Activity
From 06/22/2015 to 06/28/2015
06/28/2015
-
10:24 PM Revision 52436ce1 (git): sizes.c.tmpl: extract RUBY_DEFINT
- * template/sizes.c.tmpl: extract RUBY_DEFINT to define sizes of
types checked by configure.in, and fix size of intptr_t in
universal binary.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51056 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * insns.def (defineclass): preserve encoding of name in error
messages for super class mismatch.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51055 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
05:10 PM Feature #11315: [PATCH] Add Array#^ for parity with other set-like operations.
- I forgot to recycle the `seen` hash. This last fix should finalize the patch, unless there is some very clever way to XOR two arrays without requiring two tables or a second pass through one of the arrays.
-
- The original implementation did not correctly handle the case of repeated elements occurring an even number of times in the second array. Attached is an updated version which does not present this defect, courtesy of @apeiros and his 50-...
-
- ## Proposal
It is proposed that `Array` be given an "XOR" method which returns the symmetric difference between itself and another array.
## Rationale
`#&` and `#|` need a friend to complete the set of set-like operations on `Ar... -
- * insns.def (defineclass): preserve encoding of name in error
messages for non-class super.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51054 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * insns.def (defineclass): preserve encoding of name in error
messages when already defined but type mismatch.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51053 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 04:37 PM Revision 3ed6f43b (git): * 2015-06-29
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51052 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * test/ruby/test_{class,module}.rb (test_redefinition_mismatch):
add tests of redefinition mismatch.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51051 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
11:46 AM Bug #11314 (Third Party's Issue): [BUG] Segmentaton fault
- I'm working in Windows7 with VirtualBox,Vagrant.
But after I installed vagrant plugin "vagrant-centos7_fix", "vagrant up" command become to failed.
Error log is attached here as "error.log".
after installed the plugin, this occurs... -
- Why do you need to give a `nil`?
-
05:26 AM Bug #11001: 2.2.1 Segmentation fault in reserve_stack() function.
- > Any criteria where works on or not?
> ...
Segmentaion fault
* Ubuntu 12.04 - kernel 3.2+
* Ubuntu 12.10 - kernel 3.5
* Debian 7.7 - kernel 3.2.63
* Debian 7.8 - kernel 3.4.105
No problem
* Ubuntu 13.04 - kernel 3.8
* Ubuntu 1... -
- Takehiro Kubo wrote:
> IMO, this may be an issue of OS, not of ruby.
Any criteria where works on or not?
Runtime kernel versions?
-
- * class.c (rb_check_inheritable): preserve encoding in an error
message when the superclass is not a class.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51050 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 03:07 AM Revision c70572d5 (git): * 2015-06-28
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51049 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * class.c (rb_define_class_id_under): raise TypeError exception
same as ruby level class definition when superclass mismatch.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51048 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
06/27/2015
-
11:38 PM Bug #11001: 2.2.1 Segmentation fault in reserve_stack() function.
- On 2015/06/27 15:08, Eric Wong wrote:
> Fwiw, valgrind chokes here, too.
Try it before valgrind.
$ ulimit -s unlimited
--
// SASADA Koichi at atdot dot net -
06:08 AM Bug #11001: 2.2.1 Segmentation fault in reserve_stack() function.
- kubo@jiubao.org wrote:
> IMO, this may be an issue of OS, not of ruby.
Fwiw, valgrind chokes here, too.
I tested valgrind 3.9.0 on CentOS 7.0 and also backported to
Debian 7.0 (wheezy).
I comment reserve_stack out when I'm de... -
- Same issue with recent ruby versions on Ubuntu 12.04.
I tried it with recent OS versions and found that
* Ubuntu 12.04 - Segmentation fault
* Ubuntu 12.10 - Segmentation fault
* Ubuntu 13.04 - No problem
* Ubuntu 14.04 - No prob... -
02:13 PM Bug #11291: [PATCH] Check SPECIAL_CONST_P before RBASIC_CLASS to avoid SEGV
- Backported into `ruby_2_2` branch at r51047.
-
- * ext/objspace/objspace_dump.c(dump_object): Return empty JSON object when
passed object is a special const, instead of SEGV.
Based patch by Kohei Suzuki (eagletmt). [ruby-core:69692] [Bug #11291]
* test/objs... -
- It seems that SEGV occurred in mysql.gem extension. Please report to the gem project.
-
10:13 AM Bug #11260 (Closed): [PATCH] Net::FTP keeps a TCPServer opened when sending the data port fails
- Thanks, committed in r51046.
-
09:04 AM Bug #11260 (Assigned): [PATCH] Net::FTP keeps a TCPServer opened when sending the data port fails
- Benoit Daloze wrote:
> @shugo, can I commit this?
The patch looks fine. Please commit it. -
- when sending the port fails.
* test/net/ftp/test_ftp.rb: test for above.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51046 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
09:55 AM Feature #10594: Comparable#clamp
- Here's a pull request for this: https://github.com/ruby/ruby/pull/947
-
02:26 AM Feature #11308: Optional `include_super=true` parameter for `*method_defined?` methods
- Rationale? I don't see any real-world use-case for them.
Matz.
06/26/2015
- 10:52 PM Revision b8d54483 (git): Revert "st.c: use ccan linked-list (try 2)"
- This reverts commit r51044
Still getting failure notices from ko1's CI machine.
ref: g3qkqn
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51045 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
10:47 PM Bug #11313 (Closed): Fixed a couple of typos
- 10:32 PM Revision d3725a86 (git): st.c: use ccan linked-list (try 2)
- This improves the bm_vm2_bighash benchmark significantly by
removing branches during insert, but slows down anything
requiring iteration with the more complex loop termination
checking.
Speedup ratio of 1.10 - 1.20 is typical for the vm... - 08:19 PM Revision 08414a6a (git): common.mk: add ccan/list headers as deps for st.c
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51043 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 07:01 PM Revision c803bf3a (git): * 2015-06-27
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51042 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 06:59 PM Revision c94a9480 (git): st.c: include ccan/list/list.h to test build
- I suspect the build failures with r51034 ("st.c: use ccan linked-list")
on ko1's CI machine was due to me forgetting to update common.mk :x
Lets see what happens when I only include ccan/list/list.h
Will followup with appropriate revert... -
06:58 PM Feature #11312 (Assigned): Add Resolv::DNS::Resource::IN::SPF
- The SPF type record (RFC4408 Section 3.1.1) is quite common now. The resolv library should probably support it. As a workaround I monkey patched in the resource class:
class Resolv::DNS::Resource::IN::SPF < Resolv::DNS::Resource::... -
04:13 PM Bug #11311: STRING::Scan method badly interpret i option - Ruby 2.1.*
- ~~~ruby
'SKYPE'.scan(/[^aeiouyAEIOUY\W]/i) #returns ["P"] instead of ["S", "K", "P"]
'SKYPE'.scan(/[^aeiouyAEIOUY\W]/) #returns ["S", "K", "P"]
~~~
Tested on Ruby 2.1.3 / 2.1.5 / 2.1.6
regexp i option: case insensitive
https://gi... -
- ~~~ruby
'SKYPE'.scan(/[^aeiouyAEIOUY\W]/i) #returns ["P"] instead of ["S", "K", "P"]
~~~
Tested on Ruby 2.1.3 / 2.1.5 / 2.1.6
regexp i option: case insensitive
https://github.com/ruby/ruby/blame/trunk/string.c#L7681 -
03:16 PM Bug #11310 (Closed): CSV::Table#to_csv should use concat not +
- The current implementation of CSV::Table#to_a method uses Array#+ to which is really bad on performance. It seems like it should use Array#concat.
Relevant Github PR: https://github.com/ruby/ruby/pull/946 -
02:58 PM Feature #11309: Iterator over string matches
- Oops, I meant `pre_match`, not `post_match`. Sorry.
-
- This was hinted from a problem in stackoverflow (http://stackoverflow.com/questions/31074050/build-list-of-strings-containing-substrings-separated-by-an-from-a-string/31075511#31075511).
Suppose there is a string:
s = "a_b_c_d_... -
10:16 AM Bug #10460: Segfault instead of stack level too deep
- I think the bug is still there. I'm able to reproduce it on the nightly snapshot (dated 25 June 2015 on the FTP server).
# How to Reproduce
Use the following snippet:
~~~
require 'minitest'
require 'minitest/autorun'
class ... -
- * sprintf.c (rb_str_format): rb_inspect() always returns a string.
rb_obj_as_string() has no more effect.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51040 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * object.c (convert_type): check common prefix of converter
method names first. shrink conv_method_names.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51039 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 06:22 AM Revision e3815c47 (git): ChangeLog: doc accidentally squashed commit in r51036
- Oops :x
* hash.c (rb_hash_fetch_m): remove unnecessary volatile since r41597
(env_reject_bang): trade volatile for GC guard
(env_select): ditto
(env_select_bang): ditto
(env_keep_if): ditto
(rb_env_clear): ditto
git-svn-id: s... -
- I request the following methods in the `Module` class:
* `method_defined?`
* `private_method_defined?`
* `protected_method_defined?`
* `public_method_defined?`
to take an optional `include_super` parameter with the default value... -
- Applied in changeset r51037.
----------
transcode.c: fix race condition
* transcode.c (load_transcoder_entry): fix transcoder loading race
condition, by waiting in require. [ruby-dev:49106] [Bug #11277] -
- * transcode.c (load_transcoder_entry): fix transcoder loading race
condition, by waiting in require. [ruby-dev:49106] [Bug #11277]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51037 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
01:25 AM Bug #11306: Segmentation fault
- Thank you. This has now been reported as a gem error: https://github.com/m0wfo/cups/issues/21
and I added Eric Wong's response as a comment to that issue.
06/25/2015
- 11:06 PM Revision 9f1c0094 (git): enum.c (enum_minmax): simplify return value creation
- No need to call three functions on success when one will do.
This results in less LoC and smaller object code, too:
text data bss dec hex filename
33860 0 296 34156 856c gcc/enum.o-before
33852 ... -
- As we learned from the nonblocking IO APIs, exceptions are noisy with debugging
enabled and waste allocations on backtraces. We should have a better way to do
non-blocking Queue#pop without raising exceptions, but I don't know what th... - 07:56 PM Revision cc5547a4 (git): Revert r51034 "st.c: use ccan linked-list"
- Maybe this will stop mysterious CI failures from ko1@sasada-8core:
リビジョン 51034 です。
make[1]: ディレクトリ `/mnt/sdb1/ruby/build' に入ります
../trunk/revision.h unchanged
make[1]: ディレクトリ `/mnt/sdb1/ruby/build' から出ます
make[1]: ディレクトリ `/mnt/sdb1/ruby/b... - 07:01 PM Revision d8748874 (git): st.c: use ccan linked-list
- This improves the bm_vm2_bighash benchmark significantly by
removing branches during insert, but slows down anything
requiring iteration with the more complex loop termination
checking.
Speedup ratio of 1.10 - 1.20 is typical for the vm... - 03:19 PM Revision bbe6aa72 (git): * 2015-06-26
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51033 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
03:17 PM Revision d8a203fd (git): fix a typo [ci skip]
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51032 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
02:12 PM Bug #11277: "code converter not found" error with multi-thread (high occurrence rate since r50887)
- r50887 以降では、IOを使わなくても以下のようにすれば高確率で再現できました。
Converter が見つからない Error だけでなく、 Encoding のロードに失敗した Warning も出ているのがわかります。
~~~
% ruby --disable=gems -e '(0..2).collect { |_| Thread.new { p "\u3042".encode("EUC-JP") }}.each { |t| t.join }'... - 12:43 PM Revision c00e1880 (git): * properties.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51031 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * ext/-test-/popen_deadlock/infinite_loop_dlsym.c: new ext to call
dlsym(3) infinitely without GVL, used in the above test.
* ext/-test-/popen_deadlock/extconf.rb: extconf.rb for the above
ext. Currently, only enabled on Solaris (ma... -
- Applied in changeset r51029.
----------
* configure.in: not to use vfork on Solaris to avoid deadlock
occurred in vfork(2) with multi-threading and dynamic linker
on Solaris. [Bug #11265] [ruby-dev:49089] -
- > fork だと、fork 時に運悪くロックが確保されていて子プロセスにそのロックがそのままコピーされると、
> ...
それは確かに心配なので、簡易な実験をしてみました。
添付の、別スレッドで dlsym をループして呼び続けながら、メインスレッドで fork または vfork を呼ぶプログラムを何回も実行したところ、
vfork の場合は、3%程度の割合でデッドロックに陥り親子とも停止しましたが、
fork の場合は、1000回以上実行しても1回もデッド... -
01:52 AM Bug #11265: deadlock on Solaris 10 since r50900
- ふと、fork にすると本当に解決するのだろうか、という疑問が浮かびました。
vfork でデッドロックになるのは、
(Solaris では親プロセスでvforkを呼び出したスレッドだけでなく、他のスレッドも停止するため)
ロックを持っているスレッドが動かないからだと思うのですが、
fork だと、fork 時に運悪くロックが確保されていて子プロセスにそのロックがそのままコピーされると、
子プロセスにはロックを持っているスレッドが存在しないため、
やはり... -
- occurred in vfork(2) with multi-threading and dynamic linker
on Solaris. [Bug #11265] [ruby-dev:49089]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51029 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- A bug of that gem.
-
- dsaronin@gmail.com wrote:
> ~~~c
> static VALUE cups_get_device_uri(VALUE self, VALUE printer)
> {
> if (!printer_exists(printer))
> {
> rb_raise(rb_eRuntimeError, "The printer or destination doesn't exist!");
>... -
- bug is detected during CUPS call in cups gem (https://github.com/m0wfo/cups) v0.1.10. The complete core dump is attached as file 'dump.txt'.
the following works okay:
~~~
2.2.1 :001 > list = Cups.show_destinations
=> ["laserjet_... -
- NO_MEMORY_LEAK_ENVS is moved to Memory::NO_MEMORY_LEAK_ENVS
to reduce child executions during test-all on Solaris.
* test/lib/memory_status.rb (Memory::NO_MEMORY_LEAK_ENVS): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@5... -
- method_definition_reset() if def is given.
Actually, `me' is a new object, so we don't need to call it.
It is just to make sure.
* vm_method.c (method_definition_reset): remove duplicated insertion.
* vm_method.c (rb_method_entry_clo... -
- a rb_method_entry_t data (me).
Normally, `me' points `def'. Some Ruby objects pointed from `def'
and objects are marked by `me' (mark_method_entry() in gc.c).
However, `def' is built before making a `me', then nobody can mark
obj... -
- * error.c (rb_name_err_mesg_new): remove singleton method, and
unused argument.
* vm_eval.c (make_no_method_exception): call rb_name_err_mesg_new
directly instead of constant lookup and method invocation.
git-svn-id: svn+ssh://ci.r... -
- * random.c (fill_random_bytes): separate non-raced and raced
conditions, on Windows.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51024 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * test/test_prime.rb (test_eratosthenes_works_fine_after_timeout):
use assert_raise to check timeout.
* test/test_securerandom.rb: check if results are hexadecimal
strings, and refine failure messages.
* test/test_tracer.rb (test_t... -
- * test/test_prime.rb (test_eratosthenes_works_fine_after_timeout):
use spaces instead of TABs in ruby codes. [Fix GH-944]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51022 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- The ipaddr.rb library has a father complex and custom parser. If an IP address fails to parse, it will raise an InvalidAddressError from one of it's many parsing methods. The IP address should be included in the exception message for con...
-
- The default behavior of accept4() is fine.
Although most Ruby-level API is nonblocking flag insensitive,
there is nonblocking flag sensitive API such as IO#syswrite.
Also, the flag affects other (non-Ruby) programs when the FD is in...
06/24/2015
-
- instead of a number.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51021 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51020 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- VM_METHOD_TYPE_ATTRSET.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51019 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- rich obj_info() output.
At the default, the value of RGENGC_OBJ_INFO is
(RGENGC_DEBUG | RGENGC_CHECK_MODE).
* gc.c (RGENGC_OBJ_INFO): force enable it to debug #11244.
* gc.c (gc_mark_ptr): print more details with obj_info().
* gc.c (... - 05:44 PM Revision 8b192636 (git): * 2015-06-25
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51017 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 05:44 PM Revision 3d7473e1 (git): ext/openssl/ossl_ssl.c: raise if kwargs given in blocking mode
- OpenSSL::SSL::SSLSocket#sysread does not accept kwargs in blocking
mode, inform users if they make an error.
* ext/openssl/ossl_ssl.c (ossl_ssl_read_internal):
do not process kwargs in blocking mode
* test/openssl/test_ssl.rb: test sy... -
03:44 PM Feature #11302: Dir.entries and Dir.foreach without [".", ".."]
- Nobuyoshi Nakada wrote:
> Candidates for the methods or options?
> ...
`Dir.foreach` and `Dir.entries` both support a second hash argument for options: as of 2.2.2, the docs only mention the `:encoding` key in the options hash.
Basi... -
- Candidates for the methods or options?
I prefer a same option for both methods, but no concrete idea. -
06:24 AM Feature #11302 (Closed): Dir.entries and Dir.foreach without [".", ".."]
- Dir.entries returns an array of its content with "." and "..".
But as far as I met, almost all cases don't need them.
How about adding such new method or options? -
- funny.falcon@gmail.com wrote:
> Did you make a mistake, or `after` is really slower?
Good catch! I mislabeled the lines. This patch makes things faster. -
09:10 AM Feature #11298: [PATCH] socket: memoize common socket families in fptr->mode
- Did you make a mistake, or `after` is really slower?
-
- The added complexity does impact performance, but the difference is negligible, and the method should return what it says it does.
-
- imho, 0001 is very important, cause it gives most of performance improvement for usual case.
For now, I think twice: use struct or object, despite that struct is looking better and provides smaller and cleaner code.
If setter's will ... -
08:41 AM Bug #10398: Server Name Indication support broken when reusing a (dead) session
- Same issue: #10533
-
- Related i.e same issue: #10398
-
- * vm_method.c (rb_attr): simply use the current visibility,
instead of tests for each visibilities.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51015 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * vm_method.c (rb_attr): remove redundant check. attribute names
given in ruby level should be checked before calling this
function.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51014 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * load.c (rb_feature_provided): use RB_GC_GUARD for fullpath after
last use of feature.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51013 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
08:25 AM Bug #11303 (Third Party's Issue): [BUG] Segmentati on fault ruby 2.2.2p95 (2015-04-13 revision 50295) [x64-mingw32]
- $ rails server
c:/Ruby22-x64/lib/ruby/gems/2.2.0/extensions/x64-mingw32/2.2.0/mysql-2.9.1/mysql/mysql_api.so: [BUG] Segmentati
on fault
ruby 2.2.2p95 (2015-04-13 revision 50295) [x64-mingw32]
-- Control frame information ----------... -
- Updated v2 patch.
I care about this more, now, since I want to try to make unordered hash an
option with st.c in the future for internals. This should make future changes
easier-to-understand with less code.
I'm willing to trade... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51012 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 07:17 AM Revision a12fb26f (git): string.c (rb_str_justify): use RB_GC_GUARD
- See doc/extension.rdoc for justification using RB_GC_GUARD instead
of volatile.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51011 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- I think we have tried it some times but not made a conclusion.
Change `Module#prepend` only? -
- I don't think it's ideal:
~~~ruby
module M
end
class A
prepend M
end
A.ancestors
# => [M, A, Object, PP::ObjectMixin, Kernel, BasicObject]
class B<A
prepend M
end
B.ancestors
# => [B, M, A, Object, PP::Obje... -
- Marc-Andre, thank you for the info. You have made the issue clear.
Now I can safely close the issue.
Matz.
-
05:25 AM Bug #11296: Proc#arity returns weird value for non-lambda proc with optional arguments
- This is intentional, as per #5694 (note 22). It is not really related to #7765.
FWIW, I still believe it was not the best decision.
-
- Koichi Sasada wrote:
> Recent fix changed behavior of this script.
> ...
It's unacceptable. CREF should be copied at least when refinements are activated in the context.
-
- Koichi Sasada wrote:
> The following program making two refinements refine class C.
> ...
It's intentional.
In refine blocks of a module X, all refinements defined in X are activated, and other refinements previously activated are d... -
- It should be removed later. But we can remain this check
because it is only a branch.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51010 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
04:37 AM Bug #11244: [BUG] rb_gc_mark(): 0x00000001f34020 is T_NONE
- その後、かなり高い頻度で発生するようになっています。
毎回 gc_mark_stacked_objects_all() の呼び出しから起きていて、incremental では起きないようなので、たぶんなんかその場合のみ引っかかるオブジェクト絡みで根本的な欠陥があるのでしょう。 -
- * string.c (rb_fstring_cstr): new function to make a fstring from
a string literal.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51009 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * internal.h (rb_fstring_lit): new macro to make a fstring from a
string literal.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51008 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * include/ruby/intern.h (rb_strlen_lit): new macro to get the
length of a string literal, borrowed from mruby/mruby@e4afd53.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51007 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
03:21 AM Revision 7fac6912 (git): * re.c: Update documentation for Regexp class.
- [fix GH-937][ci skip] Patch by @davydovanton
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51006 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- matz@ruby-lang.org wrote:
> Two points. Unlike `#hex` and `#oct`, there's little demand for 0/1 strings conversion.
> Besides that, `String#bin` can be easily confused with "binary (raw) encoding" and cause more surprise.
> Any op... -
- Two points. Unlike `#hex` and `#oct`, there's little demand for 0/1 strings conversion.
Besides that, `String#bin` can be easily confused with "binary (raw) encoding" and cause more surprise.
Any opinion?
Matz.
-
- ## Proposal
It is proposed that `String#bin` be added to return the receiver's value when interpreted as a string of binary digits.
## Rationale
This method's absence is surprising in the presence of `#hex` and `#oct`. -
- matz@ruby-lang.org wrote:
> I agree with Ko1 here. GC pressure from object allocation is far
> higher than the pressure from byte allocation.
This depends on the application, of course; and I expect GC will improve.
Reducing cod... -
- ko1@atdot.net wrote:
> I'm weakly against this proposal because it will consumes more objects
> and increase GC pressure.
Existing code with xrealloc still has GC pressure, too.
> I don't think it is valuable compare with reduci... -
- I agree with Ko1 here. GC pressure from object allocation is far higher than the pressure from byte allocation.
Matz.
- 12:24 AM Revision 1782dd8c (git): variable.c: remove FL_ABLE check for generic ivars
- Missed this in r50758.
* variable.c (generic_ivar_set): remove FL_ABLE check
(gen_ivar_copy): ditto
[ruby-core:69715]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51005 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
06/23/2015
-
- I'm weakly against this proposal because it will consumes more objects and increase GC pressure.
I don't think it is valuable compare with reducing binary size by 1KB.
BTW,
```ruby
- if (FL_ABLE(obj)) RB_OBJ_WRITTEN(obj, Qunde... -
- I'll commit this in a few days unless there's an objection.
~~~
This reduces both code and object size while reducing the cognitive
overhead necessary to understand the code. Memory usage should be
slightly higher due to Array ove... -
11:41 PM Feature #11263 (Closed): [PATCH] ext/socket/ancdata.c (bsock_recvmsg_internal): reduce stack use - Applied in changeset r51003.
----------
ext/socket/ancdata.c (bsock_recvmsg_internal): reduce stack use
Using 8K stack is probably too much. As reference, ALLOCV falls
back to heap allocation at a mere 1K. Since
bsock_recvmsg_interna... - 11:41 PM Revision 3f172356 (git): * 2015-06-24
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51004 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 11:40 PM Revision d9d1b20d (git): ext/socket/ancdata.c (bsock_recvmsg_internal): reduce stack use
- Using 8K stack is probably too much. As reference, ALLOCV falls
back to heap allocation at a mere 1K. Since
bsock_recvmsg_internal is a function which will always allocate
and can trigger GC, it is in our best interest to minimize
stac... -
- I noticed an unnecesary getsockname() syscalls for every
sendmsg_nonblock call I make on SOCK_SEQPACKET pair.
This made for a noisy strace.
I also plan on following up and setting the flag on
rsock_init_* functions upon initializat... -
02:47 PM Feature #6284: Add composition for procs
- Attached patch to support composing with any object that responds to `call` (rather than raising a `TypeError` if the object was not a `Proc` or `Method`), e.g.
~~~ruby
class Foo
def call(x, y)
x + y
end
end
f = proc {... -
- Hans, thank you. That's what I meant.
Kazuhiro, #7765 suggests that seems intentional.
After hearing opnion from others, I will close the issue (unless meaningful discussion starts).
Matz.
-
- I think #5694 and #7765 are related issues.
-
12:55 PM Bug #11296: Proc#arity returns weird value for non-lambda proc with optional arguments
- i think matz means:
~~~ruby
Proc.new(){|x=0|}.arity # => 0
lambda {|x=0|}.arity # => -1
~~~ -
- `Proc.new(x=0){}` raises an `ArgumentError` for me, in 1.8..2.3.
-
- `Proc.new(x=0){}.arity` returns 0, not -1 unlike lambdas.
I don't remember if it's intentional or not. Anyone remember?
Matz.
-
01:27 PM Feature #11297 (Closed): Allow private method of self to be called
- Ruby does not allow private method to be called if receiver is given. Calling private method with receiver is prohibited even if it is written as self, though the fact that the receiver is self is still clear.
This ticket is to propos... -
12:00 PM Bug #11294: Possible bug in Object.const_get
- Thank you, Nobuyoshi, for your answering. Now I see the reason.
When I told that `Foo::Bar::Qux` is "closest" to `Foo::Baz` I meant that in a namespace:
~~~ruby
module Foo
module Baz
p Bar::Qux
end
end
# => Foo::Bar::... -
- 親プロセス側のスレッド l@5 内で呼ばれている dlsym() で要求しているシンボルは _ex_unwind であるのがわかりました。
（下記のdbxの出力は、上記とは別の実行時になりますが、おおむね同じです。）
~~~
t@1 (l@1) stopped in _vfork at 0x7fffffff7e8dd440
0x7fffffff7e8dd440: _vfork+0x0024: bcc,a,pt %icc,_vfork+0x4c ... -
- * error.c (name_err_receiver): raise ArgumentError if no receiver
is available on this exception object. [Feature #10881]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51002 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * error.c (rb_name_err_mesg_new, name_err_mesg_to_str): name magic
numbers of indexes.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51001 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- ko1@atdot.net wrote:
> There are several proposals to extend error messages.
>
> * https://github.com/charliesome/better_errors
> * https://github.com/yuki24/did_you_mean
> * https://github.com/ko1/pretty_backtrace
What perform... -
- (This is not a proposal, bug reports)
Ruby shows error messages when something wrong.
There are several proposals to extend error messages.
* https://github.com/charliesome/better_errors
* https://github.com/yuki24/did_you_mean... - 12:49 AM Revision 3fe2773c (git): dir.c (check_dirname): avoid volatile, use return value
- volatile is unnecessary since we use rb_sys_fail_path nowadays
and that prevents the path argument from being GC-ed.
Using a return value instead of modifying the argument directly
simplifies the generated code (on 32-bit x86):
text ...
06/22/2015
-
- Toplevel constants are defined under `Object` context, which are inherited by every contexts.
Your results are same as:
~~~ruby
module Foo::Baz; p Bar::Qux; end #=> Bar::Qux
~~~
> * or find the closest `Bar::Qux` to `Foo::Bar`... -
- ~~~ruby
module Foo; end
module Foo::Baz; end
module Bar; end
module Bar::Qux; end
Object.const_get "Foo::Baz::Bar::Qux"
# => Bar::Qux
~~~
Why on earth it is found at all?
It seems pretty weird to me.
The real problem aris... -
- Try `Integer('0123', 10)`.
-
03:20 PM Bug #11293: Integer('0123') => 83
- Matt Di Pasquale wrote:
> Is this the correct behavior?
> ...
No, 83 would be the expected result, just as if you specified 0123 in ruby code. Kernel#Integer treats a leading 0 to mean octal, leading 0x to mean hexidecimal, and leadin... -
03:02 PM Bug #11293 (Rejected): Integer('0123') => 83
- Is this the correct behavior?
irb(main):001:0> Integer('0123')
=> 83
Shouldn't the return value be `123`? -
- Thanks, I've committed 0002, 0003, and would like to commit a
slightly-updated 0004 soon unless somebody else objects:
Updated 0004:
http://80x24.org/spew/m/f5c7f28bff5e378b71425ac78f043b5093dac1f4.txt
I'm not a big fan of 00... - 09:50 PM Revision 099e6e78 (git): struct.c: cache member definition in a subclass
- Since getting Qnil is already error, it is safe to use rb_attr_get.
* struct.c (struct_ivar_get): cache member definition in a subclass
Thanks to Sokolov Yura aka funny_falcon <funny.falcon@gmail.com>
in https://bugs.ruby-lang.org/i... - 08:10 PM Revision 8b71b4e5 (git): * properties.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50998 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 08:10 PM Revision 1a8e3ba1 (git): * 2015-06-23
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50997 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 08:10 PM Revision a63c5b26 (git): add benchmarks for struct[:name]
- * benchmark/bm_vm2_struct_big_href_hi.rb: new benchmark
* benchmark/bm_vm2_struct_big_href_lo.rb: ditto
* benchmark/bm_vm2_struct_big_hset.rb: ditto
* benchmark/bm_vm2_struct_small_href.rb: ditto
* benchmark/bm_vm2_struct_small_hset.rb: ... -
- I don't think the flag matters for second-tier OSes, as the underlying
descriptor does not matter with our current APIs.
So better to leave the OS default if sock_nonblock kwarg is not
specified. -
- SASADA Koichi <ko1@atdot.net> wrote:
> I have no opinion about that, if there are no behavior changes.
> All of operations are , same behavior and same computational complexity?
Same behavior for enabled features. I was not able t... -
06:07 PM Feature #11158: Introduce a Symbol.count API as a more efficient alternative to Symbol.all_symbols.size
- Marc-Andre Lafortune wrote:
> Franck Verrot wrote:
> ...
Sorry I haven't formulated this right :-) I was only wondering if including `Enumerable` in `Symbol` could lead some of us to rely on methods (like `map` as you said) that weren'... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50995 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- r36385 (2012/7/14) にて、launch_worker の行が watchdog = start_watchdog の後ろに移動していました。
大丈夫そうなので、r50994 にて削除しました。
-
- because it has been meaningless since r36385. [Bug:11288]
* test/lib/test/unit.rb (Test::Unit::Parallel#_run_parallel): delete
lines related to the removed start_watchdog method
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@5099... -
- Could you add use-case of this patch?
- 12:21 AM Revision cc934fc3 (git): * 2015-06-22
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50993 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * win32/win32.c (open_dir_handle): use numberof() macro.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50992 b2dd03c8-39d4-4d8f-98ff-823fe69b080e