Activity
From 06/04/2014 to 06/10/2014
06/10/2014
-
10:00 PM Feature #8259: Atomic attributes accessors
- thedarkone2@gmail.com wrote:
> I am all for addition of `ConcurrentHash` or `Concurrent::Array` (these
> new data structures would have `cas` and `swap` methods), but for
> performance reasons plain old `Hash` and `Array` should be k... -
09:35 PM Feature #8259: Atomic attributes accessors
- Eric Wong wrote:
> Right. There's no way I will ever advocate a memory barrier of any
> ...
My thesis and objection to adding "atomic" methods to `Array` or `Hash`
is that it would necessarily entail making them thread-safe as a who... -
- thedarkone2@gmail.com wrote:
> Eric Wong wrote:
> > I do atomic operations all the time in C on arbitrary addresses.
> > Lazy, non-atomic accesses run without speed penalty if I don't need
> > up-to-date data.
>
> Are you talk... -
- > I do atomic operations all the time in C on arbitrary addresses.
> ...
Are you talking about Ruby with GVL or C in general? If C in general then
I don't understand how barrier-less access to concurrently updatable
data does not res... -
08:18 PM Bug #9927: webrick does not unset content-length when responding to HEAD requests
- My error, sorry for the distraction.
-
07:58 PM Bug #9927: webrick does not unset content-length when responding to HEAD requests
- Adrien Thebo wrote:
> I didn't make this very clear, but the `curl` invocation hangs for about 6 seconds until the webrick server is killed, and then prints '`transfer closed with 23 bytes remaining to read`' which indicates it was hang... -
07:58 PM Bug #9927 (Rejected): webrick does not unset content-length when responding to HEAD requests
- This appears to be a bug in curl.
RFC 7230 says:
3.3. Message Body
The message body (if any) of an HTTP message is used to carry the
payload body of that request or response. The message body is
... -
- I didn't make this very clear, but the `curl` invocation hangs for about 6 seconds until the webrick server is killed, and then prints '`transfer closed with 23 bytes remaining to read`' which indicates it was hanging.
-
- When Webrick responds to HEAD requests it omits the body (per RFC2616 -- 4.4 Message Length). However when setting up the response headers the content-length field is set to the length of the body, which means that the resulting response...
-
08:07 PM Bug #9928 (Closed): Fiddle::TestHandle#test_NEXT fails on AIX due to unexported symbols of extension libraries
- `Fiddle::TestHandle#test_NEXT` fails on AIX.
~~~
[ 4/21] Fiddle::TestHandle#test_NEXT = 0.00 s
1) Error:
Fiddle::TestHandle#test_NEXT:
Fiddle::DLError: unknown symbol "Init_objspace"
/ss/home/rayod/... -
- SASADA Koichi <ko1@atdot.net> wrote:
> (2014/05/31 4:56), Eric Wong wrote:
> ...
I think it is important to tell users how it is different from
`rb_global_variable` and when it is OK or not OK to use. So I think
the implementation d... -
08:58 AM Feature #9894: [RFC] README.EXT: document rb_gc_register_mark_object
- (2014/05/31 4:56), Eric Wong wrote:
> ```diff
> ...
How about that?
> Tells GC to protect the object referenced by val.
Another things are implementation details.
And I'm not sure the following sentence is needed.
> ...
I t... -
- SASADA Koichi <ko1@atdot.net> wrote:
> (2014/03/15 4:07), normalperson@yhbt.net wrote:
> > Also limits symbol space to prevent OOM.
>
> What is OOM?
> Out of memory?
Yes, out-of-memory.
> Symbol GC doesn't help?
It does; ... -
- (2014/03/15 4:07), normalperson@yhbt.net wrote:
> Also limits symbol space to prevent OOM.
What is OOM?
Out of memory?
Symbol GC doesn't help?
--
// SASADA Koichi at atdot dot net -
06:07 PM Feature #9781: Feature Proposal: Method#super_method
- I am also not in love with the naming. I would have preferred `Method#super` but we don't want to overwrite `super` or confuse functionality here. Maybe something along the lines of `get_super`. Does anyone have a better naming suggestion?
-
08:51 AM Feature #9781 (Feedback): Feature Proposal: Method#super_method
- Having something like Method/UnboundMethod#super_method for debugging purpose is OK, but
* Is super_method is the right name? It sounds little bit weird for me.
* I don't like Object#super_method.
Matz.
-
05:43 PM Revision ad5214df (git): merge revision(s) r45308,r45316: [Backport #9621]
- * ext/objspace/objspace_dump.c: Check fptr before trying to dump RFILE
object fd. [GH-562]
* test/objspace/test_objspace.rb: add test
* test/objspace/test_objspace.rb (TestObjSpace#test_dump_uninitialized_file... -
- Backported into `ruby_2_1` branch at r46402.
-
- * configure.in: Fix a build problem with clang and --with-opt-dir.
If ruby is configured with --with-opt-dir=dir when using clang
as compiler, a warning `clang: warning: argument unused during
compilation: '... -
- * numeric.c: Create var for rb_intern("<=>")
* numeric.c: Fix Numeric#step with 0 unit [Bug #9575]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@46401 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Applied in changeset r46399.
----------
* gc.c: invoke GC before memory allocation (xmalloc/xrealloc)
when GC.stress = true.
[Bug #9859]
* test/ruby/test_gc.rb: add a test. -
05:59 AM Bug #9859: An object with 6 instance variables causes SEGV
- 補正が入らなくなるというのがどのくらい問題があるのかわかっていませんが、
SEGV するよりはずいぶんとマシなんじゃないですかねぇ。 - 04:55 PM Revision 16b512b3 (git): * 2014-06-11
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46400 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- when GC.stress = true.
[Bug #9859]
* test/ruby/test_gc.rb: add a test.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46399 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- this test fails on some CI environment by timeout.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@46398 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
02:44 PM Revision 9ada2641 (git): envutil.rb: avoid error on OS X 10.10
- * test/ruby/envutil.rb (diagnostic_reports): Mac OS X 10.10 may
not create plist file with crash file together.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46397 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * bootstraptest/runner.rb (exec_test, show_progress): erase with
white spaces, and adjust each results at same column.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46396 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
01:54 PM Bug #9926 (Closed): class_eval : string and block - I'm not sure, is this a bug or a feature ?
class Foo; @@bar = 'bar'; end
p Foo.class_eval('@@bar') # => "bar"
p Foo.class_eval { @@bar } => uninitialized class variable @@bar in Object (NameError) -
12:23 PM Bug #9922: Compiling Ruby 1.9.3p194 from source code with openssl crashe
- Are you sure you disabled "Program Has Stopped Working Error" dialog and Just-in-Time Debugging?
Then you can not see any pop-up message.
-
11:16 AM Bug #9922: Compiling Ruby 1.9.3p194 from source code with openssl crashe
- 'Nothing in debugger' by this i mean, when i see pop-up msg that ruby has stoppped working, it gives me option to debugor close the program. I choose to debug with Visual studio8 with Just in time debugger option enabled. But i did not s...
-
- I am not sure what you mean by "nothing in debugger".
You must run ruby on command prompt without any debugger.
-
- Yes i did. It dod not show pop-up. but in command prompt i got same error only. And nothing in debugger
-
- Did you try to disable "Program Has Stopped Working Error" dialog?
For Vista / Win7 / Server 2008 etc:
1. Open start menu and type “gpedit.msc”. The Local Group Policy editor will appear.
2. Navigate to Computer Configuration / Admi... -
- debugger opens blank file. Also tried installing eventmachine and thin again with version used in my application. Here is gem list:-
>gem list
*** LOCAL GEMS ***
bigdecimal (1.1.0)
bundler (1.3.4)
daemons (1.1.9)
em-synchrony (... -
- It seems that Just-In-Time Debugging of Visual Studio inhibit the backtrace output of ruby.exe.
You can get the full backtrace by disabling Just-In-Time Debugging.
Refer to http://msdn.microsoft.com/en-us/library/5hs4b7a6(v=vs.90).... -
07:36 AM Bug #9922: Compiling Ruby 1.9.3p194 from source code with openssl crashe
- Pooja Saxena wrote:
> Yes I compiled openssl also with same version of visual studio. Did not do anything related to event machine library
thin depends on eventmachine, if you haven't compiled eventmachine dependency, can you tell us... -
- Dont know much about dependency checking, hence did not get much. But understood that three of them depends on both msvcrt.dll and msvcr90.dll. But logs in windows say faluting module path to msvcr90.dll. Attaching screenshot of dependen...
-
- You can check the MSVCRT dependencies with Dependency Walker(http://www.dependencywalker.com).
Please check the dependency of ruby.exe, libeay32.dll and ssleay32.dll.
-
- Yes I compiled openssl also with same version of visual studio. Did not do anything related to event machine library
-
- Was OpenSSL also compiled with the same version of Visual Studio?
Thin depends on both OpenSSL and Eventmachine library, the last being a bit problematic for some versions of Visual Studio (and also MinGW).
Can you confirm all the ... -
- due to other projects dependecies, I am bound to use that version only. From all the informaiton i explored, got to know that it has to do with compiling dependency libraries (\MSVCR90.dll) same msvcrt version. And that is working correc...
-
05:34 AM Bug #9922: Compiling Ruby 1.9.3p194 from source code with openssl crashe
- I have not been able to reproduce this problem at this time.
From the information that is provided by you, I can not debug unfortunately.
It should be noted, ruby 1.9.3 has ended the normal maintenance phase already,
we do not pla... -
- Any update on this??
-
12:12 PM Revision 26848ef8 (git): fix a typo [ci skip]
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46395 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- (2014/05/31 7:28), Rei.Odaira@gmail.com wrote:
> How about this patch? I tested this on x86/Linux with -DTHREAD_DEBUG=-1. No error happened in make test or test-all.
i have no objection.
--
// SASADA Koichi at atdot dot net -
- * win32/setup.mak (verconf.mk): rename RUBY_PROGRAM_VERSION as
RUBY_RELEASE_DATE, same as other platforms.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46394 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * win32/setup.mak (verconf.mk): separate RUBY_RELEASE_DATE from
Makefile so that build would not stop everyday.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46393 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
04:29 AM Feature #8370 (Closed): Constants MAX_MULTIPART_LENGTH in cgi/core.rb
- Applied in changeset r46392.
----------
* lib/cgi/core.rb: Provide a mechanism to specify the
max_multipart_length of multipart data.
[Feature #8370] patch by Leif Eriksen <leif.eriksen.au@gmail.com> - 04:29 AM Revision 5c99f241 (git): * lib/cgi/core.rb: Provide a mechanism to specify the
- max_multipart_length of multipart data.
[Feature #8370] patch by Leif Eriksen <leif.eriksen.au@gmail.com>
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46392 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- (2014/05/23 15:17), jjyruby@gmail.com wrote:
> For example, I want write a method, when under `EM::Synchrony` environment it should use `EM::Synchrony` call(in fact is async call, and writing in sync, power by fiber), and when under `... -
02:03 AM Feature #9925: rsock_addrinfo uses DNS family AF_UNSPEC for lookup causing high IPv6 AAAA volume
- Similar configuration is available in the "/etc/gai.conf" system (it does not apply in this use case, however). It is exactly the global nature of the setting that helps me the most. In many cases in the raddrinfo.c file, Ruby is hardcod...
-
- It doesn't feel a good idea to change a behavior in a library by an environment variable implicitly.
-
- Suggested code change: https://github.com/ruby/ruby/pull/636
-
- In ext/socket/raddrinfo.c, the function `rsock_addrinfo()` always uses `AF_UNSPEC` for DNS queries. This is causing me a very high volume of IPv6 DNS lookups. `rsock_addrinfo()` is used by TCPSocket (and all other Socket base classes, e....
-
- Applied in changeset r46391.
----------
csv.rb: honor encoding option
* lib/csv.rb (CSV#<<): honor explicity given encoding. based on
the patch by DAISUKE TANIWAKI <daisuketaniwaki AT gmail.com> at
[ruby-core:62113]. [Bug #9766] -
- * lib/csv.rb (CSV#<<): honor explicity given encoding. based on
the patch by DAISUKE TANIWAKI <daisuketaniwaki AT gmail.com> at
[ruby-core:62113]. [Bug #9766]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46391 b2dd03c8-39d4-4... - 01:45 AM Revision 47c64fdf (git): * 2014-06-10
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46390 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * test/ruby/envutil.rb (assert_valid_syntax): use keyword argument.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46389 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
06/09/2014
-
- thedarkone2@gmail.com wrote:
> Adding "atomic" operations to built in `Array`s, `Hash`es etc. is a very bad idea
> because that would force that *all* of the methods on `Array`s and `Hash`es be
> concurrency friendly and would result... -
- In my opinion an ideal API would have "atomic" attributes declared at the class level,
yet can be used as normal `@foo` i-vars:
~~~
class MyNode
atomic :item, :successor
def initialize(item, successor)
@foo = :foo # t... -
- Joel VanderWerf <joelvanderwerf@gmail.com> wrote:
> On 06/09/2014 04:06 AM, Eric Wong wrote:
> >I'm not sure if setting the attribute on the ivar is a good way to go.
> >Entries in structs, arrays, hashes, etc may also benefit from a... -
- I'm not sure if setting the attribute on the ivar is a good way to go.
Entries in structs, arrays, hashes, etc may also benefit from atomic
operations (or at least I would like that).
Maybe something like:
old = hash["foo"] x= ... -
10:09 AM Feature #8259: Atomic attributes accessors
- I should also point out that my implementation is using referential equality. I now think the default should be referential equality, perhaps (after some discussion and debate) with an option to do value equality. Of course, the value-eq...
-
- I have done a prototype of atomic variable accessors in JRuby.
Here's the IRB session. I opted to go with an :atomic option passed to attr_accessor rather than a separate method, but I don't have a strong preference. Atomic operations... -
- Waking this up again. Interest in gems like "atomic" and "concurrent-ruby" has continued to grow rapidly, and Ruby needs to start provided low-level concurrency constructs users need to build high-level constructs. That means atomics, tu...
-
- Now, GC.stat returns internal information abuot GC.
However, key names are from internal implementations fields,
so that it is difficult to understand what they mean.
To solve this problem, I reconsider about key names.
https://d... -
- Now, GC.stat returns internal information abuot GC.
However, key names are from internal implementations fields,
so that it is difficult to understand what they mean.
To solve this problem, I reconsider about key names.
https://d... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46388 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
01:05 PM Bug #9321: rb_mod_const_missing does not generate a c-return event
- Is there an official way to request a backport? Thanks!
-
- Applied in changeset r46387.
----------
* gc.c: change full GC timing to keep lower memory usage.
Extend heap only at
(1) after major GC
or
(2) after several (two times, at current) minor GC
Details in https://bugs.ruby-lang.o... -
- After introducing AGE2_PROMOTION patch, only a few old objects are added for each minor GC.
The impact of this issue (too many slots) are not critical now. But some application can not protect increasing total slots.
I rewrite algori... -
- Extend heap only at
(1) after major GC
or
(2) after several (two times, at current) minor GC
Details in https://bugs.ruby-lang.org/issues/9607#note-9
[Bug #9607]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46387 b2dd03... -
- Usaku NAKAMURA wrote:
> Pooja Saxena wrote:
> ...
Yes only this comes in console without any backtrace. Pop only only shows "ruby193p194 has stopped working". I am attaching screenshot of pop-up for your reference
-
- Pooja Saxena wrote:
> I tried compiling with p547 also. I get same error. And in ruby console I only see the msg below and no backtrace.
Oh, really?
I guess that the message ("This application ...") is shown in a popup window. Isn'... -
- Usaku NAKAMURA wrote:
> 1. Your ruby is too old. Please check p547, and report if you still have the same problem.
> ...
I tried compiling with p547 also. I get same error. And in ruby console I only see the msg below and no backtrace... -
- 1. Your ruby is too old. Please check p547, and report if you still have the same problem.
2. When ruby crashes, ruby outputs backtrace to the console. Please attach it. -
- Hi,
I am trying to compile ruby from source code. And i am using thin server. Problem is when i am trying to start thin with ssl option by specifying :-
ruby bin/thin --ssl -a 127.0.0.1 -p 44466 start
I starts the server, but on... -
- * gc.c (gcdebug_sentinel): fix typo, "sentinel" not "sential".
[fix GH-634]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46386 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
03:55 AM Bug #9904 (Assigned): remove rubyforge entries from rake.1
- Thanks for the report!
-
- Thanks for the report!
-
03:46 AM Feature #8370: Constants MAX_MULTIPART_LENGTH in cgi/core.rb
- OK noted - just trying to be super polite. Please lt me know if there is anything else I need to complete or consider for this fix.
-
- I meant "adding new feature".
And you didn't have to resubmit, pushing new changeset should be enough. -
02:42 AM Bug #9921: gmtime_r, win32.h, and <string>
- Interesting indeed, I wonder how feasible it would be within my actual program...
Heesob Park wrote:
> I guess error messages about gmtime_r must be solved with `#include <ctime>`
> ...
-
- I guess error messages about gmtime_r must be solved with `#include <ctime>`
~~~
#include <ctime>
#include <string>
#include "ruby.h"
int main() {
return(0);
}
~~~
06/08/2014
-
- Found an interesting/annoying issue, using ruby 2.1.2p95 on windows via the msys2 project;
seems there is an issue regarding compiling programs including <string> and "ruby.h" using mingw's g++.
The following test program fails to comp... -
- https://github.com/ruby/ruby/pull/633
-
- OK I'll remove the getter/setters and resubmit.
-
- Adding new methods is a feature request.
-
- OK done.
Used the pattern for passing accept_charset, which can be read/modified by a class method, and read by an instance method. That seems reasonable, let me know if a setter instance methods is also required, will do as a separat... -
- Well I'm only trying to solve the problem in the CGI library, not the underlying web server, and the only place that will reference MAX_MULTIPART_LENGTH is in the request parsing. That happens only in the CGI constructor. I dont think I ...
- 03:04 PM Revision 71992ebd (git): * 2014-06-09
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46385 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * configure.in (posix_fadvise): disable use of posix_fadvise
itself on 32-bit AIX. [ruby-core:62968] [Bug #9914]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46384 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * compile.c (new_insn_body): use enum instead of bare int.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46383 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
02:39 PM Bug #9914 (Closed): posix_fadvise() does not work correctly with _LARGE_FILES on 32-bit AIX
- Applied in changeset r46382.
----------
io.c (rb_io_advise): AIX currently does not support a 32-bit call to
posix_fadvise() if _LARGE_FILES is defined. Patch by Rei Odaira.
[ruby-core:62968] [Bug #9914] - 02:39 PM Revision 1c0b507b (git): io.c (rb_io_advise): AIX currently does not support a 32-bit call to
- posix_fadvise() if _LARGE_FILES is defined. Patch by Rei Odaira.
[ruby-core:62968] [Bug #9914]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46382 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- `sizeof` expression is not a preprocessor constant.
-
- Greetings,
While this may be a small issue, I think many c-language libraries would benefit from this suggestion;
While compiling ruby from source, `SIZEOF_VOIDP` is determined and hard-coded into $prefix/ruby-$version/$RUBY_PLATFO... -
05:10 AM Bug #9920 (Rejected): Ruby does not support Unicode 6.0 emoji, only private-use area Unicode emoji
- Translating emoji characters between Unicode and encodings like Shift JIS uses Unicode characters in the private-use range, instead of Unicode emoji characters.
Here's an example:
The SJIS character F6DB is an emoji representing an... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46381 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * bootstraptest/test_thread.rb: remove a garbage line for debug.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46380 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
06/07/2014
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46379 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * bootstraptest/runner.rb (show_progress): show each elapsed times
in verbose mode.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46378 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 07:52 PM Revision 35386858 (git): * 2014-06-08
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46377 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * string.c (rb_str_slice_bang): [DOC] update return value against
a fixnum, which has changed because of M17N. [fix GH-631]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46376 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
01:16 PM Revision a21ac07f (git): * numeric.c (do_coerce): Add a warning when an exception is raised
- or an invalid value is returned in #coerce called by
numeric comparison operators and the exception
thrown by the caller has no information on the failure.
In the next release such exception should not be rescued or
should be the... -
- Do you want to set the value of MAX_MULTIPART_LENGTH every request ?
I think request length is only one value in web server.
-
- OK, yeah I got the wrong end of the conversation - CGI is all about the server side.....
OK an idea comes to mind
A new option in the options_hash in the constructor - currently only :tag_maker and :accept_charset are defined
Th... -
04:35 AM Feature #8370: Constants MAX_MULTIPART_LENGTH in cgi/core.rb
- > There is no reason, that I can see, to artificially limit the upload size on the client side
I think that MAX_MULTIPART_LENGTH is the limit on the server side.
The limit itself is actually required to prevent shortage of server res... -
- From my reading of RFC-1867, this constant/constraint is not required.
If a server detects space constraints for an upload, it can terminate the connection at any time.
Servers can indicate a maximum length via the MAXLENGTH attrib... -
- I had another look at the other cases mentioned above.
* Comparable#==: A warning has been added when rescuing an exception of #<=>. There should be no more "rescue" after 2.2.0.
* numeric.c and #coerce: The cases where an error is... -
- Exception#cause has been added in Feature #8257.
But the cause is never shown, one needs to inspect exc.cause to know the cause.
The cause should be shown by default, for instance:
~~~ruby
begin
raise "cause"
rescue
raise ... -
- On AIX, time-out for `TestIO#test_io_select_with_many_files` expires because it takes more than 10 seconds to remove 1024< temporary files. The test itself succeeds if the time-out period is extended.
~~~
[ 85/170] TestIO#test_io_sel... -
12:08 PM Feature #9916 (Closed): Hide Struct internal
- is a bit big so I attach the proposal but basically same as my previous proposals. I'd like to make strct RStruct opaque. The benefit is that by hiding its `const VALUE* ptr` part we can optimize GC pressure. That GC tweaks are not in...
-
- OK, lets just abandon them. Patch updated.
-
- (always TRUE) of bit_coerce().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46374 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 07:11 AM Revision 3acff92b (git): cont.c : keep context.uc_stack.ss_sp and context.uc_stack.ss_size for later use.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46373 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- Applied in changeset r46372.
----------
io.c: truncate before appending
* io.c (read_all): truncate the buffer before appending read data,
instead of truncating before reading.
[ruby-core:55951] [Bug #8625] -
- r46360 breaks some tests on mswin.
-
- * io.c (read_all): truncate the buffer before appending read data,
instead of truncating before reading.
[ruby-core:55951] [Bug #8625]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46372 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * ext/digest/digest.c (rb_digest_instance_equal): no need to call
`to_s` twice. [Bug #9913]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46371 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- I decided to fix this problem because to_s method is useful for debugging and exception on to_s can be
irritating. -
- Applied in changeset r46370.
----------
* object.c (rb_mod_initialize_clone): Override Kernel#initialize_clone
to avoid an exception on Class.new.freeze.clone.to_s.
Reported by Andrew Grimm. [ruby-core:41858] [Bug #5828] -
- to avoid an exception on Class.new.freeze.clone.to_s.
Reported by Andrew Grimm. [ruby-core:41858] [Bug #5828]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46370 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
02:18 AM Feature #9909: why shouldn't constant lookup check the nesting of module's name
- See discussion in feature #6810.
-
12:01 AM Bug #9908: [mingw] openssl.so LoadError when built with 1.0.0m or 1.0.1h
- My custom built openssl libs are guilty, not the openssl ext.
The upgraded 32bit 4.9.0 toolchain I used to build the openssl libs uses dw2 rather than sjlj used by the previous version. The libeay32.dll implicitly linked against libgc...
06/06/2014
-
- Applied in changeset r46368.
----------
digest.c: fix #== for non-string arguments
* ext/digest/digest.c (rb_digest_instance_equal):
fix #== for non-string arguments. [ruby-core:62967] [Bug #9913]
* test/digest/test_digest.rb: add te... -
07:21 PM Bug #9913 (Closed): Digest == throws TypeError when testing against nil
- Also present in ruby -v: jruby 1.7.10 (1.9.3p392) 2014-01-09 c4ecd6b on OpenJDK 64-Bit Server VM 1.7.0_55-b14 [linux-amd64]
To recreate:
require 'digest'
x = Digest::SHA256.new()
x == nil
TypeError: can't convert nil into Strin... - 09:05 PM Revision aae18778 (git): * 2014-06-07
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46369 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * 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.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46368 b2dd03c8-39d4-... -
- `test_advise` in `test/ruby/test_io.rb` fails on AIX.
~~~
[ 1/170] TestIO#test_advise = 0.01 s
1) Error:
TestIO#test_advise:
Errno::EINVAL: Invalid argument - /tmp/test_io20140606-7733390-1xgpc07
/ss/home/rayod/Dev/Contrib... -
05:27 PM Bug #9912 (Closed): URI#to_s with an empty host formats the string incorrectly
- maciek@gamera:~$ ruby -e "puts URI.parse('postgres:///foo').to_s"
postgres:/foo
The string of the formatted URI should be 'postgres:///foo'.
Postgres uses empty-host URIs to indicate a Unix socket connection should be used instead... -
- signal.c: check stack overflow by SP
* signal.c (check_stack_overflow): raise SystemStackError if SP
register and fault address is in the same page, on x86 linux.
[EXPERIMENTAL]
* signal.c (check_stack_overflow): Don... -
03:00 PM Bug #9774: Net::HTTP failure to validate certificate
- I can confirm this bug on Mac OS X Mavericks with Ruby 2.1.1.
I actually didn't specify a ca_path, but imported the self-signed certificate into the Mac OS X keychain. The behaviour is the same however:
~~~
require 'net/http'
htt... -
02:43 PM Feature #9909 (Open): why shouldn't constant lookup check the nesting of module's name
- module A
module B
end
end
above, B will have access to A's constants
but below, it won't
module A::B
end
is there a reason why the nesting of the name should not be part of the constant resolution algorithm?
when add... -
02:05 PM Bug #9705: Installing ruby 2.1.1 fails while compiling (after: rvm install 2.1.1)
- I tried it via rvm, but it still does not work, see this thread on github:
https://github.com/wayneeseguin/rvm/issues/2850#issuecomment-45339676
Regards,
Carsten -
- * compile.c (private_recv_p): check by node type, instead of a
magic number.
* node.h (NODE_PRIVATE_RECV), parse.y (attrset_gen): remove
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46366 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Applied in changeset r46365.
----------
compile.c, parse.y: private op assign
* compile.c (iseq_compile_each), parse.y (new_attr_op_assign_gen):
allow op assign to a private attribute.
[ruby-core:62949] [Bug #9907] -
- * compile.c (iseq_compile_each), parse.y (new_attr_op_assign_gen):
allow op assign to a private attribute.
[ruby-core:62949] [Bug #9907]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46365 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * test/ruby/test_assignment.rb (test_assign_private_self): test
for r3509, assignment to private attribute is allowed iff its
receiver is literal `self`.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46364 b2dd03c8-39d4-4d8f-98f... -
- * test/ruby/test_assignment.rb (test_assign): split by assignment
forms.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46363 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * parse.y (aryset_gen): use `attr_receiver()`.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46362 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * node.h (NODE_PRIVATE_RECV): name a magic number, `self` as the
receiver of a setter method call.
* compile.c (private_recv_p), parse.y (attr_receiver): use the
named macro.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46361 ... -
- Applied in changeset r46360.
----------
io.c: not shorten buffer unless succeeded
* io.c (io_setstrbuf, io_read): should not shorten the given buffer until
read succeeds. [ruby-core:55951] [Bug #8625] -
- * io.c (io_setstrbuf, io_read): should not shorten the given buffer until
read succeeds. [ruby-core:55951] [Bug #8625]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46360 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- While testing an MRI rebuild to use bug fixed openssl libs, I get the following runtime fail during load of openssl ext.
MRI built on Win8.1 64bit using mingw-w64 32bit 4.9.0 toolchain and rubyinstaller recipes. Custom OpenSSL libs bu... -
- Seems fine.
06/05/2014
- 10:43 PM Bug #9873 (Closed): YAML conversion of empty symbol
- Applied in changeset r46358.
----------
* ext/psych/lib/psych/visitors/yaml_tree.rb: dump empty symbols with a
tag so that they can be parsed on input. [Bug #9873] [ruby-core:62825]
* test/psych/test_symbol.rb: test for change - 10:43 PM Revision 68a06368 (git): * 2014-06-06
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46359 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
10:42 PM Revision 3f6eafaa (git): * ext/psych/lib/psych/visitors/yaml_tree.rb: dump empty symbols with a
- tag so that they can be parsed on input. [Bug #9873] [ruby-core:62825]
* test/psych/test_symbol.rb: test for change
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46358 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
03:50 PM Bug #9907 (Closed): Abbreviated method assignment with private attr_writer/attr_reader does not work.
- This looks like a hole in the specification:
~~~ruby
private def foo=(*) end
public def foo; 0 end
self.foo = 42
self.foo += 42
# private method `foo=' called for main:Object (NoMethodError)
private :foo
self.foo += ... -
03:35 PM Bug #9906 (Rejected): duplicated require of '.so' files?
- Dear all,
We are trying to get rid of autoload;
We could require everything in the first line of the file but instead we
prefer to require only what we need, when we need it.
So, instead of
require "digest/md5"
def foo(n)
... -
- Fixed in 2.1.
-
09:48 AM Bug #9897: Build fails for fiddle with libffi version 3.1
- Yes, that works :) Thank you very much for the quick fix.
-
- `test/ruby/test_fiber.rb` fails on AIX.
~~~
[ 2/23] TestFiber#test_argument/ss/home/rayod/Dev/Contribution/ruby-trunk-blue1/test/ruby/test_fiber.rb:19: [BUG] machine_stack_cache size is not canonicalized
ruby 2.2.0dev (2014-06-02 tr... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46357 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * gc.c (gc_page_sweep): should not set, but add final_slots into
sweep_page->final_slots.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46356 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
06:55 AM Bug #7733: Segfault in ::JSON.parse(nil, :quirks_mode => true)
- I can reproduce this.
```
hsbt@chkbuild003:~$ ruby -rjson -e '::JSON.parse(nil, :quirks_mode => true)'
ruby 1.9.3p484 (2013-11-22 revision 43786) [x86_64-linux]
/usr/lib/ruby/1.9.1/json/common.rb:148: [BUG] Segmentation fault
```
... -
- ref. https://github.com/ruby/ruby/pull/549
-
- rake.rubyforge.org is sunset now.
I replaced rubyforge entry to github on rake.1 and removed dead links.
-
- * configure.in (jemalloc): check for the header regardless drop-in
libjemalloc is found, for `malloc_conf` declaration.
* version.c (ruby_show_version): show `malloc_conf` if set.
[Feature #9113]
git-svn-id: svn+ssh://ci.ruby-lang.o... -
- Applied in changeset r46354.
----------
configure.in, missing.h: jemalloc mangling
* configure.in (with-jemalloc): also check for header, for ABIs
which JEMALLOC_MANGLE is needed, i.e., Mach-O and PE-COFF
platforms. [ruby-core:629... -
- jemalloc prefixes "je_" to each public symbols at Mach-O (Mac OS X) and PE-COFF (Windows), so it can't be drop-in by default.
https://github.com/jemalloc/jemalloc/blob/dev/configure.ac#L494L498
Or you have to install it with `--wit... -
- I tried to build with --with-jemalloc option.
* OS X 10.9.4 pre-release
* jemalloc-3.6.0 via homebrew
* r46350
but it got build failure.
```
checking for malloc_conf in -ljemalloc... no
configure: error: jemalloc reques... -
- * configure.in (with-jemalloc): also check for header, for ABIs
which JEMALLOC_MANGLE is needed, i.e., Mach-O and PE-COFF
platforms. [ruby-core:62939] [Feature #9113]
* include/ruby/missing.h: include alternative malloc header to
... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46353 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46352 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46351 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
06/04/2014
-
- KOSAKI Motohiro <kosaki.motohiro@gmail.com> wrote:
> So, I'd suggest two phase action.
>
> 1. Commit your patch, but disable by default.
OK, r46349
> 2. Gather more use-case and performance data.
> Note: They should be re... - 07:34 PM Revision 045aef85 (git): * 2014-06-05
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46350 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
- 07:34 PM Revision 22d8480e (git): configure.in: add --with-jemalloc option
- * configure.in: add --with-jemalloc option
[ruby-core:62912]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46349 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- It's not a security issue but affects other platforms.
You need to make a 1GiB string on 32-bit platforms or on Windows, and a 4EiB string on other 64-bit platforms. -
02:49 PM Bug #9709: Large string causes SEGV with x64-mingw32
- Hi,
I'm seeking explanation if this [security issue] only applies to mingw, and if so, why, as the same code appears to run on other platforms as well.
Thank you,
Christian
-
- Rewrite finalizing logics with this type.
* gc.c (gc_verify_internal_consistency): verify zombie (finalizing)
objects count.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46348 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46347 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- Applied in changeset r46346.
----------
re.c: consider name encoding
* re.c (match_aref, rb_reg_regsub): consider encoding of captured
names, encoding-incompatible should not match.
[ruby-dev:48278] [Bug #9903] -
- Regexp#[]が引数の文字列やシンボルを考慮していません。
~~~ruby
key = "\xb1\xb2".force_encoding(Encoding::EUC_JP)
m = /(?<#{key}>.*)/.match("xxx")
p m[key.dup.force_encoding(Encoding::Shift_JIS)]
# => "xxx"
~~~ -
- * re.c (match_aref, rb_reg_regsub): consider encoding of captured
names, encoding-incompatible should not match.
[ruby-dev:48278] [Bug #9903]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46346 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * re.c (match_aref, rb_reg_regsub): reduce new strings creation
for exceptions.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46345 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Applied in changeset r46344.
----------
re.c: fix name with NUL
* re.c (match_aref): should not ignore name after NUL byte.
[ruby-dev:48275] [Bug #9902] -
- Regexp#[]に"\0"を含む文字列やシンボルを指定したとき、それ以降が無視されます。
~~~
$ ruby -w -e 'p(/(?<a>.*)/.match("foo")["a\0foo"])'
"foo"
$ ruby -w -e 'p Regexp.new("(?<foo\0bar>.*?)").match("xxx")["foo\0bar"]'
-e:1:in `[]': undefined group name reference: f... -
- * re.c (match_aref): should not ignore name after NUL byte.
[ruby-dev:48275] [Bug #9902]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@46344 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
07:05 AM Bug #9901 (Closed): Clarify documentation for Array#insert
- The current documentation for Array#insert is misleading regarding negative indices. The attached patch clarifies the documentation.
-
- It may have been fixed already, but not sure because the link doesn't reveal the revision and the platform.
-
04:10 AM Bug #9900 (Closed): Segmentation fault with recursive reference of a hash
- The following code raises a segmentation fault.
Hash.new{|h,k| h[k]}[:foo]
# => [BUG] Segmentation fault at 0x007fff5f3fff80
This bug was found on stackoverflow http://stackoverflow.com/questions/24028557/irb-crash-segment... -
- solved by r46312
-
- duprecated #9893