19938-21072

19748-21854 subjects 20151-20712

Fibers in 1.8
19938 [rubytalk@tm ] Are there any plans to backport Fiber to ruby 1.8?
19948 [nobu@ru y- a] Fiber in 1.9 equals to Thread of 1.8.
+ 19975 [ko1@at ot ne] nearly equals.  Fiber + preemption = 1.8 Thread.
+ 19999 [rubytalk@tm ] How can I implement the Fiber API (Fiber.yield and Fiber#resume) in
  20000 [binary42@gm ] It is possible to do something like Fibers quite easily with callcc
  20003 [shyouhei@ru ] I'm sorry to inform you that 1.8 threads are implemented on top of
  20004 [binary42@gm ] 1.8 threads do have efficiency problems as well but 1.9 Fibers seem to
  20016 [shyouhei@ru ] Ko1 explained me about this.  According to him, 1.8 threads are slower
  20017 [pbrannan@at ] 1.8 threads do a lot of memory copies on context switch, but they are
  20024 [binary42@gm ] From what I understand, the NeverBlock [0] has used 1.9 Fibers to
  + 20027 [pbrannan@at ] If the GC is invoked while in another thread, the calculated stack end
  + 20031 [rubytalk@tm ] NeverBlock uses my pure ruby implementation
    20037 [rogerpack200] Wow that's intense :)

unable to search on redmine
19944 [rogerpack200] searching redmine for "leak" yields 0 results.  Is that expected?

[Bug #744] memory leak in callcc?
19945 [redmine@ru y] Issue #744 has been updated by Roger Pack.
+ 19967 [kbloom@gm il] I can't reproduce http://redmine.ruby-lang.org/issues/show/649
+ 19968 [brent@mb ri ] Roger,
  19969 [duerst@it ao] I haven't read the spec, but my guess (having worked on other specs)
  19970 [brent@mb ri ] Martin,
  19972 [kurt@ca hn t] ...

[Bug #755] select in windows errantly accepts too many fd's
19946 [redmine@ru y] Bug #755: select in windows errantly accepts too many fd's

windows linking question
19959 [rogerpack200] It turns out that if I compile an extension (win32) that uses "select"

[Bug #755] select in windows errantly accepts too many fd's
19960 [redmine@ru y] Issue #755 has been updated by Roger Pack.

ext/curses ESCDELAY path
19961 [support@co x] ...

any ideas on why quoted parameters fail in mingw?
19962 [rogerpack200] Anybody have any thoughts why
19964 [luislavena@g] I'm still trying to see why this happens to you and nobody else :-)
19973 [rogerpack200] Interesting.
19974 [luislavena@g] All the XP boxes has SP3?
20001 [rogerpack200] yeah all SP3
20002 [luislavena@g] You forgot to include in your previous replies the console and code
20030 [rogerpack200] Here they are...

ruby "[BUG] " and backtrace of native function call - addr2line - useless info
19965 [radek.bulat@] SSBhdHRlbXB0IHRvIHVzZSBydWJ5MS45IGZyb20gdHJ1bmsgdG8gaGF2ZSBmdW4gd2l0aCBpdCBv
19976 [ko1@at ot ne] 1) Same binary?
19977 [radek.bulat@] PiAxKSBTYW1lIGJpbmFyeT8KWWVzLCBpdCdzIHRoZSBzYW1lLiBJIGRvdWJsZSBjaGVja2VkIHdp
19978 [radek.bulat@] Pj4gQlRXIGl0J3Mgc2VlbXMgR0MgcHJvYmxlbS4gIERvZXMgbm9rb2dpcmkgdXNlIEMgZXh0ZW5z

NULL pointer emerging from empty regexp match!?
19971 [jens.wille@u] recently, there were some weird segfaults with Unicode [1] in our
20018 [matz@ru y- a] unicode module needs update.  In the past RSTRING_PTR() never be NULL,
+ 20019 [pbrannan@at ] It seems like this could break a lot of extensions.  Could RSTRING_PTR
+ 20072 [jens.wille@u] ok, i contacted yoshida about the issue. thank you for clearing this up.

[Bug #760] Net::FTP does not work when "mathn" is loaded.
19979 [redmine@ru y] Bug #760: Net::FTP does not work when "mathn" is loaded.
19981 [transfire@gm] by-lang.org/issues/show/760

[Bug #760](Closed) Net::FTP does not work when "mathn" is loaded.
19980 [redmine@ru y] Issue #760 has been updated by Yukihiro Matsumoto.

[patch] Net::FTP SOCKS support is broken.
19982 [alan.wayne.j] ...
19984 [matz@ru y- a] Just checked in.  Thank you!

[Bug #761] Identical docs for Kernel#autoload? and Kernel#autoload
19983 [redmine@ru y] Bug #761: Identical docs for Kernel#autoload? and Kernel#autoload

[Bug #726](Closed) erb references $KCOD
19985 [redmine@ru y] Issue #726 has been updated by Yukihiro Matsumoto.

[Bug #691](Closed) Time::zone_utc? does not follow rfc2822
19986 [redmine@ru y] Issue #691 has been updated by Yukihiro Matsumoto.

[Bug #731](Feedback) Net::HTTPHeader []= method does not work as expected
19987 [redmine@ru y] Issue #731 has been updated by Yukihiro Matsumoto.

[Bug #691] Time::zone_utc? does not follow rfc2822
19988 [redmine@ru y] Issue #691 has been updated by Yukihiro Matsumoto.
20005 [duerst@it ao] I can see no other way to interpret this.

[Bug #743] Socket.gethostbyname returns odd values
19989 [redmine@ru y] Issue #743 has been updated by Alan Johnson.
20136 [matz@ru y- a] It's implemented using getaddrinfo because at the time of coding (long
20154 [alan.wayne.j] (Replying via mailing list because the topic has strayed from that of
20156 [matz@ru y- a] It's to help IPv6 transition.  Ordinary programmers still tend to use

[Bug #763] Mini-unit should not auto-run tests when exiting due to an exception
19990 [redmine@ru y] Bug #763: Mini-unit should not auto-run tests when exiting due to an exception

[Bug #764] Undefined method `chomp` while requiring logger.rb
19991 [redmine@ru y] Bug #764: Undefined method `chomp` while requiring logger.rb

[Bug #764] Undefined method `chomp` while requiring logger.rb
19992 [redmine@ru y] Issue #764 has been updated by Paul Brannan.

[Bug #765] Mismatched indentation warning in minitest/unit.rb
19993 [redmine@ru y] Bug #765: Mismatched indentation warning in minitest/unit.rb

[BUG?] rdoc/ri on solaris 8 (i386 vs sparc)
19996 [bwalton@ar s] I'm attempting to package 1.8.7p72 for solaris as part of the
20034 [rogerpack200] I assume they both use the same version of rdoc?  Do they both
20078 [bwalton@ar s] Well, the process I'm using to build everything is managed by GAR, so

[Bug #764](Closed) Undefined method `chomp` while requiring logger.rb
19997 [redmine@ru y] Issue #764 has been updated by Yukihiro Matsumoto.

[Bug #765](Closed) Mismatched indentation warning in minitest/unit.rb
19998 [redmine@ru y] Issue #765 has been updated by Yukihiro Matsumoto.

Re: OpenSolaris support  Was: Re: Definition of "Support levels", 1.9.1 supported platforms and recruitment for platform maintainers
20006 [yugui@yu ui ] Great! thanks!
20021 [florian.ebel] What are the official rules for supported platforms? I didn't
20023 [florian.ebel] Found it myself, sorry for the noise.

Etc behavior on Windows
20007 [jdeville@mi ] Can I confirm that Etc methods always return nil on Windows?
20020 [charles.nutt] That's what the code appears to say :) And I modified JRuby's Etc to
20026 [Daniel.Berge] Etc.getlogin works on Windows, actually, but none of the other methods

[Bug #766] 'Not enough space' error on windows
20008 [redmine@ru y] Bug #766: 'Not enough space' error on windows

[Bug #768] Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7
20009 [redmine@ru y] Bug #768: Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7

[Bug #769] Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7
20010 [redmine@ru y] Bug #769: Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7

[Bug #769] Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7
20011 [redmine@ru y] Issue #769 has been updated by Jim Deville.

[Bug #768] Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7
20012 [redmine@ru y] Issue #768 has been updated by Jim Deville.

[Bug #770] assert_match should escape string argument
20013 [redmine@ru y] Bug #770: assert_match should escape string argument

[Bug #771] test/unit wrapper breaks compatibility
20014 [redmine@ru y] Bug #771: test/unit wrapper breaks compatibility

Fiber using a Proc object with a parameter having default value doen't work
20015 [ed.odanow@wo] Is this...

[Bug #742] 1.9.1-preview1 build fails on Mac OS X 10.4
20022 [redmine@ru y] Issue #742 has been updated by Florian Ebeling.

Fiber as Superclass - initialize of derived Class will not be executed
20025 [ed.odanow@wo] The following code...

[Bug #694](Closed) eof? call on a pty IO object causes application to exit
20028 [redmine@ru y] Issue #694 has been updated by Yukihiro Matsumoto.

[Bug #745] IO associates with the file descriptor, not the stream
20032 [redmine@ru y] Issue #745 has been updated by Roger Pack.

[Bug #743] Socket.gethostbyname returns odd values
20033 [redmine@ru y] Issue #743 has been updated by Roger Pack.

[Bug #766] 'Not enough space' error on windows
20035 [redmine@ru y] Issue #766 has been updated by Roger Pack.

build error mingw
20036 [rogerpack200] gcc -O2 -g -Wall -Wno-parentheses   -L.
20075 [matz@ru y- a] Does this still happen after r20341?

better error messages?
20038 [rogerpack200] "Many of the compiler diagnostics I got during the Ruby porting simply
20076 [B.Candler@po] I find this generally happens with a missing do or end, especially with
20733 [rogerpack200] <cough> this would be very nice for 1.8.x, too [if easy].

printing more output on unrescued exceptions
20039 [rogerpack200] I have been contemplating creating a patch which would make the output
20064 [wycats@gm il] ...
+ 20065 [jim.weirich@] YES!  Please.
| + 20067 [duerst@it ao] As far as I understand, this has been fixed. The stack trace is
| + 20915 [rogerpack200] Looks like this has been fixed in 1.9 and 1.8.7 at least. Thanks
+ 20116 [rogerpack200] I would even love it for 187 branch :)

[Bug #775] trunk build fails on Mac OS X 10.5.5
20040 [redmine@ru y] Bug #775: trunk build fails on Mac OS X 10.5.5

[Bug #775] trunk build fails on Mac OS X 10.5.5
20041 [redmine@ru y] Issue #775 has been updated by Roger Pack.

[Bug #743] Socket.gethostbyname returns odd values
20042 [redmine@ru y] Issue #743 has been updated by Alan Johnson.

[Bug #776] File#flush doesn't flush in windows
20043 [redmine@ru y] Bug #776: File#flush doesn't flush in windows

[Bug #766] 'Not enough space' error on windows
20044 [redmine@ru y] Issue #766 has been updated by Ittay Dror.

[Bug #775](Closed) trunk build fails on Mac OS X 10.5.5
20045 [redmine@ru y] Issue #775 has been updated by Yuki Sonoda.

ruby19 13% slower running rexml benchmark than ruby 1.8.6 p114
20046 [stephen.bann] I just added ruby 1.9 (svn rev 20317) to a simple xml processing
20050 [luc@ho k- on] REXML uses regular expressions for parsing and regular expressions are
20058 [charles.nutt] Why is that? JRuby uses a port of Oniguruma for all regexp support, and
20059 [lopx@ga et .] It's an additional huge String overhead itself (not just Regexp) and

1.9 method argument binding question
20047 [dblack@ru yp] => nil
+ 20052 [radek.bulat@] T24gU3VuLCBOb3YgMjMsIDIwMDggYXQgMTA6MDAgQU0sIERhdmlkIEEuIEJsYWNrIDxkYmxhY2tA
| 20053 [dblack@ru yp] You know, I should have just gone back and read Chapter 2 of the new
+ 20054 [ko1@at ot ne] class C
| 20055 [dblack@ru yp] Interesting. I wish I'd known that yesterday (I did a presentation on
+ 20060 [binary42@gm ] Arguments should never be reordered (ignoring named argument
  20061 [dblack@ru yp] irb(main):006:0> def m(a,b='b',c); p [a,b,c]; end
  20062 [binary42@gm ] I would say I was carefully differentiating arguments and parameters
  20063 [dblack@ru yp] I'm still seeing it (my last example) as a potentially very confusing

Unexpected Performance of Symbol Construction
20048 [kurt@ca hn t] Evaluating :°∆foo_bar°« is 1.5 to 2.5 times faster than :foo_bar. This is
+ 20051 [kurt@ca hn t] Looks like GC overhead/interaction was the culprit.  Note to self,
| 20057 [charles.nutt] A benchmark that doesn't consider GC is a bad benchmark. Don't disable
+ 20066 [ryand-ruby@z] This =20
  20074 [kbloom@gm il] Frankly I'm surprised that the simpler tests aren't intentionally equal
  20080 [kurt@ca hn t] ...

[Bug #766] 'Not enough space' error on windows
20049 [redmine@ru y] Issue #766 has been updated by Ittay Dror.

[Feature #578] add method to disassemble Proc objects
20056 [redmine@ru y] Issue #578 has been updated by Koichi Sasada.

[Bug #745](Rejected) IO associates with the file descriptor, not the stream
20068 [redmine@ru y] Issue #745 has been updated by Yukihiro Matsumoto.

[Bug #769] Mathn's Rational#power2 is broken on 1.8.6 and 1.8.7
20069 [redmine@ru y] Issue #769 has been updated by Yukihiro Matsumoto.

[Bug #731](Closed) Net::HTTPHeader []= method does not work as expected
20070 [redmine@ru y] Issue #731 has been updated by Yukihiro Matsumoto.

Is missing documentation considered a bug?
20071 [flo@an er gr] as some of you might have noticed, Wolfgang N=E1dasi-Donner and I had a =20=
+ 20081 [duerst@it ao] This is a purely personal opinion, but my suggestion is that
| + 20084 [flo@an er gr] #constants now returns symbols instead of strings, but no one documented
| + 20089 [ed.odanow@wo] Sorry, but this is a really frustrating answer. It is a fact that
+ 20103 [meta@po ox c] ...
  20105 [flo@an er gr] ...

Ruby 1.9.1 Testing Libraray
20073 [james@gr yp ] Is test/unit gone for good from Ruby 1.9?  Is it all going to be
20077 [jeremy@bi sw] Yes, minitest is the stdlib testing framework now, with the test-unit

Again: Questions about Fiber behaviour
20079 [ed.odanow@wo] I am a little bit surprized to get no (=0) answer for my questions in
20117 [rogerpack200] are you referring to ruby talk questions?
20124 [flo@an er gr] He's referring to ruby-core.
+ 20173 [rogerpack200] Hmm I didn't see those for some reason.  Not that I'm a fiber expert at all.
+ 20181 [matz@ru y- a] Both are bugs, and the former is fixed in the trunk.

[Bug #788] r20287 broke rb_str_replace
20082 [redmine@ru y] Bug #788: r20287 broke rb_str_replace

[Bug #788](Closed) r20287 broke rb_str_replace
20083 [redmine@ru y] Issue #788 has been updated by Nobuyoshi Nakada.

[Bug #790] Unexpected Behaviour: Fibers as superclass don
20086 [redmine@ru y] Bug #790: Unexpected Behaviour: Fibers as superclass don

[Bug #790] Unexpected Behaviour: Fibers as superclass don
20087 [redmine@ru y] Issue #790 has been updated by Florian Gilcher.

[Bug #791] Fiber using a Proc object with a parameter having default value doesn't work
20088 [redmine@ru y] Bug #791: Fiber using a Proc object with a parameter having default value doesn't work

[Bug #728] Segmentation fault with Ruby 1.8.7-p22 from st_lookup
20090 [redmine@ru y] Issue #728 has been updated by Martin Krauskopf.

[Bug #796] dynamic constant assignment
20091 [redmine@ru y] Bug #796: dynamic constant assignment
20095 [matz@ru y- a] You have Module#const_set, and to avoid duplicate assignment warning,
20096 [jdeville@mi ] What about a conditional dynamic assignment. It would be nice to have an id=
20099 [matz@ru y- a] Please understand constants are constants.  Adding such features might
20114 [transfire@gm] ent"
+ 20120 [michael.s.kl] I agree here, sounds like a safety measure that still can be worked =20
+ 20122 [matz@ru y- a] At least we can keep "illusion" of constants being constants for

[Bug #797] bug or feature: local method ?
20092 [redmine@ru y] Bug #797: bug or feature: local method ?
20097 [matz@ru y- a] Even though it's syntactically allowed, we have never made it
20098 [dave@pr gp o] I don't think it's useless: it's a great way of having a method
+ 20100 [matz@ru y- a] Ah, maybe.  But this syntax can be spared for more fantastic purpose,
| + 20102 [dave@pr gp o] I'm OK if you deprecate it in 1.9.3: I've used in just a few times,
| | 20104 [ezmobius@gm ] Rails routing code uses this method redefinition inside of itself
| | 20106 [jeremy@bi sw] Rails defines routing methods using instance_eval. It doesn't use
| + 20127 [francois.pr@] Well, it's seem that i am the last man standing to save the underdog. To
|   + 20128 [halostatue@g] Nested methods in Ruby don't work like they do in Pascal; they appear in
|   + 20132 [B.Candler@po] f = lambda { |a|
|   | 20142 [francois.pr@] technic 1 (bad)
|   | 20143 [B.Candler@po] Perhaps you've misunderstood how assignments work in Ruby. If you do
|   | 20155 [francois.pr@] I was aware that could used a class to do the trick, but a like to stay
|   | 20157 [B.Candler@po] Beware (a) premature optimisation, and (b) using your intuition to decide
|   + 20144 [diogoslisboa] This isn't possible though.
|     + 20145 [dev@me ne .c] Isn't the current behavior the same as using def self.foo?
|     | 20146 [dblack@ru yp] They're not the same. self.bar is a singleton method on whatever
|     | 20150 [dev@me ne .c] Ugh, my bad. I was under the strong impression using def in this
|     + 20152 [francois.pr@] ...
|       + 20153 [dblack@ru yp] That's very different, though, because you're redefining your inner
|       + 21061 [murphy@ru yc] ...
|         21072 [rogerpack200] #instance_method does
+ 20101 [meinrad.rech] ...

[Bug #796] dynamic constant assignment
20093 [redmine@ru y] Issue #796 has been updated by Tomas Matousek.
20094 [matz@ru y- a] Why?

[Bug #798] r19320 updated REXML but didn't increment REXML::VERSION
20113 [redmine@ru y] Bug #798: r19320 updated REXML but didn't increment REXML::VERSION

[Feature #578] add method to disassemble Proc objects
20115 [redmine@ru y] Issue #578 has been updated by Roger Pack.

link problem embedding ruby interpreter
20119 [pbrannan@at ] I've copied code from main.c in an attempt to embed the ruby interpreter
20123 [pbrannan@at ] Never mind, need to use -rdynamic -Wl,-export-dynamic (actually

Playing with String#bytes
20125 [emiel@re no ] Today I was playing around with manipulating strings containing
+ 20126 [radek.bulat@] V2hhdCBhYm91dDoKZGF0YS5mb3JjZV9lbmNvZGluZygiQVNDSUktOEJJVCIpWzEsM10uYnl0ZXMu
| 20130 [B.Candler@po] But that changes the encoding of 'data' as a side-effect. To prevent that,
| 20133 [michael.seli] In retrospect it might have been nice to have String#force_encoding! doing
| 20134 [matz@ru y- a] It appeared nice at first sight, but it turned out to cause troubles
+ 20140 [kbloom@gm il] In most cases, it probably isn't, so trying to change things around may

Ruby class variable access from C
20129 [cthompson@ne] class Myclass
+ 20135 [matz@ru y- a] * rb_eval_string()
+ 20138 [why@ru y- an] If you create the class in your C extension, you'll get a VALUE
  20139 [cthompson@ne] Ahh, but I'm not creating it in the C extension, for reasons that won't
  20141 [lukfugl@gm i] VALUE klass = rb_const_get("Myclass");

[Bug #799] cross compile 1.9.1-preview1 is not possible
20131 [redmine@ru y] Bug #799: cross compile 1.9.1-preview1 is not possible

Re: ASCII-8BIT String literals (Was: Re: Playing with String#bytes)
20137 [michael.seli] Thank you very much for explaining this.

[Feature #715](Rejected) Ruby interpreter should understand UTF-8 symbols with special meaning
20147 [redmine@ru y] Issue #715 has been updated by Yukihiro Matsumoto.

[Bug #743](Rejected) Socket.gethostbyname returns odd values
20148 [redmine@ru y] Issue #743 has been updated by Yukihiro Matsumoto.

Promising C coding techniques to reduce MRI's memory use
20149 [brent@mb ri ] After a couple weeks of long nights and false starts, I feel I may have come
+ 20168 [nobu@ru y- a] In shorter, if you use gcc, can't you try -mpreferred-stack-boundary=2 option?
| + 20169 [brent@mb ri ] Before hacking rb_eval(), I first tried finding some compiler
| + 20175 [B.Candler@po] The problem can be demonstrated with a very simple program (attached), and
|   20178 [brent@mb ri ] Brian,
|   20185 [B.Candler@po] Did you replace the whole switch statement with a dispatch table? That
|   + 20186 [pbrannan@at ] 1.9 does have callcc (require 'continuation').  It's probably not good
|   + 20193 [ezmobius@gm ] Actually I think you will find a *ton* of interest in this for the
|   | 20195 [sdsykes@gm i] I would like to second that.  1.8.7 patches would be very interesting indeed.
|   + 20194 [brent@mb ri ] Brian,
|     20196 [B.Candler@po] I know what you mean. My own small patches (just to fix compatibility for
|     20197 [duerst@it ao] Please don't assume that this was on purpose. With that much going
|     20207 [B.Candler@po] I posted it twice to ruby-core, once to rubyforge tracker and then migrated
|     20208 [matz@ru y- a] My bad, somehow I (we) missed all of your posts.  I am awfully sorry.
|     20209 [matz@ru y- a] * I missed the original report in the rubyforge tracker
|     20211 [B.Candler@po] Many thanks - I hadn't noticed that you had applied the patch to 1.9
|     20261 [rogerpack200] Few people use webrick maybe that's why there's no discussion :)
+ 20170 [rogerpack200] Wow thanks for doing that. I'd say please create a redmine bug for it
| 20179 [brent@mb ri ] Roger,
| + 20192 [ezmobius@gm ] Brent-
| + 20210 [matz@ru y- a] We are troubled by the "ghost references from the machine stack"
| + 20263 [rogerpack200] I was wondering why the older version :)
|   20268 [brent@mb ri ] Roger,
|   20272 [calcifer@ru ] I'll definitely try it out, too.
|   20314 [calcifer@ru ] Sorry, my bad, was jumping to conclusions too fast. Ignore that :)
+ 20379 [ks@ku ts ep ] The "initialization holes" that leave potential pointers on the stack occur
+ 20517 [rogerpack200] This is sweet.  I liked the idea so much I coded my own [perhaps much
  20534 [brent@mb ri ] Roger,
threads.html
top