Activity
From 12/29/2013 to 01/04/2014
01/04/2014
-
10:46 PM Bug #9353 (Open): Struct setter methods now return the struct object rather than the value - 10:44 PM Bug #9353 (Closed): Struct setter methods now return the struct object rather than the value
- This issue was solved with changeset r44501.
Charlie, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* struct.c (rb_struct_set): return assigned value from setter ... -
08:59 PM Feature #9362: Minimize cache misshit to gain optimal speed
- On 01/04/2014 06:14 PM, Eric Wong wrote:
> Potential for future improvement:
>
> st_table and st_table_entry are both 48 bytes on 64-bit. That means
> those allocations may use 64-byte object slots and avoid going through
> ... -
- Sweet! Merged. Thank you.
On 01/04/2014 08:12 PM, Eric Wong wrote:
> Eric Wong <normalperson@yhbt.net> wrote:
>> OK, last update of the night :o I think everything is good on 32-bit...
>
> Gah, I decided to play on 64-bi... -
08:23 PM Feature #9362: Minimize cache misshit to gain optimal speed
- Eric Wong <normalperson@yhbt.net> wrote:
> OK, last update of the night :o I think everything is good on 32-bit...
Gah, I decided to play on 64-bit and fixed one more bug:
commit 87f13024862fe33bd2588013b833c64fbb2ef95a
... -
- Potential for future improvement:
st_table and st_table_entry are both 48 bytes on 64-bit. That means
those allocations may use 64-byte object slots and avoid going through
normal malloc. AFAIK, glibc malloc (and probably oth... -
- OK, last update of the night :o I think everything is good on 32-bit...
The following changes since commit 67a86ca145e50c77e36d95b16e58c8eea5edea6b:
Merge e5ed75dee0c334e8b14dcf987440500d1b70f80f into 8f04556111b25d336838f4... -
- Btw, I just pushed a few trivial fixes up (a few more failures below):
The following changes since commit 67a86ca145e50c77e36d95b16e58c8eea5edea6b:
Merge e5ed75dee0c334e8b14dcf987440500d1b70f80f into 8f04556111b25d336838f40a... -
- Eric Wong <normalperson@yhbt.net> wrote:
> 3 bits for embedded array/struct length does not seem to be enough :<
Six bits is enough for 32-bit and 64-byte cache line sizes. Hopefully
not clobbering anything else that's importa... -
- Hi, I noticed a trivial typo in array.c, and it fails building struct.c
and array.c on 32-bit x86 (with 64-byte cache line).
--- a/array.c
+++ b/array.c
@@ -28,7 +28,7 @@ VALUE rb_cArray;
static ID id_cmp, id_div, id_p... -
- Cool. I didn't expect the improvement for largely single-threaded
workloads. I'm not sure if it's feasible, but it might be
better to detect cache line size with:
sysconf(_SC_LEVEL1_DCACHE_LINESIZE);
At least on glibc-b... -
- Main features:
- Applies cleanly onto trunk,
- Passes tests,
- RUNS FASTER.
Detailed concepts, the patches, and benchmark results can be
obtained from: https://github.com/ruby/ruby/pull/495 -
07:39 PM Bug #9312: Build the ruby executable in bin/
- mkrunnable.rb creates symlinks only.
-
- Thanks for trying. This is probably specific to *BSD sockets
implementation, so I can't reproduce it at the moment.
Can you try reverting parts of r36944 which affect IO#write?
("avoid unnecessary select() calls before doing I... - 02:05 PM Bug #9356: TCPSocket.new does not seem to handle INTR
- Eric: Unfortunately your patch doesn't fix it, still getting the same ENOTCONN error.
-
- This might be a bug exposed due to r36944
("avoid unnecessary select() calls before doing I/O")
which I don't want to revert.
Does the following fix it?
--- a/ext/socket/init.c
+++ b/ext/socket/init.c
@@ -400,12 +400,6... -
02:29 PM Feature #7274: UnboundMethods should be bindable to any object that is_a?(owner of the UnboundMethod)
- Hi,
matz (Yukihiro Matsumoto) wrote:
> OK, I misunderstood something.
> ...
I agree. I can't really think of an actual use-case. I was asking for consistency's sake.
- 01:44 PM Revision 306c154c (git): * struct.c (rb_struct_set): return assigned value from setter method
- rather than struct object. [Bug #9353] [ruby-core:59509]
* test/ruby/test_struct.rb (test_setter_method_returns_value): add test
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44501 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
12:47 PM Revision 54f6b3fe (git): test/ruby/test_gc.rb: use GC.stat(:key) instead of GC.stat
- * test/ruby/test_gc.rb (TestGc#test_latest_gc_info): use
GC.stat(:key) instead of GC.stat.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44500 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44499 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * ext/tk/extconf.rb: use -rpath linker option instead of -R, to
fix build error with tk variant. incorporate from MacPorts.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44498 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
10:33 AM Bug #9117 (Closed): 43656 breakes openssl in stalic-ext ruby
-
12:57 AM Bug #9117: 43656 breakes openssl in stalic-ext ruby
- it works fine for me with 2.0.0-p353 and 2.1.0 - you can close this issue
-
10:16 AM Bug #9351 (Closed): Make Rinda::RingFinger work correctly on OpenBSD and NetBSD
- This issue was solved with changeset r44492.
Jeremy, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* lib/rinda/ring.rb (Rinda::RingFinger#make_socket): Use
ipv... -
-
- * ext/socket/socket.c (rsock_syserr_fail_host_port): use format flags,
'+' to inspect, ' ' to quote unprintables.
* ext/socket/socket.c (rsock_syserr_fail_path): ditto.
* ext/socket/socket.c (rsock_syserr_fail_raddrinfo): ditto.
git-s... -
- * ext/socket/socket.c (rsock_syserr_fail_host_port): add errno
argument version anduse rb_syserr_fail_str() instead of
rb_sys_fail_str() with restoring errno.
* ext/socket/socket.c (rsock_syserr_fail_path): ditto, and
rb_syserr_fai... -
08:22 AM Revision de7e488b (git): * vm.c (RubyVM::OPTS): get rid of a garbage character.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44495 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
06:52 AM Bug #9361 (Closed): gem install --ignore-dependencies fails to install local gems
- $ rvm ruby-2.1 do gem install --ignore-dependencies /Users/headius/projects/jruby/lib/ruby/gems/shared/cache/dicks-0.03.gem
ERROR: While executing gem ... (NoMethodError)
undefined method `full_name' for nil:NilClass
This is a... -
02:10 AM Bug #9321: rb_mod_const_missing does not generate a c-return event
- But when raising an exception from ruby, return events are emitted... Shouldn't this 2 cases be consistent?
-
02:10 AM Bug #8182: XMLRPC request fails with "Wrong size. Was 31564, should be 1501"
- Today I upgraded a ruby app that uses xmlrpc/client from 1.9.3-p374 to 2.0.0-p353, still not fixed in it…
-
- Forwarding problem from https://github.com/wayneeseguin/rvm/issues/2511 - the binary ruby used can be found here https://rvm.io/binaries/osx/10.6/x86_64/ruby-2.1.0.tar.bz2 (it should work on 10.6-10.9) versions ... the build was done on ...
-
01:23 AM Bug #9269: class variable is not being handled correctly
- ah, sorry, I saw the comment on the gist. It doesn't notify me by e-mail.
-
- Why do you think so? This is not my understanding. Could you please explain why the threads are not joining? I still think this is a bug. Please consider reopening it.
-
- ipv4_multicast_ttl option for portability.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44494 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 01:16 AM Revision 864e0c90 (git): * 2014-01-04
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44493 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- ipv4_multicast_loop option for portability. Patch by Jeremy Evans.
[ruby-trunk - Bug #9351]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44492 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
12:58 AM Bug #9359 (Rejected): indirect require of OpenSSL issues warning: assigned but unused variable - id
- =begin
On OpenSUSE:
(({ruby1.9 -wU -e "require 'resolv'"})) => (({/usr/lib64/ruby/1.9.1/openssl/ssl-internal.rb:91: warning: assigned but unused variable - id}))
(({ruby2.0 -wU -e "require 'resolv'"})) => (({/usr/lib64/rub...
01/03/2014
- 11:53 PM Bug #9357: TracePoint's c_return traces return from call to 'trace'
- Makes sense.
Thanks.
Will close this now.
On 3 January 2014 14:42, deivid (David Rodríguez) <
deivid.rodriguez@gmail.com> wrote:
>
> Issue #9357 has been updated by deivid (David Rodríguez).
>
>
> Yes, because traci... -
- Yes, because tracing is enabled after the 'c_call' and before the 'c_return' event.
- 11:23 PM Bug #9357: TracePoint's c_return traces return from call to 'trace'
- It doesn't cause any issues, however that doesn't seem to happen when one
traces on c_call. Here's the code:
https://eval.in/85228
Thanks.
On 3 January 2014 11:51, deivid (David Rodríguez) <
deivid.rodriguez@gmail.com> wr... -
- I think this is expected because tracing is enabled _inside_ the trace method, not after, nor before.
Does that behaviour cause any trouble? -
07:58 PM Bug #9357 (Closed): TracePoint's c_return traces return from call to 'trace'
- =begin
Hello,
In TracePoint, trace on 'c_call' doesn't trace the call made to 'trace' method, where as trace on 'c_return' does, is this expected behaviour?
Here's the code I used to reproduce it:
=== Code (c_return)
clas... -
10:39 PM Bug #9341: IMAPTest#test_imaps_with_ca_file test failed after 2014-01-01.
- r44488 and r44489 were backported to ruby_2_0_0 branch at r44491.
-
10:47 AM Bug #9341 (Closed): IMAPTest#test_imaps_with_ca_file test failed after 2014-01-01.
- This issue was solved with changeset r44488.
Heesob, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
test/net/imap/test_imap.rb: fix test failures due to expired ce... -
-
- I do not know how, but 2.1.0 and 2.0.0-p353 worked just fine without removing this `require` - can be closed I guess.
-
09:18 PM Bug #9259: "Could not find 'minitest': test suite errors
- This is still issue for ruby 2.1.0p0 (2013-12-25 revision 44422) [x86_64-linux]
-
- =begin
Hello,
In TracePoint class, if a particular introspection method is not supported then 'not supported by this event (RuntimeError)' is thrown.
=== Line event doesn't support raised_exception
TracePoint.trace(:line)... -
08:24 PM Bug #9352: [BUG] rb_sys_fail_str(connect(2) for [fe80::1%lo0]:3000) - errno == 0
- Thanks Eric, patch fixes the issue for me in ruby 2.2.0dev (2014-01-03 trunk 44490) [x86_64-darwin13.0]
-
- Thanks for testing. I can't say I like my patch, however, as
it's still error-prone and fragile.
How about some new functions to the C API:
rb_syserr_fail_str(VALUE mesg, int errnum);
rb_syserr_fail(const char *mesg, in... - 07:15 PM Bug #9352 (Open): [BUG] rb_sys_fail_str(connect(2) for [fe80::1%lo0]:3000) - errno == 0
- 07:12 PM Bug #9352 (Closed): [BUG] rb_sys_fail_str(connect(2) for [fe80::1%lo0]:3000) - errno == 0
- This issue was solved with changeset r44490.
Claudio, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* ext/socket/socket.c (rsock_sys_fail_host_port): save and res... -
- I can confirm Eric's patch fixes this [BUG] for me in faraday's test suite.
-
- I can't reproduce the issue, but I think it has to do with calling functions which may clobber errno before
finally calling rb_sys_fail.
Can you try the attached patch (or pull request) against trunk?
$ git pull git://80x24.org/ru... -
- Issue persists with ruby 2.2.0dev (2014-01-03 trunk 44488) [x86_64-darwin13.0]
-
- I'm seeing this crash attempting to connect to an http unreachable host (127.0.0.1:3000) through Excon, which in turn uses `connect_nonblock`.
I'm compiling ruby-trunk at the moment, to try on it as well. -
08:12 PM Bug #9262: global_method_cache should be configurable or grow automatically
- FYI it seems perf of method lookups has regressed in 2.1:
https://gist.github.com/SamSaffron/8232978
This makes this change particularly important - 07:29 PM Bug #9356 (Closed): TCPSocket.new does not seem to handle INTR
- TCPSocket.new does not seem to handle EINTR properly.
In the attached test script, I try to open a TCP connection to my server and make an HTTP request while a background thread continually sends a signal to the process.
This cause... -
- Hello,
I'd like to suggest renaming method_id (http://www.ruby-doc.org/core-2.1.0/TracePoint.html#method-i-method_id) to method_name in TracePoint class. There may be a particular reason for keeping it 'method_id', and I'd appreciate ... -
- The following code works on 2.0.0, but raises on 2.1.0:
require 'timeout'
class MyError < StandardError
def initialize(msg)
super
end
end
timeout(1, MyError){}
$ ruby -v
ruby 2.1.0p0-github (2013-12-31 revision 44488... - 06:31 PM Bug #9353 (Closed): Struct setter methods now return the struct object rather than the value
- ~ ruby-2.1.0 λ irb
>> Struct.new(:foo).new.send(:foo=, 123)
=> #<struct foo=123>
~ ruby-2.0.0-p247 λ irb
> ...
=> 123
This looks like a regression introduced by r41534, where rb_struct_set was changed:
- return pt... -
- Still observe the issue with ruby 2.1.0p0 (2013-12-25 revision 44422) [x86_64-linux]
-
05:58 PM Bug #9327: Why FileUtils.rmdir ignores Errno::ENOTEMPTY ?
- It is because we added :parents option in FileUtils.rmdir. See this ticket for detail: https://bugs.ruby-lang.org/issues/3178
I already created patch to improve this situation on that ticket. -
- > * Should string freeze-duping in Hash#[]= always dedup, regardless of
> whether the string has already been frozen or the Hash has
> compare_by_identity == true? Or else what combination of those
> conditions should lead to dedu... -
- I ran into a small snag so I think it will be useful to have my JRuby commits here.
First, a background commit... adding a weak dedup cache: https://github.com/jruby/jruby/commit/926ca89075a4a4c84592add729531189263c143f
It's not pa... -
- I've implemented something like this for JRuby 9k. Both the interpreter and the compiler pre-freeze and dedup the keys, allowing them to go straight into literal hashes.
Performance is significantly improved by doing this:
Before:
... -
02:38 PM Bug #9321: rb_mod_const_missing does not generate a c-return event
- @tmm1's suspicion was right: c-return is also missing in the case when you try to call a nonexistent method.
It seems that c-return will be missing when raising an exception from C land.
Which means my patch doesn't really solve th... -
- @ko1: Why do you think the second patch creates a performance problem?
Module#const_missing is an exceptional case. Am I missing something?
@tmm1: the stack tracking approach breaks down if the method id reported on return by ruby ... -
02:05 PM Bug #9309 (Closed): Crash while running tests
- fixed by r44458
-
06:42 AM Bug #9309: Crash while running tests
- Hi,
I had the same problem and reported in #9313.
I tested with r44487 and it's working now.
Thank you -
- fixed by r44458
-
- Noted that same issue is reported in issue #9309 and resolved with r44458
-
01:38 PM Bug #9351 (Closed): Make Rinda::RingFinger work correctly on OpenBSD and NetBSD
- The current Rinda::RingFinger code assumes that the IP_MULTICAST_LOOP socket option takes int, but in NetBSD and OpenBSD it takes unsigned char. Switch to using Socket::Option.ipv4_multicast_loop for better portability.
To get this t... -
- * test/net/imap/cacert.pem: generate new CA cert, since the last one
expired. [Bug #9341] [ruby-core:59459]
* test/net/imap/server.crt: new server cert signed with updated CA.
* test/net/imap/Makefile: add `ma... -
- OpenBSD is similar to NetBSD in that the IP_MULTICAST_LOOP and IP_MULTICAST_TTL socket options take unsigned char and not int arguments. So replace
#ifdef __NetBSD__
with
#if defined(__NetBSD__) || defined(__OpenBSD__)
-
01:23 PM Feature #9347: Accept non callable argument to detect
- I think, if it was possible, a call like this `ary.detect(:sym)` would make more sense by using case equality for comparison instead of #call. It would be useful with regexes for example. There's a ticket proposing this but I can't seem...
-
11:38 AM Feature #9347: Accept non callable argument to detect
- I am afraid that kind of conditional behavior could error prone (e.g. accidental misbehavior when one forget having #call in argument object). I'd rather have #detect with keyword argument, like ary.detect(ifnone: :notfound).
Matz.
-
- Currently, the only argument that `Enumerable#detect` accepts is a callable object.
Shouldn't we accept non callable objects too?
[42].detect(:not_found){} # => NoMethodError: undefined method `call' for :not_found:Symbol
... -
- Btw, I took some time to work on this further. Only _very_ lightly
tested (make check passes)
# GC::Profiler.report doesn't even show anything with this patch applied
# because GC never happens.
h = {}
GC::Profiler.enable
... - 10:12 AM Revision 8f045561 (git): * ext/socket/socket.c (rsock_sys_fail_host_port): save and restore errno
- before calling rb_sys_fail_str to prevent [BUG] errno == 0.
Patch by Eric Wong. [ruby-core:59498] [Bug #9352]
* ext/socket/socket.c (rsock_sys_fail_path): ditto
* ext/socket/socket.c (rsock_sys_fail_sockaddr): ditto
* ext/socket/socke... -
- Could you try to test with ruby trunk?
-
- I created backport issue #9349 for Ruby 2.1.
-
03:49 AM Bug #9326: No superclass method in ruby 2.1
- Confirmed that it works on trunk. Looks like the fix was r44458. It still fails on ruby_2_1.
I can't explain the 'ruby -v' discrepancy. I did a fresh download of 2.1.0 and confirmed that's what it reports. -
- Could you try to trunk version of Ruby?
-
- Could you try to trunk version of Ruby? It seems to be fixed by r44458 probably.
-
- duplicated #9345
-
- We only install system-level signal handlers during init (before any
threads or user code is running), so there's no need to repeatedly
enable/disable the interrupts at that time. We also do not change
system-level sig handlers once ... -
- When using the --with-tcllib=tcl85 -with-tklib=tk85 configure options, ruby still tries to use -ltcl -ltk as linker options for the tk extension (instead of -ltcl85 -ltk85), which causes the tk extension not to build. This is due to a b...
-
- We only install system-level signal handlers during init (before any
threads or user code is running), so there's no need to repeatedly
enable/disable the interrupts at that time. We also do not change
system-level sig handlers once ... -
- On 1 Jan 2014, at 13:19, Avdi Grimm <avdi@avdi.org> wrote:
> I believe I can make a solid case for this as an accidental feature rather than a bug. Should I open a new ticket to do so?
I think so. -
04:15 AM Misc #8835: Introducing a semantic versioning scheme and branching policy
- vo.x (Vit Ondruch) wrote:
> Do I understand correctly that patch level won't be anymore part of the released tarball name? E.g. retrofitting semver on Ruby 2.0, the first released bugfix release had been ruby-2.0.0-p195.tar.gz, so with ... -
- Do I understand correctly that patch level won't be anymore part of the released tarball name? E.g. retrofitting semver on Ruby 2.0, the first released bugfix release had been ruby-2.0.0-p195.tar.gz, so with the semver, it would have be ...
-
- Sorry, but this is not Mingw specific issue. I am facing the same issue on Fedora [1]. The test passes on rubyci.org, because there seems to be explicitly specified -msse2 configuration option for some reasons (may be to make this test p...
-
- That code just starts threads but wait nothing.
Makes no sense. -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44489 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * test/net/imap/cacert.pem: generate new CA cert, since the last one
expired. [Bug #9341] [ruby-core:59459]
* test/net/imap/server.crt: new server cert signed with updated CA.
* test/net/imap/Makefile: add `make regen_certs` to automat...
01/02/2014
-
09:17 PM Bug #9344 (Closed): warning origin incorrect with instance_eval
- File 'test.rb' contains:
~~~ruby
$-w = 2
code = <<-RUBY
x = /]]/
y = /[a-z]+*/
RUBY
instance_eval code, 'foo.rb'
~~~
The execution reports:
~~~
foo.rb:1: warning: regular expression has ']' without escape: /]]/
test... -
- In rblineprof, I maintain a stack of CALL events. When a RETURN event comes in, CALL entries on the internal stack are popped until one matching the RETURN is found.
https://github.com/tmm1/rblineprof/blob/32795afc748f2f51cf76ce20ef... -
- We were accessing the wrong array and trying to wake up
elements stored in the queue :x
$ git pull git://80x24.org/ruby.git szqmaxset-wakeup
The following changes since commit d1cc0ebb38a23feb37bb16ff2df3137c3cead069:
mkrunna... -
- I agree this is a problem. This also affects 2.1.0 and trunk (where
SizedQueue is C). -
03:53 PM Bug #9342: [PATCH] SizedQueue#clear does not notify waiting threads in Ruby 1.9.3
- Looks like this is also a problem in 2.0.0, although the implementation is a little different. Attaching patch for 2.0.0 too although it's probably clear.
-
- In Ruby 1.9.3, when SizedQueue#clear is called, it empties the queue but does not notify waiting threads that the queue is empty. This typically leads to deadlock when the queue is full.
For example:
sq = SizedQueue.new(1)
sq << 1... - 03:10 PM Revision 834ea58a (git): * 2014-01-03
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44487 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
03:10 PM Revision 1cc709a8 (git): * ext/bigdecimal: update class method call style from :: to .
- in documentation and usage.
* ext/bigdecimal/lib/bigdecimal/math.rb: [DOC] fix examples values.
Computations were made using ruby 2.0.0p247 to ensure
no effect of the recent BigDecimal bug.
* ext/bigdecimal/sample/nlsolve.rb: fix ind... -
-
12:12 PM Bug #9341: IMAPTest#test_imaps_with_ca_file test failed after 2014-01-01.
- You must recreate the certificates.
Refer to http://www.lwithers.me.uk/articles/cacert.html -
- Yes, I know it, but not how to refresh it.
-
- The test IMAPTest#test_imaps_with_ca_file failed with the following error after 2014-01-01 on all platforms in rubyci.org
IMAPTest#test_imaps_with_ca_file [D:/tmp/mswin-build20140102-4508-1arnv64/ruby/test/net/imap/test_imap.rb:74]:
... -
- If you want to extract from a tarball, you need 'p' option.
I recommend not to include drive letter and even use of environment dependent path.
As windows binaries are always relocatable, I configure with --prefix=/. and use DESTDIR ... -
- Hello,
Even that you're using MSYS bash and make to trigger the compilation process, please use Windows-style paths for the installation directory.
configure --prefix=C:/some/folder
Depending on the way the source code was extra... -
08:30 AM Bug #9339 (Closed): Can't Compile Ruby 2.1 without baseruby (Win7, MSYS/MinGW)
- On my Win7 System in an MSYS/MinGW-environment (more precisely the 'mingw-builds' version) I have been trying to compile Ruby 2.1 from source.
I unpacked and entered:
configure --prefix=/c/path/to/my/ruby/install
make
The f... -
09:11 AM Bug #9340 (Closed): Document order related behavior in Array#uniq
- The behavior of Array#uniq without a block in Ruby 2.1.0 is inconsistent with the behavior of Array#uniq in Ruby 2.0.0, and with the behavior of Array#uniq with a block in Ruby 2.1.0.
Array#uniq without a block in Ruby 2.1.0 returns t... -
07:29 AM Revision 72385b04 (git): * io.c (io_fwrite): freeze converted str.
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44485 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
06:54 AM Bug #9334: activesupport-4.0.2/lib/active_support/dependencies.rb:229: [BUG] Segmentation fault at 0x00000000000018
- Noted that same issue, I think, is reported in issue #9309
-
06:23 AM Bug #9335: dynamic rescue regression in Ruby 2.1
- I believe I can make a solid case for this as an accidental feature rather
than a bug. Should I open a new ticket to do so?
On Wed, Jan 1, 2014 at 1:28 AM, nobu (Nobuyoshi Nakada)
<nobu@ruby-lang.org>wrote:
>
> Issue ... -
- This issue was solved with changeset r44483.
Heesob, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
eval.c: extra modifier
* eval.c (rb_longjmp): remove an extra ... -
- * tool/mkrunnable.rb: DLL needs placed at same directory as
executables on Windows.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44484 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 12:19 AM Feature #9336 (Closed): [PATCH (trivial)] remove rb_hash_keys from internal.h
- This issue was solved with changeset r44478.
Eric, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* hash.c (rb_hash_keys): make rb_hash_keys() static.
it is no l... -
01/01/2014
-
- The revision r44473 introduced the following error.
compiling eval.c
eval.c
eval.c(558) : error C4028: formal parameter 3 different from declaration
NMAKE : fatal error U1077: '"C:\Program Files (x86)\Microsoft Visual Studio 12.0\V... -
08:49 PM Bug #9337 (Closed): [BUG] Segmentation fault at 0x00000000000018 ruby 2.1.0p0 (2013-12-25 revision 44422) [x86_64-linux]
- Tried to switch to ruby 2.1.0 on my rails app that runs on a linux server.
During deployment when executing rake assets:precompile I got a Segmentation fault error.
Tried to to the same thing locally (on a mac os system) and I got ... -
- This is no longer needed since r43969
If you prefer: git pull git://80x24.org/ruby.git rb_hash_keys-static
The following changes since commit ad0980504b355485b848395f1bf4f067e6b23d3d:
* 2014-01-01 (2014-01-01 02:12:43 +0000)
... -
- * eval.c (rb_longjmp): remove an extra modifier from the forward
declaration to match the actual definition. [ruby-core:59451]
[Bug #9338]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44483 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * ext/dbm/dbm.c (fdbm_fetch): yield dup of keystr, to make it shared
and get rid of use of uninitialized variable.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44482 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- This issue was solved with changeset r44476.
Wataru, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
configure.in: LDFLAGS and DLDFLAGS for opt-dir
* configure.in:... -
- ALLOCA_N() and rb_ary_tmp_new().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44481 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44480 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- It had been allowed on 1.9.3 accidentally.
In other words, it's a fixed bug.
-
12:47 PM Bug #9335: dynamic rescue regression in Ruby 2.1
- Some other person on the internet also took the time to track this change in behaviors over time:
https://gist.github.com/gmallard/7034128 -
- This seems to apply to 2.0 as well. Here's the commit found by git bisect:
commit 93b6f8d6195564d0884ee00f536a951f4c26288c
Author: ko1 <ko1@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Wed Aug 8 07:52:19 2012 +0000
* compil... -
- The following type of rescue block appears broken:
begin
raise 'hello'
rescue ->(e) { true }
end
In Ruby 1.9.3p484, this exits with a normal exit code and no output.
On Ruby 2.1.0, it instead yields:
/tmp/badin21.rb:4:in... - 03:19 PM Revision edc5eb0d (git): * 2014-01-02
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44479 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- it is no longer used from array.c since r43969.
the patch is from normalperson (Eric Wong).
[ruby-core:59449] [Feature #9336]
* internal.h: remove definition of rb_hash_keys().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@4447... -
- * test/ruby/test_transcode.rb (test_valid_dummy_encoding): add
assertion messages and suppress a warning.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44477 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- This issue was solved with changeset r44474.
Heesob, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
configure.in: use SSE2
* configure.in: use SSE2 instructions f... -
- Thanks! I will check this out soon
-
- * configure.in: reset LDFLAGS and DLDFLAGS for opt-dir again after
LIBPATHFLAG and RPATHFLAG are set. [ruby-dev:47868] [Bug #9317]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44476 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- I have just upgraded to ruby 2.1.0 and running in Rails foreman (unicorn) and received a long error report, which I quite frankly do not understand, attached an output and diagnostics report
-
- I like the second patch, but now I'm not so sure about this technique.
Technically, const_missing never returns. So emitting a c_return event seems wrong. In fact, a raise event should be emitted from inside const_missing.
The "rig... - 02:12 AM Revision ad098050 (git): * 2014-01-01
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44475 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- * configure.in: use SSE2 instructions for drop unexpected
precisions. [ruby-core:54738] [Bug #8358]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44474 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
12/31/2013
-
10:41 PM Bug #9332 (Closed): win32/Makefile.sub malconfiguration
-
- win32/Makefile.sub defines HAVE_FSEEKO because win32/win32.c defined the function.
However since r40661 the function was removed from win32/win32.c.
-
-
- win32/Makefile.sub defines HAVE_FSEEKO because win32/win32.c defined the function.
However since r40661 the function was removed from win32/win32.c.
It must be removed from Makfile.sub too. -
10:36 PM Bug #9326: No superclass method in ruby 2.1
- Seems to be fine on ruby_2_1, after r44410 (merging r44380).
I confirmed NoMethodError is raised on 2.1.0 (r44340) and not on ruby_2_1 (r44443).
But 'ruby -v' of this ticket is revision 44422 (> 44410).
Really test.rb fails on r4442... -
- Does it work on the trunk?
-
- Runing ruby 2.1 plus nobu's patch from issue 9315. Without the patch it segfaults.
Found this testing testing test-redef with ruby 2.1. Works on 2.0.0-p353 and fails on 2.1.0. When the replacement method calls the alias directly it... -
09:24 PM Bug #9321: rb_mod_const_missing does not generate a c-return event
- The last patch is not acceptable because of performance reason.
I'll think about it.
-
- Here's the new patch. I have moved the hook to rb_frame_pop().
The new patch makes sure that the method names sent along with call and return events are identical.
But I'm not sure whether the current ruby behavior in the presence ... -
- The patch has some issues with redefined Module#const_missing.
I'm in the process of preparing a new one. -
- I can reproduce with 2.1.0p0 but not with r44459 (trunk).
And git-bisect said it hasn't occurred after r44455.
I guess the ticket is a duplication of #9315. -
05:37 PM Misc #9331 (Closed): Add examples to Set#intersect? and Set#disjoint?
-
- This issue was solved with changeset r44469.
Hal, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
mkmf.rb: expand RUBY_SO_NAME
* lib/mkmf.rb (RbConfig): expand RUB... -
- I can reproduce this with Ubuntu 9.04 also.
Here is a patch:
diff --git a/enc/Makefile.in b/enc/Makefile.in
index a64785d..378f4a5 100644
--- a/enc/Makefile.in
+++ b/enc/Makefile.in
@@ -28,6 +28,7 @@ BUILTIN_ENCS = ascii.c us_a... -
11:36 AM Bug #9329: ./configure --enable-shared --enable-relocatable && make fails
- I can reproduce this with `./configure --enable-shared`. My platform is Fedora Linux 19 x86-64.
-
- Not --enable-relocatable, --enable-load-relative.
-
- What's your platform?
-
- Steps to Reproduce:
$ ./configure --enable-shared --enable-relocatable
$ make
Actual Result:
compiling ./enc/encdb.c
linking encoding encdb.so
/usr/bin/ld: cannot find -l-lpthread
collect2: error: ld returned 1 exit status
-
- * eval.c (rb_f_raise): add cause: optional keyword argument.
[ruby-core:58610] [Feature #8257] [EXPERIMENTAL]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44473 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44472 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44471 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- commit 52525b673669019dc3c03474e613937fd3587187
Author: Eric Wong <e@80x24.org>
Date: Tue Dec 31 02:09:27 2013 +0000
socket: avoid redundant fcntl calls for FD passing
If MSG_CMSG_CLOEXEC works on the first recvmsg ... -
- * lib/mkmf.rb (init_mkmf): fix for 1.8 baseruby. RbConfig of 1.8
did not have CXXFLAGS.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44470 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- Running `make runnable` appears to add a `bin/goruby` hardlink to `../goruby`.
-
- $ ./configure --enable-shared --enable-load-relative
$ make
...
making enc
make[1]: Entering directory `/vault/0/src/ruby/trunk'
linking encoding encdb.so
/usr/bin/ld: cannot find -l-lpthread
collect2: error: ld returned 1 exit st... -
- To make relocatable, configure needs --enable-shared and --enable-load-relative.
-
- Test on trunk (r44467)
$ ./configure
$ make
$ make runnable
make: *** No rule to make target `un-runnable', needed by `runnable'. Stop. -
- Already.
-
- Add an uninstall target to the Makefile, so ruby can be uninstalled if installed manually.
-
- It is pointless to check if O_CLOEXEC works after every single
open() call. If the kernel supports O_CLOEXEC for one open(), it is
supported for every open() syscall.
I mainly made this to make my strace less noisy. Also similar c... -
06:59 AM Bug #9327 (Closed): Why FileUtils.rmdir ignores Errno::ENOTEMPTY ?
- It catches Errno::ENOTEMPTY and Errno::ENOENT and just does nothing.
Most of other FileUtils methods have option "force" which is false by default
but rmdir behaves as if it is true implicitly.
Imho it looks unexpected.
-
- * lib/mkmf.rb (RbConfig): expand RUBY_SO_NAME for extensions
backward compatibility. [ruby-core:59426] [Bug #9329]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44469 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- * tool/mkconfig.rb (RbConfig): version numbers are also needed by
RUBY_SO_NAME.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44468 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- I created the following file, but I'm not sure if we should add anything else. https://gist.github.com/zzak/8187842
The old 1.9 file which converted these keywords to methods and added expanded documentation for each one is nice. I'm ... -
-
- Thanks for the report, I can put something together for this soon
-
03:30 AM Bug #9313: Segmentation fault with active_record
- Possible duplicate of https://bugs.ruby-lang.org/issues/9315
-
01:28 AM Feature #9278: Magic comment "immutable: string" makes "literal".freeze the default for that file
- > Perhaps if you show concrete gains in a large code-base like rails the case will be more compelling.
Working on it! First to see how many tests actually fail from trying to mutate string literals, then to get some general performan... -
- Hello Nobu, Usa,
This is still happening in 2.1.0 release, and the test is blocking me from releasing RubyInstaller.
We need a response on this, is the test valid or not? can be ignored or the issue needs to be fixed?
Thank you.
12/30/2013
-
11:39 PM Bug #9210 (Closed): Adding extra example in docs
- This issue was solved with changeset r44463.
Steve, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* variable.c: Adding extra example in docs. [Bug #9210] -
- I propose addition of IO#writev.
It enables gather output from multiple buffers.
If writev(2) is not available, IO#writev uses Array#join to emulate atomic write. -
- rb_frame_pop is called in only one place, and that's the one which the patch addresses.
So yes, we could move the code to rb_frame_pop(), if we change it's signature.
I'm not sure, under which circumstances the problem can show it'... -
- Thanks for adding the test.
Seems like `rb_frame_pop()` should probably emit the event, but I'm not sure how to tell klass/mid in there.
BTW, doesn't this problem happen anytime an exception is raised? One or more stack frames can ... -
- I believe this bug also exists in 1.9.3 and 2.0.0.
Any chance to backport?
-
- added tests to patch.
-
- I believe the attached patch fixes the problem.
-
- We have had an issue reported for ruby-prof where execution times were attributed incorrectly in the call graph.
It turned out that the problem is caused by a missing c-return event for Module#const_missing.
ruby-prof simulates the... -
09:16 PM Revision 41deb12e (git): fix previous commit
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@44467 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- the behavior is under 2.2.0
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_1@44466 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
08:12 PM Feature #9322 (Closed): method_defined? family of of methods should support the exclusion of ancestors
- getting an array of all instance methods (instance_methods(false)) and scanning through it for a match seems like a poor substitute
- 07:48 PM Revision d6c7ab9f (git): * 2013-12-31
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44465 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44464 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- Can you try to reproduce this with trunk? This issue should be resolved with r44458 (dupe of #9315).
-
- >> It seems like it would be logical for the VM to automatically
> ...
I implemented such an optimization for Hash#[]= in https://bugs.ruby-lang.org/issues/9188#note-8
A similar patch could be developed for the aref case, but I'm not... -
- "jacknagel (Jack Nagel)" <redmine@ruby-lang.org> wrote:
> It seems like it would be logical for the VM to automatically
> freeze/dedup string literal keys in hash literals and hash lookups. Is
> there a reason that can't be done?
... - 02:16 PM Feature #9278: Magic comment "immutable: string" makes "literal".freeze the default for that file
- It seems like it would be logical for the VM to automatically freeze/dedup string literal keys in hash literals and hash lookups. Is there a reason that can't be done?
-
- colindkelley (Colin Kelley) wrote:
> > * A magic comment should not completely change the semantics of a literal type. Encoding magic comments do not suffer from the same issue since they only change how the bytes of the strings are int... -
- Thanks for the quick reply, Charles. Here are my thoughts:
> * A magic comment should not completely change the semantics of a literal type. Encoding magic comments do not suffer from the same issue since they only change how the byte... -
- This issue was solved with changeset r44462.
Krists, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
encoding.c: mask dummy flags
* encoding.c (`must_enci... -
04:14 PM Bug #9320 (Closed): Nightly Snapshot is not found
- This issue was solved with changeset r44460.
Heesob, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
* tool/make-snapshot: needs CXXFLAGS. [ruby-core:59393][Bug #9... -
- >phasis68
Thank you for your report. 404 error of ruby-lang.org is caused by r44459 too. -
- If I click the Nightly Snapshot link on the http://www.ruby-lang.org/en/downloads/, It raised "404 Not Found" error.
When I try to build the latest snapshot,
# ruby tool/make-snapshot /tmp/snapshot/ trunk
Exporting trunk@44459
E... -
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44463 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- I'm still against this (and having aliases at all for standard).
It makes code harder to search/read, especially when they have
short names like "f". -
- I'd like to see this get into 2.2, and it sounds like Matz wants it too.
-
- * encoding.c (must_encindex, rb_enc_from_index, rb_obj_encoding): mask
encoding index and ignore dummy flags. [ruby-core:59354] [Bug #9314]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44462 b2dd03c8-39d4-4d8f-98ff-823fe69b080e - 07:14 AM Revision 8e0013ea (git): * 2013-12-30
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44461 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
- git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44460 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-
05:11 AM Feature #9319: Add constant RUBY_KEYWORDS or a similar way to output which keywords exist at runtime
- Oops, I meant "IRC", not "IRB", but yeah from within running IRB it would be also nice of course. :)
-
- Hi,
Would it be possible to list all keywords in Ruby, from within a running program?
Like RUBY_KEYWORDS or similar?
Reasoning: On IRB someone asked me if "in" is a keyword. I was not sure so I googled
because I could not remem... -
- @tmm1 Thanks for the feedback!
If anyone else has any pointers on this, its appreciated! -
- Seems good, thanks nobu! <3
-
- (13/12/29 8:45), Zachary Scott wrote:
> You mean manually?
By the "Mail to issue" link. -
- Indeed, I did not know of $~[:b] either. Quite nice.
-
- ko1: I have no problem whatsoever with backtrace_locations being read-only now and forever. I'm glad to see this made it into 2.1 as at least a basic feature.
-
- Sorry this didn't get into 2.1 and I was unable to review. December was probably too late to get it in anyway.
Nobu's patch looks fine. If other ruby-core folks really want to keep Weakref as-is for compatibility and introduce a new t... -
- Hello,
r44175 and r44179 were already backported to ruby_2_0_0 branch at r44345.
But I cannot reproduce SEGV on 2.0.0-head(r44393) with testcase in r44455.
I wonder if I should backport r44458 to ruby_2_0_0. Any idea? -
- r44175 and r44179 were backported to ruby_2_0_0 at r44345, and additionally r44198 was backported at 44366.
See #9236 too. -
- Unfortunately it doesn't look like anything other than $! logic for this made it into 2.1. I was hoping we'd get either the constructor or a one-time-only #cause= but there was still some debate about which way to go.
Since the base #...
12/29/2013
-
- Adding .freeze to the frequently-used strings has the same result for me, and it doesn't also gather up other strings that may not need/want to be immutable.
https://gist.github.com/ColinDKelley/8156708
I have the same objections t... -
- With the 'immutable: string' magic comment, this benchmark runs 1.6X faster (that is, in 61% of the time) compared to stock Ruby 2.1.0.
https://gist.github.com/ColinDKelley/8156708 -
- +1. Some notes:
* To add -I/opt/include and -L/opt/lib, use --with-opt-dir=/opt
* Do not override CFLAGS. If CFLAGS is provided, it is used as a *replacement* for default flags (i.e., '-ggdb3 -O3' will not be used unless you specify ... -
11:26 AM Bug #9317 (Closed): --with-opt-dirがLDFLAGS, DLDFLAGSに反映されない
- ./configureに--with-opt-dirで与えたパスがCPPFLAGSには設定されますが、
LDFLAGSとDLDFLAGSに設定されません。
opt-dirの処理をしている時点で、LIBPATHFLAGやRPATHFLAGに値がまだないためだと思います。
trunkと2.1で発生しています。2.0では起きません。 -
09:43 AM Revision e8792774 (git): * lib/mkmf.rb (configuration): Make CXXFLAGS customizable.
- Patch by Kohei Suzuki (eagletmt). [Fixes GH-492]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@44459 b2dd03c8-39d4-4d8f-98ff-823fe69b080e -
- This issue was solved with changeset r44455.
Jan, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.
----------
vm_insnhelper.c: missing super in method module
* vm\_in... -
12:38 AM Bug #9315 (Assigned): Segfault when calling super in rspec mocks with and_call_original [linux]
- Good catch. Here is a smaller example:
~~~ruby
module A
def foo
super
end
end
class B
include A
end
B.new.method(:foo).call
~~~
I think r44179 caused this issue.
--
Yusuke Endoh <mame@tsg.ne.jp> -
-
- * proc.c (mnew_from_me): keep iclass as-is, to make inheritance
chain consistent. [ruby-core:59358] [Bug #9315]
* proc.c (method_owner): return the original defined_class from
prepended iclass, instead.
* vm_insnhelper.c (vm_searc... -
12:58 AM Bug #8591: [BUG] Segmentation fault ruby 2.0.0p195 (2013-05-14) [x64-mingw32]
- Let's move the conversation to https://github.com/brianmario/mysql2/issues/469