Activity
From 07/08/2014 to 07/14/2014
07/14/2014
-
08:10 PM Bug #10009: IO operation is 10x slower in multi-thread environment
- Good to know it works for you. Keep in mind TIME_QUANTUM_USEC=1000 is
very low and may cause problems on some systems, too.
My gut feeling is 100ms (default) is too high, but 10ms is too low
(based on kosaki's comment). Maybe 20m... -
01:31 PM Bug #9759: [TracePoint API] return event missing when raising exception
- Yes, maybe he is aware of this because he only asked for backport to the 2.1 branch. Can you help us, ko1?
-
12:21 PM Bug #10035 (Closed): Find.find no longer accepts Pathname type as argument
- In 2.1, a check for encoding of the paths parameter has been added to File.find. This works perfectly well for String parameters, but this results in a NoMethodError when a Pathname parameter is given. In my opinion a Pathname is a bette...
-
11:51 AM Bug #10027: make at windows , err : _rb_file_expand_path_internal already defined in file. obj
- Those functions are in msvcrt.lib.
Have you applied service packs? -
01:34 AM Bug #10027: make at windows , err : _rb_file_expand_path_internal already defined in file. obj
I have delete all source , and run configure.bat . have an other problem :
```
G:\dev-tools\ruby-2.1.2>win32\configure.bat --disable-install-doc
Creating Makefile.new
"type `nmake' to make ruby."
G:\dev-tools\ruby-2.1.2>nmak...-
- Andrii Tereshchenko wrote:
> kk kk wrote:
> ...
thank you . how to clean all config files ? -
- * win32/Makefile.sub (LIBS): oldnames.lib is not used already.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46817 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
10:17 AM Revision c3000ed8 (git): * ext/win32ole/win32ole.c: modify WIN32OLE class document and
- add comment for constants of WIN32OLE.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46816 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
10:02 AM Feature #8895: Destructuring Assignment for Hash
- Sean Linsley wrote:
> I don't follow. Can't this assignment behave the same way that method argument destructuring does?
I agree. Destructuring should work as method & block passing works (apart from block passing) - 09:48 AM Revision f747a231 (git): ChangeLog: document size changes [ci skip]
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46815 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- It may need further consideration whether `instance_exec` should affect the wrapped proc.
-
- Reducing struct re_pattern_buffer from 464 to 448 bytes allows it to fit
into neatly into seven 64-byte cache lines on x86-64 systems.
ABI is changed in 2.2.0dev anyways, so we may do this.
Will commit in a few days.
C extension us... - 07:42 AM Revision ff9046ad (git): vm_core.h (struct rb_iseq_struct): reduce to 288 bytes on 64-bit
- * vm_core.h (struct rb_iseq_struct): reduce to 288 bytes on 64-bit
(found with pahole, from the dwarves package)
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46814 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
07:22 AM Feature #10030: [PATCH] reduce rb_iseq_struct to 296 bytes
- (2014/07/14 16:12), Eric Wong wrote:
> OK, r46811 (I also made trivial commit r46813 to shrink catch entry)
Thanks.
> What should we do about rb_iseq_build_for_ruby2cext? I think it is
> broken, now (as is anything else which m... -
- ko1@atdot.net wrote:
> Thank you.
> I will modify style after your commit.
OK, r46811 (I also made trivial commit r46813 to shrink catch entry)
What should we do about rb_iseq_build_for_ruby2cext? I think it is
broken, now (as... -
- Thank you.
I will modify style after your commit.
-
- Changes from v2:
- iseq_catch_table_size removed, use if (...) for (;...;)
I used a short local variable in some places to keep the "if (...) for (...)"
sequences from wrapping on long lines.
-
- Eric Wong wrote:
> Eric Wong <normalperson@yhbt.net> wrote:
> ...
There are two reasons I'm opposed to:
(1) I think writing "if (...) for (...)" is more clear (easy to understand).
I don't care to adding one line to check exisiting... -
- Updated patch
Changes from v1:
- renamed iseq->_catch_table to iseq->catch_table
- iseq_catch_table_bytes: made a static inline function
- iseq_catch_table_size: new function replaces the
iseq_catch_table_each iterator macro
- 07:08 AM Revision 34f7e908 (git): iseq.h (struct iseq_catch_table_entry): shrink to 32 bytes on 64-bit
- * iseq.h (struct iseq_catch_table_entry): shrink to 32 bytes on 64-bit
This drops a few minor page faults on my system at startup.
Using 64-bit start, end, cont, and sp fields are unnecessary as
label_get_position and label_get_sp only ... - 07:06 AM Revision b811194e (git): * 2014-07-14
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46812 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 07:06 AM Revision e91ae784 (git): vm_core.h (struct rb_iseq_struct): reduce to 296 bytes on 64-bit
- Most iseq do not have a catch_table, so avoid needlessly adding
4-8 bytes to the struct for the common case.
Changes from v2:
- iseq_catch_table_size removed, use if (...) for (;...;)
Changes from v1:
- renamed iseq->_catch_table to is... -
- Nobuyoshi Nakada wrote:
> How did you get the source, and what is the exact revision?
I delete all source, and run `nmake array.obj -n` ,the result is same to you. fixed.
( I think the reason is I run configure.bat at win32 dir , ... -
- Setting SO_KEEPALIVE on the socket will help, but that still takes at
least 2 hours to detect on a stock Linux system. Do you want to set a
shorter keepalive, or is ~2 hours OK?
You can change the interval using OS-specific knobs ...
07/13/2014
-
11:54 PM Misc #10032: Matrix classについて
- Marc-Andre Lafortune wrote:
> Since this request talks about Matrix, of which I'm the maintainer, and refers to a PR I reviewed, I'd very much like an english translation.
In short, he is interested in improving Matrix class. And to... -
11:00 PM Misc #10032: Matrix classについて
- >こちらのPRなど最初のレビューを頂くまでに一ヶ月半を要しました. 問題だと感じました.
対面でもお伝えしましたが、 github は補助的なものなので、レビューを要するものは
オフィシャルの issue tracker である redmine に最初にパッチを投稿することをお勧めします。
また、Ruby の標準添付ライブラリのメンテナンスはライブラリごとにメンテナが
担当することになっています。
つまり、パッチをどうするかはメンテナ次第です。まず最初... -
- Since this request talks about Matrix, of which I'm the maintainer, and refers to a PR I reviewed, I'd very much like an english translation.
Thanks -
07:47 PM Misc #10032 (Closed): Matrix classについて
- こんにちは. 突然のご連絡失礼します.
もしこちらでお伺いするには不適切な内容であれば、お叱りと共にその旨をお伝え頂ければなと思います.
* 以下本文
Matrixクラスの貧弱さに憂いてGithub上からシコシコとPRを投げているのですが、
https://github.com/ruby/ruby/pull/594
こちらのPRなど最初のレビューを頂くまでに一ヶ月半を要しました. 問題だと感じました.
こちらは単純にGithubの扱... -
11:05 PM Feature #8895: Destructuring Assignment for Hash
- Koichi Sasada wrote:
> Problem is what happen when `h' is not a hash object (and doesn't have to_hash method).
> ...
I don't follow. Can't this assignment behave the same way that method argument destructuring does? This currently work... -
07:35 PM Bug #10031 (Closed): Net::IMAP idle can still block a thread forever.
- When calling Net::IMAP idle often errors will go unnoticed forever, simply leaving the Thread that called "idle" blocking forever.
For example try bringing your internet connection down whilst a thread is calling the idle method, the ... -
02:35 PM Bug #8523: intermittent unit test failure in test_timeout.rb results in build failures
- r44516, r46159 and r46196 were backported into `ruby_2_1` branch at r46810.
-
- test_timeout.rb: shorten waiting times
* test/test_timeout.rb (test_timeout): inverted test condition.
[Bug #8523]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@46810 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Backported into `ruby_2_1` branch at r46809.
-
- * ext/digest/digest.c (rb_digest_instance_equal):
fix #== for non-string arguments. [ruby-core:62967] [Bug #9913]
* test/digest/test_digest.rb: add test for above.
* ext/digest/digest.c (rb_digest_instance_equ... -
- r46417 and r46418 were backported into `ruby_2_1` branch at r46808.
-
- * array.c (yield_indexed_values): extract from permute0(),
rpermute0(), and rcombinate0().
* array.c (rb_ary_combination): iterate on a shared copy, and use
array of indexes instead of array of chosen objects... -
- Backported into `ruby_2_1` branch at r46807.
-
- * array.c (rb_ary_permutation): `p` is the array of size `r`, as
commented at permute0(). since `n >= r` here, buffer overflow
never happened, just reduce unnecessary allocation though.
git-svn-id: svn+ssh://ci.rub... -
- Backported into `ruby_2_1` branch at r46806.
-
- * pack.c (encodes): fix buffer overrun by tail_lf. Thanks to
Mamoru Tasaka and Tomas Hoger. [ruby-core:63604] [Bug #10019]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@46806 b2dd03c8-39d4-4d8f-98ff-823fe69b... -
- Backported into `ruby_2_1` branch at r46805.
-
- * ext/thread/thread.c (undumpable): ConditionVariable and Queue
are not dumpable. [ruby-core:61677] [Bug #9674]
* marshal.c (w_object): internal objects are not dumpable.
[ruby-core:61677] [Bug #9674]
git-... -
10:38 AM Bug #10011: Passing a string to Pathname#relative_path_from results in NoMethodError
- https://github.com/ruby/ruby/pull/666
-
- spell out `encryption` by @vipulnsward [fix GH-664]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46804 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * ext/sdbm/init.c: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46803 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Eric Wong <normalperson@yhbt.net> wrote:
> How about using a `iseq_catch_table_size` macro instead:
Or static inline function (also for `iseq_catch_table_bytes`). -
- ko1@atdot.net wrote:
> From what size on which architecture?
x86-64
> On the code:
>
> (1) `_catch_table` should be `catch_table`
OK, I will change it back. I renamed it so it'd be easier for the
compiler to det... -
- From what size on which architecture?
On the code:
1. `_catch_table` should be `catch_table`
2. I don't like `iseq_catch_table_each()` macro.
I like to write like
```C
if (iseq->_catch_table) for (i=0; i<iseq->_catch_table-... -
- This probably breaks ruby2cext; but I'm not sure if anybody uses that.
May we remove rb_iseq_build_for_ruby2cext?
I will probably have more patches along these lines to reduce iseq-related
allocations. This is probably the most o... -
- * ext/sdbm/init.c: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46802 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- [fix GH-655]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46801 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 07:46 AM Revision c7fbc296 (git): vm_core.h (struct rb_call_info_struct): improve packing
- * vm_core.h (struct rb_call_info_struct): improve packing
On x86-64, it goes from 112 to 104 bytes. This results in
~20K of savings from just "./ruby -e exit"
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46800 b2dd03c8-39d4-4d... -
- [fix GH-659]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46799 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 07:01 AM Revision 499f6d62 (git): vm_core.h (struct rb_iseq_struct): stack_max is uint32_t
- * vm_core.h (struct rb_iseq_struct): stack_max is uint32_t
No program will ever need more than a few megabytes of stack,
so there's no sense in using a 64-bit counter for stack
accounting. Packing this with the 32-bit type enum re... -
- Applied in changeset r46796.
----------
configure.in: exit with failure
* configure.in (rb_cv_broken_backtrace): exit with failure
normally, no needs to abort. [ruby-core:63678] [Bug #10008] -
-
- It makes no sense.
- 01:56 AM Revision da16701b (git): * 2014-07-13
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46797 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * configure.in (rb_cv_broken_backtrace): exit with failure
normally, no needs to abort. [ruby-core:63678] [Bug #10008]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46796 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
07/12/2014
-
11:37 PM Bug #10008: conftest crashed: installing mri 2.1.2: EXC_BAD_ACCESS (SIGABRT)
- If I replace the whole test for "broken backtrace" with a hardcoded value, then I can run `./configure` without a crash.
See my proof of concept patch (attached) which can be compared to tag v2_1_2 in git. -
- > Which conftest?
I think conftest is defined in the `configure` script. For me, it's around line 20376, which looks like:
cat confdefs.h - <<_ACEOF >conftest.$ac_ext
Furthermore, line 20389 defines a `sigsegv` function.
... -
- Simpler steps to reproduce:
1. Download ruby 2.1.2 source
2. ./configure
So, I don't think this has anything to do with ruby-build :) -
- Steps to reproduce:
1. rbenv uninstall 2.1.2
2. rbenv install 2.1.2
Crash dump:
https://gist.github.com/jaredbeck/127e61aa73a16c614b6d
ruby_build log:
https://gist.github.com/jaredbeck/3f6de8e3e432a51ae707
versions:
gcc... -
- * win32/configure.bat: bail out if run in win32 directory.
[ruby-core:63648] [Bug #10027]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46795 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
-
02:05 PM Bug #10027: make at windows , err : _rb_file_expand_path_internal already defined in file. obj
- kk kk wrote:
> G:\ruby-2.1.2\win32>nmake
You should follow README.win32 and configure and nmake from parent folder, not from win32.
I.e. inside G:\ruby-2.1.2 call win32\configure.bat and then do nmake from same source code root. N... -
- fields Hash key.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46794 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
05:11 AM Bug #9985 (Closed): Gems documenation is not generated when installing local gem with --ignore-dependencies option
07/11/2014
-
10:31 PM Bug #10009: IO operation is 10x slower in multi-thread environment
- Alexandre Riveira wrote:
> Eric Wong wrote:
> ...
Tests completes, my system without changes join stress tests 30 secons for load page, after changes, pages loading in instant all pages loading in less than 1 second.
-
- Eric Wong wrote:
>
> ...
Hello Eric!!!!
I stayed enjoyed the result of change TIME_QUANTUM_USEC. Changed its value to 1000 only see the results:
ruby 2.
first 17434583
second 2754 <=============
third 16752441
If you have ... -
07:22 PM Revision 00eb7203 (git): * lib/net/smtp.rb (Net::SMTP#data): enable buffering while
- 'data' send for optimizing Net::SMTP#send_message.
[ruby-dev:48329] [misc #9981]
patch by Masahiro Tomita.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46793 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- コードは問題なさそうに見えるし、賛成意見もついているし、メンテナがいないモジュールのようなので、
代理コミットしときますね。 -
02:22 PM Feature #9981: Net::SMTP#send_message が大量の write(2) を発行する
- In message <redmine.journal-47699.20140711064732.e1bfc429d603f917@ruby-lang.org>
on Fri, 11 Jul 2014 06:47:32 +0000,
shibata.hiroshi@gmail.com wrote:
> Issue #9981 has been updated by Hiroshi SHIBATA.
>
>
> どなたか SMTP にお詳しい方はいら... -
- どなたか SMTP にお詳しい方はいらっしゃいませんか.
-
04:42 PM Bug #10020: TestEnv#test_memory_leak_*, Fiddle::TestPointer#test_no_memory_leak, and Test_StringModifyExpand#test_modify_expand_memory_leak on Solaris
- r46791 にて、Solaris 9 以降で、以下を満たす場合に、それらの環境変数をセットして invoke_ruby するように変更しました。
* 環境変数 LD_PRELOAD, LD_PRELOAD_64 (または LD_PRELOAD_32), UMEM_OPTIONS がセットされていない
* LD_PRELOAD=libumem.so UMEM_OPTIONS=backend=mmap を付けてrubyを実行して正常終了
* テストケース独自に、... -
- Applied in changeset r46791.
----------
* test/ruby/envutil.rb (assert_no_memory_leak): On Solaris 9 or later,
if possible, execute child ruby with environment variables
LD_PRELOAD=libumem.so UMEM_OPTIONS="backend=mmap". With these
... -
- TestEnv#test_memory_leak_* に関しては、
Illumos(旧OpenSolaris）のソース usr/src/lib/libc/port/gen/getenv.c を見る限り、
setenv() 内で malloc にて確保した領域を unsetenv() で free する事は無いように見えます。
つまり、少なくとも Solaris 10 / Illumos では、setenv(3C) と unsetenv(3C) を繰り返すと（mal... - 04:32 PM Revision f11c30e8 (git): * 2014-07-12
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46792 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- if possible, execute child ruby with environment variables
LD_PRELOAD=libumem.so UMEM_OPTIONS="backend=mmap". With these
variables, freed memory is immediately returned to the OS.
[Bug #10020] [ruby-dev:48391]
git-svn-id: svn+ssh... -
- * symbol.c (op_tbl): remove non-regular symbols.
* symbol.c (global_symbols): start from the next of the preserved
ID.
* symbol.c: (rb_id2str): op_tbl does not exceed tLAST_OP_ID.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46... -
- backported at r46789.
-
- Applied in changeset ruby-21:r46789.
----------
merge revision(s) r46780: [Backport #9499]
* lib/matrix.rb: Fix sign for cross_product [#9499] -
- * lib/matrix.rb: Fix sign for cross_product [#9499]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@46789 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
01:58 PM Revision f063a0b5 (git): fix plural form [ci skip]
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46788 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * parse.y (rb_init_parse): rename to get rid of name clash with
json/parser at statically linking.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46787 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 11:51 AM Revision 803232c8 (git): * remove trailing spaces.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46786 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- VT_RECORD OLE variables.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46785 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- How did you get the source, and what is the exact revision?
-
- ```
G:\ruby-2.1.2>nmake array.obj -n
Microsoft (R) Program Maintenance Utility Version 6.00.8168.0
Copyright (C) Microsoft Corp 1988-1998. All rights reserved.
echo compiling array.c
cl -nologo -MD -Zi -W2 -O2... -
08:50 AM Bug #10029: miss probes.h when nmake
- The version 6.00.8168.0 of nmake.exe is a part of Visual Studio 6.0.
It seems that ruby-2.1.2 cannot be built with Visual C++ 6.0.
-
08:31 AM Bug #10029: miss probes.h when nmake
- Nobuyoshi Nakada wrote:
> I can't reproduce your environment.
Me too.
-
- I can't reproduce your environment.
probes.h should be copied from probes.dmyh.
~~~
> nmake array.obj -n
echo copying dummy probes.h
if exist probes.dmyh ( @ copy > nul probes.dmyh probes.h ) else ( @ copy > nul ..... -
- ```
G:\ruby-2.1.2>nmake
Microsoft (R) Program Maintenance Utility Version 6.00.8168.0
Copyright (C) Microsoft Corp 1988-1998. All rights reserved.
CC = cl -nologo
LD = cl -nologo
LDSHARED = cl -nologo ... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46784 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- ネストしたrest argumentは通りますが
~~~ruby
def f((*a)) end
~~~
ネストしたrest keyword argumentはsyntax errorになります。
~~~ruby
def f((**a)) end #=> syntax error, unexpected **arg
~~~ -
- I can't reproduce it.
Could you try `nmake V=1`? -
- ```
G:\ruby-2.1.2\win32>nmake
Microsoft (R) Program Maintenance Utility Version 6.00.8168.0
Copyright (C) Microsoft Corp 1988-1998. All rights reserved.
CC = cl -nologo
LD = cl -nologo
LDSHARED = cl -n... -
- Applied in changeset r46783.
----------
* lib/fileutils.rb: handle ENOENT error with symlink targeted to
non-exists file. [ruby-dev:45933] [Bug #6716] -
- non-exists file. [ruby-dev:45933] [Bug #6716]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46783 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
-
- Applied in changeset r46782.
----------
* array.c: Clarify documentation for Array#insert.
[ruby-core:62934] [Bug #9901] -
- [ruby-core:62934] [Bug #9901]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46782 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
-
-
-
- Applied in changeset r46781.
----------
* io.c: Improve Documentation by @dapplebeforedawn.
[fix GH-658] [ruby-core:63579] [Bug #10012] -
- [fix GH-658] [ruby-core:63579] [Bug #10012]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46781 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 05:19 AM Revision 72d05366 (git): * lib/matrix.rb: Fix sign for cross_product [#9499]
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46780 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * benchmark/prepare_so_reverse_complement.rb: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46779 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Probably `malloc_usable_size` would access the pointer - 8, that means `ruby_xfree` was called with `(void *)1`.
Does this happen with the trunk and recent 2.1 snapshots as the target version is set to `current: 2.2.0`? -
- Applied in changeset r46778.
----------
pack.c: fix buffer overrun
* pack.c (encodes): fix buffer overrun by tail_lf. Thanks to
Mamoru Tasaka and Tomas Hoger. [ruby-core:63604] [Bug #10019] -
- * pack.c (encodes): fix buffer overrun by tail_lf. Thanks to
Mamoru Tasaka and Tomas Hoger. [ruby-core:63604] [Bug #10019]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46778 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- OK I will update to TypeError - the method argument does need to support the Pathname interface.
As for the subclass - yes that is mostly it. It's a small detail.
Thanx for the feedback.
07/10/2014
-
- eberner@gmail.com wrote:
> Here is the Other runtime information
>
> https://gist.github.com/maletor/df43091aca5e9b7ee534
URL seems invalid.
> -- C level backtrace information -------------------------------------------
> /ho... -
11:36 PM Bug #10026: Segmentation fault
- Here is the Other runtime information
https://gist.github.com/maletor/df43091aca5e9b7ee534 -
- The following segmentation fault is from running tests with RSpec on a somewhat large test suite. The results of the test are always printed *before* the segmentation fault.
~~~
/home/ubuntu/doximity/vendor/bundle/ruby/2.1.0/bin/rspe... -
11:48 PM Bug #10011: Passing a string to Pathname#relative_path_from results in NoMethodError - At the end of the day, I just want something more friendly than a `NoMethodError` from inside the method, so I won't stand in the way of progress, but...
>1. I didn't want to call Pathname::new if we already had been passed a Pathname... -
- I agree with the confusion - initially I thought :to_path could be used, but despite its name it returns a string, hence the suggestion for :to_pathname, that allows things to implement a way to have themselves converted to a Pathname.
... - 03:25 PM Bug #10011: Passing a string to Pathname#relative_path_from results in NoMethodError
- I don't think we should add a method to String; there is an existing path coercion protocol (`to_path` and `File.path`) and even though String does not implement `to_path`, I think having a `to_pathname` would be confusing.
Anyway, co... -
- ok so I have 2 possible fixes, one of which I've implemented
The unimplemented fix is maybe a new method along the lines of to_s, but for paths, called to_pathname. Then we could reopen String and define
clas String
def to_pathna... - 08:46 PM Bug #7707: FileUtils.mv fails to move symlink between filesystems
- In Homebrew, we have code that works around this by shelling out to `mv` when necessary. It would be nice to have this fixed.
-
06:45 PM Bug #10025 (Closed): Incorrect wrapping of base64 output of Array.pack()
- String format directive `m` for Array `pack()` is documented as:
~~~
m | String | base64 encoded string (see RFC 2045, count is width)
| | (if count is 0, no line feed are added, see RFC 4648)
~~~
... -
- Nobuyoshi Nakada wrote:
> But if `tail_lf` is 1, `len` is a multiple of 3, so it can't be 3070.
`len` in `encodes()` can be anything between 1 and `len` from `pack_pack()` (which is a multiple of 3). It is possible to trigger mentio... -
- Am I missing something?
-
- But if `tail_lf` is 1, `len` is a multiple of 3, so it can't be 3070.
-
09:28 AM Bug #10019: segmentation fault/buffer overrun in pack.c (encodes)
- ... and type is not 'u'.
-
- Looking at encodes(), it seems at least when
* len = 3070 (= ((4096 - 4) / 4 * 3 + 1))
* tail_lf = 1
encodes() causes segv at https://github.com/ruby/ruby/blob/trunk/pack.c#L987 -
- `<<` and `>>` are shift operators primarily, and have higher precedence.
-
04:28 PM Bug #10024: Array injection of ternary operators injects boolean rather than ternary result
- The `#<<` has a higher precedence than the ternery operator. Using #<<= seems to make it work just fine.
-
03:40 PM Bug #10024 (Rejected): Array injection of ternary operators injects boolean rather than ternary result
- The following code:
```ruby
a = []
a << ''.blank? ? 'blank' : 'not blank'
puts a
```
renders `[true]`.
I would expect it to render `['blank']`
If I wrap it in parenthesis it does work.
```ruby
a = []
a << ( ''.blank?... - 03:22 PM Revision 70451a56 (git): * 2014-07-11
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46777 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- When name contains '=', ruby_setenv raises Errno::EINVAL.
That is the same behavior as Solaris 10.
NULL check for malloc return value is also added.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46776 b2dd03c8-39d4-4d8f-98ff-82... -
- Not a bug.
`A[:b]` and `c[:b]` refer the same object.
Check its `object_id`. -
01:44 PM Bug #10023: Hash#merge fails duplicating objects
- Please, note that only Rails' `deep_dup` seems helpful to workaround the issue
-
- The same with duplication
>> c=A.dup #same .clone
=> {:b=>{}}
> ...
=> {:b=>{}, :c=>{:d=>"e"}}
>> c[:b][:d] = 'bug'
=> "bug"
> ...
=> {:b=>{:d=>"bug"}}
-
- Please, also note that Rails' `deep_merge` does not help
-
- Here is an example in IRB (ruby 2.1.2):
>> A={b:{}}
=> {:b=>{}}
> ...
=> {:b=>{}, :c=>{:d=>"e"}}
>> c[:b][:d] = "f"
=> "f"
> ...
=> {:b=>{:d=>"f"}}
Every key, value pair is expected to be duplicated during the merge? Please, inv... -
10:55 AM Bug #9985: Gems documenation is not generated when installing local gem with --ignore-dependencies option
- Fixed: https://github.com/rubygems/rubygems/commit/c16846bf9e0b90521029f0d743620f138ee2236e
-
- +1 for this proposal.
I feel it is fine for me:
```ruby
k1: 1, k2: 2 = h
kr1:, kr2: = h
#=> same as
k1 = h.fetch(:k1, 1)
k2 = h.fetch(:k2, 2)
kr1 = h.fetch(:k1)
kr2 = h.fetch(:k2)
# mixed
k1: 1, k2: 2... -
- Alexandre Riveira wrote:
> information that I consider important
> ...
My results kernel linux BFS/CK + taskset
first 103214331
second 2762 <======
third 24259986
-
- Applied in changeset r46775.
----------
vm_insnhelper.c: fix implicit splat keyword argument
* vm_insnhelper.c (vm_callee_setup_keyword_arg): adjust VM stack
pointer to get rid of overwriting splat arguments by arguments
for `to_ha... -
- * vm_insnhelper.c (vm_callee_setup_keyword_arg): adjust VM stack
pointer to get rid of overwriting splat arguments by arguments
for `to_hash` conversion. [ruby-core:63593] [Bug #10016]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/tr... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46774 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * gc.c (global_list): fix typo, capital 'L'.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46773 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- And method transplanting only works when the method does not call super nor other Enumerable methods.
Automatic resolving is possible, but it would make inclusion far more complex.
Matz.
-
- The second argument is evaluated **before** calling `gsub` method.
Call it with a block:
~~~ruby
puts str.gsub(/(CFBundleVersion<\/key>\n\t.*\.).*(<\/string>)/) {"#{$1}#{ver}#{$2}"}
~~~ -
12:39 AM Bug #10022 (Rejected): Inconsistent behavior of gsub replacement
- In the following code, the two gsub's yield different results. ([[https://gist.github.com/franklsf95/6c0f8938f28706b5644d]])
~~~ruby
ver = 9999
str = "\t<key>CFBundleDevelopmentRegion</key>\n\t<string>en</string>\n\t<key>CFBundleVer...
07/09/2014
-
08:38 PM Bug #10021: OptParse does not handle missing arguments bracketed by valid option switches
- When I try this:
~~~
--re-run x -r --rerun y -- file
~~~
Then I see this
~~~
:in `eval': Only one output filename is permitted (RuntimeError)
~~~
From which I infer that -r has 'eaten' --rerun as its argument and... -
- OS=CentOS-6.5 x86_64
Ruby=2.1.2p95
I have this in an optparse structure:
# Create parser
opts = OptionParser.new do |opts|
opts.banner = <<-BANNER
. . .
opts.on("-r", "--rerun=DATEorTRANSACTION", "-... -
07:28 PM Feature #10017: Add `Hash#fetch_values`
- Please consider the name `fetch_at` carefully: it seems easily confused with `fetch` (1 key), and the name doesn't hint that it's for multiple keys. (`values_at` is plural, so it gives a good clue.) Maybe `fetch_values`?
-
07:00 PM Feature #10017: Add `Hash#fetch_values`
- I attached another patch calling this method fetch_at and I also added missing tests (I couldn't change issue title & body to reflect this changes, though).
Also, perhaps out of scope for this ticket but I'm wondering what're your tho... -
- +1 for `fetch_at`.
-
04:04 AM Feature #10017: Add `Hash#fetch_values`
- `fetch_at` sounds much better than values_at! to me.
-
- Btw, you're already paying significant costs for a case-insensitive FS
(not just Ruby, but things like git, too). -
04:05 PM Bug #10015: Performance regression in Dir#[]
- Thanks nobu, I really appreciate it! If it can't be fixed, I have ideas for avoiding the calls to Dir#[], but it will take time to get merged and released.
-
- It might affect unintentionally.
I'll check the bottleneck. -
- @nobu this change causes a significant increase in Rails boot time on my system.
My application pays a 120ms price for this change. @hsbt will pay 520ms:
http://twitter.com/hsbt/status/486378979138367488
Is there a way we can... -
- Yes.
Or use case-sensitive platforms. -
- `encodes()` fills and flush `buff` gradually by the buffer size.
Could you show the message at the segmentation fault?
Anyway, your patch includes unnecessary garbage. -
02:40 PM Bug #10019 (Closed): segmentation fault/buffer overrun in pack.c (encodes)
- While working with an AWS sample I hit a segmentation fault. The same sample works under 1.9.3. It appeared to be coming from pack.c function encodes. After looking at the source there's a 4K buffer allocated on the stack. I made a m...
-
- fstrings refered by static symbols and pinned dynamic symbols
are registerd by rb_gc_register_mark_object().
frstring refered by dynamic symbols (not pinned symbols)
are refered from global_symbols.dsymbol_fstr_hash (Hash object).
... - 03:29 PM Revision 996ff907 (git): * 2014-07-10
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46771 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- introduced in r46768. Object files containing dtrace probes should
be listed in DTRACE_DEPENDENT_OBJS.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46770 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Solaris 10 にて、以下のように memory leak 検出テストの Failure が出ます。（r46762 で確認）
Solarisデフォルトのlibcのmallocでは、freeしてもOSにメモリを返さないのが原因と考えています。
メモリ関係デバッグ用のライブラリ libumem を使い、UMEM_OPTIONS=backend=mmap を指定すると、freeしたら直ちにOSにメモリを返すようになるようです。
dynamic link な... -
- * random.c (fill_random_seed): fix type to cast. this may or may
not suppress warnings by icc.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46769 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- You can achieve it by "method transplanting".
~~~ruby
class Foo
define_method(:any?, Enumerable.instance_method(:any?))
end
~~~
Wrapper method like `Module#mix` may help you.
http://www.slideshare.net/yukihiro_matz/rubyconf-... -
11:49 AM Feature #10018: Consider adding Sub-Includes as in include Foo::bar
- Oops sorry!
This is not a bug, it is a feature request.
I am not sure how to move it to features now, sorry. :( -
- On IRC someone asked a nice question:
> \<arup_r\> If I write `class Foo; include Enumerable; end` This code include all methods from `Enumerable`. But suppose I want to include `#all?`, `#any?` .. What's the way to include only req... -
- r46762 fix this issue
-
- Elements expanded on the VM stack are overwritten by calling `to_hash` method.
-
- * symbol.c, symbol.h: Symbol class implementation and internals,
split from parse.y.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46768 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * parse.y (symbols_i): adjust argument types for st_foreach()
using st_data_t.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46767 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * io.c (do_io_advise): more precise error message, with arguments.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46766 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * test/ruby/test_io.rb (test_advise): fix inverted arguments order
by assert_nil.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46765 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Applied in changeset r46764.
----------
* parse.y: change Symbol <-> ID relationship to avoid
exposing IDs from collectable symbols.
[Bug #10014]
Now, rb_check_id() returns 0 if corresponding symbol is
pinned dynamic symbol.
T... -
01:46 AM Bug #10014: Do not expose no-pinned down IDs
- パッチ作成と内容説明ありがとうございます。賛成します。
-
- exposing IDs from collectable symbols.
[Bug #10014]
Now, rb_check_id() returns 0 if corresponding symbol is
pinned dynamic symbol.
There is remaining intern_cstr_without_pindown(), it can return
IDs from collectable symbols. We... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46763 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
02:31 AM Revision 9db5a7f8 (git): require "rbconfig" to use RbConfig
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46762 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- information that I consider important
Kernels BFS and ruby 1.9.2 work fine as if applied taskset.
Other kernels like freebsd and macos with ruby 1.9.2 has similar behavior.
http://en.wikipedia.org/wiki/Brain_Fuck_Scheduler
https:/... -
- My application is not web-site is an ERP. So reporting and very heavy tasks are performed. Then the system crashes because only one thread using 100% cpu will damage the whole worker, passing any request for at least 2 seconds, then the ...
07/08/2014
-
- Updated patch (from testing for #10009).
Uploading for archival purposes. This version is probably less intrusive and
falls back to pipe in case of ENOSYS (in case glibc supports eventfd and the
kernel has eventfd disabled).
Thi... -
- I'm proposing to add a new method `Hash#values_at!` that's basically a combination of `#values_at` and `#fetch`.
When dealing with configuration I prefer to use `#fetch` to fail early and this would be useful to read multiple keys at ... -
09:55 PM Bug #10016 (Closed): Destructuring block arguments with a Hashable last element
- The following code:
~~~ruby
# destructure_bug.rb
def m
x = Object.new
def x.to_hash() {x: 9} end
yield [1, 2, 3, x]
end
p m { |a, *b, c, **k| [a, b, c, k] }
~~~
gives:
~~~
$ ruby -v destructure_bug.rb
ruby 2.1.... - 08:50 PM Revision b10058ae (git): thread_pthread.h: do not expose pthread type for lock
- * thread_pthread.h (struct rb_global_vm_lock_struct):
do not expose pthread type for lock
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46761 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 08:47 PM Revision 3e14cdc4 (git): thread_pthread.h: remove unneeded semaphore.h include
- * thread_pthread.h: remove unneeded semaphore.h include
We currently do not use POSIX semaphores.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46760 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- eventfd doesn't help performance (but still reduces FD count),
I never expected eventfd to improve speed, though.
Lowering TIME_QUANTUM_USEC (in thread_pthread.c) helps with the I/O case
(try it yourself if you have a 1000HZ kernel)... -
- Eric Wong wrote:
> I'll try resurrecting an old eventfd proposal and maybe also bare futexes
> ...
Tank's Eric,
If an application running Rainbows has only one thread using 100% the worker is affected greatly in the query database. ... -
- I'll try resurrecting an old eventfd proposal and maybe also bare futexes
to see if that improves things. -
- r44802 seems to have introduced a performance regression in Dir#[].
Here is the test program:
~~~ruby
require 'benchmark'
puts Benchmark.realtime {
glob = "minitest/*_plugin.rb{,.rb,.bundle}"
$LOAD_PATH.map { |load_path|
... -
03:32 PM Bug #9975: irb で Hash を 1.9 記法で記述したときに、キーがキーワードだと syntax error
- けいじゅ＠いしつかです.
takkanm@gmail.com wrote:
> Issue #9975 has been reported by 三村 益隆.
> irb で Hash を 1.9 記法で記述したとき、以下の end を伴なうキーワードをキーとして使用するとエラーとなります。
>
> ~~~
> if do case def for unless until while begin
確かにバグですが...
> 添付... - 03:19 PM Revision a402e84e (git): * 2014-07-09
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46759 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46758 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46757 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
12:12 PM Feature #9997: Specifying non volatile registers for increase performence in ppc64
- Hi Nobuyoshi, yes, my mistake. The correct results are the ones I added later.
I'm attaching the results for the "Computer Language Benchmarks Game" as well. Check the time enhancement. -
- r46756 の make test-all にて、同じ Failure がすべての test/ruby/memory_status.rb 関連の場所で出たので再オープンします。
invoke_ruby によって --diable-gems を付けて呼んだ子プロセスの ruby 内で実行される envutil.rb にて、 require 'rbconfig' より前に find_executable が実行されるため、エラーになっているように見えます。（親プロセ... -
- 以下のように test/ruby/test_env.rb を実行すると、なぜか、
uninitialized constant EnvUtil::RbConfig (NameError) になります。
実行に使ったrubyはr46751で、svnレポジトリのローカルコピーr46756 の test/ruby/test_env.rb を相対パス指定にて直接実行した結果のものです。
~~~
$ ruby test/ruby/test_env.rb
Run ... -
- Applied in changeset r46755.
----------
memory_status.rb: find proper ps
* test/ruby/memory_status.rb (Memory::PSCMD): use ps command which
outputs expected result. [ruby-dev:48370] [Bug #10010] -
- 概要
----
ID と Symbol の関係を整理し、よりバグの少ないコードへと変更する。
現状と問題点
------------
Symbol GC が導入され、Symbol に dynamic symbol と static symbol に別れる
ことになった。
pindown されていない dynamic symbol は、GC 可能であり、ここでは mortal
symbol と呼ぶことにする。
逆に、static s... -
- * sprintf.c (rb_str_format): get rid of accessing by symbol of
invalid ID, when the name is not interned.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46756 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * test/ruby/memory_status.rb (Memory::PSCMD): use ps command which
outputs expected result. [ruby-dev:48370] [Bug #10010]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46755 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * parse.y (new_op_assign_gen): expand is_asgn_or_id(), which call
id_type() function multiple times now.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46754 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 05:48 AM Revision 7defa383 (git): * 2014-07-08
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46753 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * parse.y (dsymbol_check): set RSYMBOL(sym)->fstr to 0
because we should not touch fstr after that.
* parse.y (rb_gc_free_dsymbol): skip deleting str and sym
from tables if fstr == 0.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/tr...