been destructed
* actually start using F_PRIVATE_GLOBAL

2016-09-05 21:23:37 (3 weeks ago) by Arne Goedeke <el@laramies.com>

Interpreter: set current_storage for pike functions

Currently, Pike_fp->current_storage was only initialized
for C functions. Initialize it for pike functions, aswell, to allow
the generated machine code to use it.

2016-09-05 21:09:19 (3 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: added missing argument to inc/dec opcodes

2016-09-05 12:51:48 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Cipher: Fixed max_bytes calculation for DES, DES3 and RC2.

Fixes multiple "Indexing the NULL value with "block_size".".

Sunday 04 September 2016

2016-09-04 21:05:52 (3 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: implement F_PRIVATE_GLOBAL

2016-09-04 15:56:13 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Use a pike array to hold the macro parts.

Further preprocessor cleanup.

Saturday 03 September 2016

2016-09-03 12:43:50 (3 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: added INC/DEC opcodes
ARM32: added the F_CALL_BUILTIN opcodes

It would be useful to know at assembler time, if a given efun has
a return value or not.

2016-09-03 11:58:38 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Build [hashtable]: hashtable.[ch] are no longer used.

2016-09-03 11:54:12 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Replace the struct hash_entry with a string.

The struct hash_entry is only used to hold a string, so get rid
of it and store the string directly instead.

Friday 02 September 2016

2016-09-02 12:18:42 (3 weeks ago) by Arne Goedeke <el@laramies.com>

Protocols.WebSocket: fixed handling of truncated frames

When small frames were not fully in the receive buffer, the frame
parsing could get confused.

2016-09-02 10:54:49 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Fixed indentation.

2016-09-02 10:52:08 (3 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Use the precompiler for the define class.

Thursday 01 September 2016

2016-09-01 20:26:28 (3 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: renamed load and store instruction API

This new API with explicit bit widths will make sharing code
between the two arm arches easier.

2016-09-01 14:41:42 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Renamed struct define to struct define_struct.

This is in preparation to using the precompiler for the define program.

2016-09-01 10:01:43 (4 weeks ago) by Arne Goedeke <el@laramies.com>

Math: improved spelling of documentation

2016-09-01 09:11:51 (4 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: F_NE should always return 0 or 1
Debug.get_program_layout: behave as documented

Wednesday 31 August 2016

2016-08-31 16:23:23 (4 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: fixed F_EQ/F_NE

The fast path (for everything but objects and functions) was broken
in these two opcodes.
ARM64: added F_OR_INT, F_AND_INT and F_XOR_INT
ARM64: implement F_RETURN_IF_TRUE
ARM64: fixed F_LOOP

We cannot rely on the loopcnt always being positive.
ARM64: implemented F_LOOP

2016-08-31 12:12:43 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Wrap struct define in objects.

The preprocessor macro lookup table is now a standard mapping instead of
a custom hashtable. This is a step in cleaning up the preprocessor code.

Tuesday 30 August 2016

2016-08-30 19:24:12 (4 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: fixed F_LOOP

We cannot rely on the loopcnt always being positive.

2016-08-30 11:09:22 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [cpp]: Removed some redundant declarations.

2016-08-30 09:30:37 (4 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: implemented F_LOOP

Monday 29 August 2016

2016-08-29 21:38:42 (4 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: added F_OR_INT, F_AND_INT and F_XOR_INT

2016-08-29 17:54:07 (4 weeks ago) by Arne Goedeke <el@laramies.com>

ARM32: implement F_RETURN_IF_TRUE
ARM32: fixed F_EQ/F_NE

The fast path (for everything but objects and functions) was broken
in these two opcodes.
ARM32: reordered arguments of lsl_reg_reg

It seems like lsl_reg_reg(dst, a, b) was actually dst = b << a.
This was not intended! These instructions were only used in
the F_EQ/F_NE opcodes where this "error" resulted in the slow
path being executed unconditionally. Also: the fast path is broken.

2016-08-29 17:42:24 (4 weeks ago) by Arne Goedeke <el@laramies.com>

random_string(): fixed warning

2016-08-29 16:15:43 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Constants [Documentation]: Some notes about ALPN.

Sunday 28 August 2016

2016-08-28 15:45:10 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Use sprintf %x some more.

2016-08-28 15:35:32 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

Allow sprintf %x on a string.

2016-08-28 13:10:19 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Cipher.CipherSpec: Fixed max_bytes for CIPHER_block.

Saturday 27 August 2016

2016-08-27 12:26:58 (4 weeks ago) by Martin Nilsson <nilsson@fastmail.com>

More precise return type for %X expressions.

2016-08-27 12:08:19 (4 weeks ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Documentation [Multiset]: Added minimal description.

Friday 26 August 2016

2016-08-26 21:41:35 (4 weeks ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Use tbz/tbnz instructions
ARM64: simplify some conditionals
ARM64: added comparison opcodes
ARM64: implement OPCODE_INLINE_BRANCH
ARM64: implement F_THIS_OBJECT and F_SIZEOF_LOCAL
ARM64: Use pre-decrement when popping from mark stack
ARM64: fix some warnings
ARM64: more debug prologues
ARM64: fix bug in slowpath of integer ops

The slowpath of the integer operations could sometimes change the register
state. This was incorrect because it is not always executed.
ARM64: implement fast paths for comparisons
ARM64: added low level support for shift instructions
ARM64: set return hint on function exits
ARM64: implement QUICK and comparison jumps
ARM64: introduced aliases for argument registers
ARM64: Add some custom register names
ARM64: simplify arm32_ins_maybe_exit
ARM64: Always return unconditional branches from ins_f_jump

This makes sure we get maximum avaiable range for future update_f_jumps.
ARM64: add F_ADD_INTS
ARM64: generate better statistics
ARM64: simplify prologue/epilogue generation
ARM64: use unsigned compare for pointers
ARM64: fixed svalue type in F_THIS_OBJECT
ARM64: better versions of F_STRING and F_CONSTANT
ARM64: added more CMP opcodes
ARM64: Use cbz/cbnz in more places
ARM64: added disassembler
ARM64: Initial commit
ARM64: implement F_ASSIGN_LOCAL
ARM64: correctly keep track of special registers

Since updating Pike_fp->pc would load the fp_reg, the fp_reg could
sometimes not be correctly invalidated after calling a opcode fun.
ARM64: fix several assert failure
ARM64: do some checks only during init
ARM64: Use zero register some more
ARM64: remove c++ style comments
ARM64: Fix loading of string constants with large indices
ARM64: really_free_svalue does not change fp or sp
ARM64: implement some F_RETURN_* opcodes
ARM64: implement F_POP_TO_MARK
ARM64: fix F_PROTECT_STACK

expendibles are stored as an offset in pike_frame now.

2016-08-26 17:07:39 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Cipher.CipherSpec: Added max_bytes.

This value is intended to be used as a threshold for automatic key
renegotiation (and make us proof against eg Sweet32).

cf https://sweet32.info/

Thursday 25 August 2016

2016-08-25 16:44:26 (1 month ago) by Arne Goedeke <el@laramies.com>

WhiteFish: silence warning

2016-08-25 12:48:20 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Debug: Added hexdump().

Wednesday 24 August 2016

2016-08-24 21:40:49 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: fix several assert failure

2016-08-24 21:27:23 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: better versions of F_STRING and F_CONSTANT

2016-08-24 12:49:17 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Compiler [CPP]: Fixed missing reference.

Potential fix for precompiler failing on some platforms.

Tuesday 23 August 2016

2016-08-23 12:33:37 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Regexp.PCRE: Hide `().

Hide `() so that modules that import Regexp.PCRE don't get a nasty
surprise when it shadows predef::`()...

Fixes testsuite failures for Parser.Markdown.

2016-08-23 10:59:45 (1 month ago) by Arne Goedeke <el@laramies.com>

ARM32: typo

2016-08-23 07:58:51 (1 month ago) by Arne Goedeke <el@laramies.com>

ARM32: fix several assert failure

Under some circumstances bad load/store instructions could be generated
for large offsets.
ARM32: better versions of F_STRING and F_CONSTANT

2016-08-23 00:22:55 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Use () instead of call_function.

Monday 22 August 2016

2016-08-22 20:47:48 (1 month ago) by unknown <unknown user@unknown>

release number bumped to 8.0.290 by export.pike
release number bumped to 8.0.291 by export.pike

2016-08-22 15:37:05 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.EXIF: Improved robustness.

Some EXIF files apparently terminate the attribute list chain with
an empty block with no next pointer at the end of the EXIF IFD.

Fixes [bug 7761].

2016-08-22 15:36:40 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Don't support infinite looping in exif data.
Allow for only specific tags to be parsed and the rest ignored.
Reuse the tag parsing loop.
Confine the EXIF parsing in its JFIF segment.
Converted if-train to switch.

2016-08-22 15:32:54 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

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

2016-08-22 14:56:05 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Standards.EXIF: Improved robustness.

Some EXIF files apparently terminate the attribute list chain with
an empty block with no next pointer at the end of the EXIF IFD.

Fixes [bug 7761].

Sunday 21 August 2016

2016-08-21 20:23:37 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Add some custom register names

2016-08-21 20:18:16 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Use pre-decrement when popping from mark stack

2016-08-21 20:04:46 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Use tbz/tbnz instructions

2016-08-21 19:32:24 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Use zero register some more

2016-08-21 18:25:03 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Use cbz/cbnz in more places

2016-08-21 18:22:30 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: Always return unconditional branches from ins_f_jump

This makes sure we get maximum avaiable range for future update_f_jumps.

2016-08-21 16:33:07 (1 month ago) by Marcus Comstedt <marcus@mc.pp.se>

ARM64: correctly keep track of special registers

Since updating Pike_fp->pc would load the fp_reg, the fp_reg could
sometimes not be correctly invalidated after calling a opcode fun.
ARM64: do some checks only during init
ARM64: set return hint on function exits
ARM64: added more CMP opcodes
ARM64: use unsigned compare for pointers
ARM64: implement some F_RETURN_* opcodes
ARM64: fix F_PROTECT_STACK

expendibles are stored as an offset in pike_frame now.
ARM64: generate better statistics
ARM64: add F_ADD_INTS
ARM64: implement fast paths for comparisons
ARM64: simplify some conditionals
ARM64: implement F_THIS_OBJECT and F_SIZEOF_LOCAL
ARM64: implement F_ASSIGN_LOCAL
ARM64: implement F_POP_TO_MARK
ARM64: added disassembler
ARM64: simplify arm32_ins_maybe_exit
ARM64: more debug prologues
ARM64: introduced aliases for argument registers
ARM64: really_free_svalue does not change fp or sp
ARM64: implement QUICK and comparison jumps
ARM64: fix bug in slowpath of integer ops

The slowpath of the integer operations could sometimes change the register
state. This was incorrect because it is not always executed.
ARM64: fixed svalue type in F_THIS_OBJECT
ARM64: added comparison opcodes
ARM64: fix some warnings
ARM64: implement OPCODE_INLINE_BRANCH
ARM64: remove c++ style comments
ARM64: Fix loading of string constants with large indices
ARM64: Initial commit
ARM64: simplify prologue/epilogue generation
ARM64: added low level support for shift instructions

2016-08-21 14:25:45 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Deprecate call_function as per Hubbes note from 1998.
call_function tests updated

2016-08-21 10:46:34 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Parser.Markdown: Fixed typo in regexp.

2016-08-21 04:01:50 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Don't look up - with findstring. Us lfun_strings.

2016-08-21 03:49:11 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

No need of calling findstring on __INIT. We have a reference in lfun_strings.

2016-08-21 03:29:58 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Just use pike_string for args in define magic functions, and get rid of non-generic code.

Saturday 20 August 2016

2016-08-20 13:08:25 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

binary_findstring_pcharp had one user, so simplify the arguments for both sides.
Type PCHARP a little harder.

2016-08-20 12:10:12 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

SSL.Constants: Added enum Extension.

Fixes some documentation references.

Friday 19 August 2016

2016-08-19 14:44:03 (1 month ago) by Anders Johansson <anders@roxen.com>

release number bumped to 8.0.289 by export.pike
release number bumped to 8.0.288 by export.pike

2016-08-19 11:23:08 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

MasterObject: describe_string() now shows some of the end of the string.
MasterObject: describe_string() now shows some of the end of the string.

2016-08-19 11:19:23 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

MasterObject: describe_string() now shows some of the end of the string.

Thursday 18 August 2016

2016-08-18 23:02:35 (1 month ago) by Arne Goedeke <el@laramies.com>

ARM32: fix bug in slowpath of integer ops

The slowpath of the integer operations could sometimes change the register
state. This was incorrect because it is not always executed.
ARM32: more debug prologues

2016-08-18 22:32:14 (1 month ago) by Martin Nilsson <nilsson@fastmail.com>

Improved hash_value debug.

2016-08-18 18:29:32 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite: Removed some debug.

2016-08-18 14:30:43 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

hash_value: Fixed truncation bug on 64-bit architectures.

The hash value was truncated to 32 bits.

Fixes testsuite failures on Solaris/sun4u, and most likely random
failures on other 64-bit architectures when the process has a large
memory footprint.

Wednesday 17 August 2016

2016-08-17 12:48:26 (1 month ago) by Henrik Grubbström (Grubba) <grubba@grubba.org>

Testsuite [hash_value]: Added some debug information.

[permalink]

Bugs mentioned

  7780RESOLVEDString optimization bug
  7783RESOLVEDCast to program is not thread-safe.
Bugs? Suggestions?