3729-4093 subjects 4101-8816

mkmf enhancement for Win32
3944 [Daniel.Berge] It turns out that some Win32 functions, such as AttachConsole(), are
3946 [nobu.nokada@] It doesn't seem meaningful to check the compiling environment.
3952 [halostatue@g] Yes, it should check the target environment as a native win32

Address of current client in DRb front object
3945 [drbrain@se m] DRb stores the current client in Thread.current['DRb']['client'], but

Build failure with 1.8.2 on Win32
3947 [djberg96@ya ] Win2k

[ ruby-Bugs-1188 ] test stalling for test_readline (in TestReadline ) with cvs HEAD on mingw
3948 [noreply@ru y] Bugs item #1188, was opened at 11-12-2004 12:36

[ ruby-Bugs-1189 ] test stalling for test_open_pipe (in TestReadPartial) with cvs HEAD on mingw
3949 [noreply@ru y] Bugs item #1189, was opened at 11-12-2004 12:43

[1.9] don't pass its tests
3950 [decoux@mo lo] <#<Set: {"a", "b", #<: {"a", "b", #<: {"a", "b"}>}>}>> expected but was

[ ruby-Bugs-1190 ] error using IO.popen on  ruby19 (mingw)
3951 [noreply@ru y] Bugs item #1190, was opened at 11-12-2004 14:59

[BUG] tracer.rb tries to extract source code from binary files
3953 [florgro@gm i] Moin!

DamageControl for Ruby
3957 [aslak.helles] I have set up a continuous build for Ruby at the DamageControl (DC)

Re: [ruby-cvs] ruby/lib/rexml/encodings: Merged in the changes from BSD bug report. shift-jis is now shift_jis, in
3958 [kou@co mi ng] This change causes LoadError.

[ ruby-Bugs-1200 ] Segmentation Fault with Sockets and Threads
3959 [noreply@ru y] Bugs item #1200, was opened at 2004-12-13 10:18
3960 [matz@ru y- a] Could someone check these bugs (#1200, #1190, #1189, #1188) against
3961 [usa@ga ba ec] With 1.8.2, it causes IOError on mswin32/mingw32.

[BUG] webrick
3962 [cmills@fr es] Webricks logging may cause it to fail on a valid http requests (on Mac
3963 [gotoyuzo@no ] Sorry, I'm not familiar with DNSSD. But isn't the error of
3964 [rich@in oe h] DNS Service Discovery gives you the ability to resolve LAN-local addresses
3965 [cmills@fr es] That fixed it.  Thanks,

[BUG] unknown node type 0
3966 [andrew@wa ro] I still get this happening a lot with my Rubyx linux ruby script.

How does Ctrl-C on Win32 work?
3967 [djberg96@ya ] I'm just curious how trap("INT") works on Win32.  I

[ ruby-Bugs-1208 ] webrick test_client_server does not pass
3968 [noreply@ru y] Bugs item #1208, was opened at 15-12-2004 11:05
3969 [mneumann@nt ] This is due to method fork. I expect that when a block is given to fork,

3970 [andrew@wa ro] This is a long standing bug in Ruby, and has been reported hundreds of times
3971 [matz@ru y- a] Try recent snapshot.  The following change might fix your problem.
+ 3972 [andrew@wa ro] Thanks matz. That certainly looks a likely candidate.
+ 4001 [andrew@wa ro] Just an update; Almost a week now and no problems so far on the four servers I
  4005 [joaopedrosa@] That's a very good news. :-)

YAML::Syck::Parser#load dumps core
3973 [akr@m1 n. rg] Ruby dumps core as follows.

Patch to enable innetgr() & getnetgrent() suport in Etc
3974 [jfh@ci e. fl] ...

Patches to test/unit
3975 [ryand-ruby@z] I believe these are the minimal patches needed to make it possible to
+ 3979 [ryand-ruby@z] would it be better to start submitting the patches to the rubyforge
| 3983 [matz@ru y- a] No.  We use rubyforge to track issues, not patches.
+ 3980 [nobu.nokada@] $ ruby -rtest/unit -rtest/unit/autorunner -e Test::Unit::AutoRunner.run
  3988 [ryand-ruby@z] Given that the code in test/unit.rb is only used by autorunner.rb, I

/lib/ruby/1.8/rational.rb:374: [BUG] Segmentation fault
3976 [andrew@wa ro] I have been unable to reproduce this so far, but I guess the file/line-number
3977 [matz@ru y- a] I'm afraid it's not. :-<

[ ruby-Bugs-1232 ] The latest 1.8.2 cvs generates RI documentation really long time
3978 [noreply@ru y] Bugs item #1232, was opened at 2004-12-18 21:09

patch to cgi/session.rb
3981 [ksibilev@be ] The recent change to the CGI::Session
3986 [nobu.nokada@] Thank you, fixed now.

[BUG] Win32: rb_sys_fail() - errno == 0
3982 [florgro@gm i] Moin!
+ 3984 [djberg96@ya ] Which version of Windows are you running?  And did you
+ 3987 [usa@ga ba ec] Moin! Moin!
  3989 [florgro@gm i] Moin! :)
  3990 [usa@ga ba ec] Moin!
  4075 [Daniel.Berge] Although Florian's original code seems to work fine, I still see this
  4078 [usa@ga ba ec] How should I do to reproduce the bug?
  4080 [Daniel.Berge] Download win32-process-0.3.1
  4111 [usa@ga ba ec] Hmm, is the point here?
  4117 [Daniel.Berge] I'm not sure I follow, because this has worked as expected in certain
  4118 [pbrannan@at ] Perhaps you need an abstraction like this one (from
  4123 [Daniel.Berge] Thanks for this Paul - I'll take a look.
  4124 [pbrannan@at ] errno is a modifiable lvalue and is declared in errno.h.  It may be an

[ ruby-Bugs-1236 ] DL: structure with member of pointer type gives error
3991 [noreply@ru y] Bugs item #1236, was opened at 2004-12-20 13:49

Hash#delete - inconsistent with docs
3992 [cmills@fr es] irb(main):001:0> h = Hash.new("go fish")
3994 [dblack@wo bl] ruby -ve 'h=Hash.new("x"); h[1]=2; p h.delete("y"); p h.delete(1)'
3998 [dooby@d1 .k ] AFAICS, this was changed 3 years ago in rev1.58 ...

[bug] serious memory leak + continuations
3993 [mneumann@nt ] ruby 1.9.0 (2004-12-20) [i386-freebsd5.2.1]
3995 [mneumann@nt ] The memory leak probably results from the fact that the continuations

Updated ssh key
3996 [Dave@Pr gP o] Sorry to be an idiot, but I've just updated machines. Who do I send a
3997 [matz@ru y- a] For CVS account?  Send the key to cvs-admin@ruby-lang.org

[ ruby-Bugs-1243 ] ruby 1.8.1 installation problems on ia64/altix
3999 [noreply@ru y] Bugs item #1243, was opened at 2004-12-21 19:23

1.8.2 preview4
4000 [matz@ru y- a] Ruby 1.8.2 preview 4 is available from
+ 4002 [tom@in oe he] Yours,
| 4004 [halostatue@g] -austin
+ 4003 [holmberg@ia ] The following example (taken from the PickAxe2 documentation of
  4007 [nobu.nokada@] Sorry, it was a fault at this time.

cgi.rb GET and POST values patch
4006 [eustaquioran] ...

Question about the latest change to test/unit library
4008 [ksibilev@be ] It used to be possible to run unit test scripts by simply requiring
4012 [nobu.nokada@] I didn't think that usage, but testrb script, for such purpose.
4014 [ksibilev@be ] I didn't know about testrb script. Wouldn't it make sense to keep that

[PATCH] cgi.rb -- more GET/POST stuff
4009 [mde@st te 6.] First of all, I think it would be great, as Eustaquio suggests, to
4016 [sera@fh an .] GETs and POSTs are defined to be fairly different actions. I'd read
+ 4017 [eustaquioran] But what you think about keeping the values of GET and POST on different
| + 4018 [sera@fh an .] Personally, I don't have a really strong opinion about it ... Myself, I
| + 4019 [transami@ru ] Offer both ways. And it can still be backward compat.
|   4023 [eustaquioran] params, get_params and post_params.
|   4245 [patrick@he a] This is the architecture I am looking to for Narf.  You have a hash for
+ 4020 [mde@st te 6.] Some more thoughts / clarification on cgi.rb. This might be
+ 4024 [ruby@ja es r] <snip/>
  4026 [eustaquioran] Yeah, I think that *comparison* between languages sucks, because it
  4028 [mde@st te 6.] First of all, the entire discussion of when GET is appropriate
  + 4029 [eustaquioran] Yeah, that's what I did, also. :-)
  + 4031 [ruby@ja es r] <snip/>
    4032 [eustaquioran] Where should we ask for something like that so?
    4035 [ruby@ja es r] My 2-cents: If you have questions or comments on the implementation
    4054 [mde@st te 6.] Okay, the last little bit of follow up here ...
    4056 [sera@fh an .] Much of the discussion about web programming techniques can be moved to

YAML::Syck::Loader#transfer dumps core
4010 [akr@m1 n. rg] Ruby dumps core as follows.
4011 [nobu.nokada@] Moin,
4013 [matz@ru y- a] Commit the fix, please.

[ ruby-Bugs-1251 ] backtrace lost when throwing exceptions from procs defined in extensions
4015 [noreply@ru y] Bugs item #1251, was opened at 2004-12-23 14:55

[BUG] File#reopen bug in 1.8.2
4021 [dblack@wo bl] Sorry to be Scrooge... :-)
4022 [dblack@wo bl] Never mind -- I think this is covered in SCSC (Second Coming of Santa

[ ruby-Bugs-1254 ] Rdoc crash when running on /usr/local/lib/ruby with --ri-site
4025 [noreply@ru y] Bugs item #1254, was opened at 2004-12-25 14:23

Allowing custom number literal suffixes?
4027 [florgro@gm i] ...
+ 4030 [Peter.Vanbro] ...
+ 4057 [transami@ru ] All this talk about such trivial and obviously reasonable changes to cgi.rb
+ 4070 [nobu.nokada@] As for Rational, almost cannot be represented in that way.
  + 4072 [matju@sy pa ] Here are practical ways to achieve similar results using only standard
  | 4079 [florgro@gm i] While this is a nice idea I don't feel sure about polluting the global
  | 4081 [matju@sy pa ] Hahaha, so you prefer to pollute the Ruby grammar instead?
  | 4082 [florgro@gm i] I prefer not having global methods with one-character names as those
  | + 4084 [brent@mb ri ] I'm not sure I would advocate making Ruby's grammar even more
  | | + 4086 [matju@sy pa ] Funny that you mention this, because this actually changed significantly
  | | | 4107 [brent@mb ri ] Thanks for the warning. I'm still using 1.6.8.  Another reason to hold
  | | | + 4109 [florgro@gm i] But if it's already so hard to find support for the relatively simple
  | | | | 4112 [brent@mb ri ] By "String suffixes" I assume you mean numeric literal suffixes
  | | | | 4113 [florgro@gm i] Incorrect, I was referring to something like "foobar"Q which could work
  | | | | 4115 [brent@mb ri ] see those implemented;
  | | | | + 4122 [florgro@gm i] I do not insist on that. It was just not necessary for my proposal.
  | | | | + 4127 [Peter.Vanbro] ...
  | | | |   + 4129 [florgro@gm i] So maybe we should go back to the String.literal_methods style?
  | | | |   | 4131 [transami@ru ] Perhaps
  | | | |   | 4132 [brent@mb ri ] Could such specially named methods be invoked by anything other than
  | | | |   + 4130 [brent@mb ri ] This looks great.  Thank you.
  | | | + 4136 [matju@sy pa ] It's three less letters, and five less chars if you drop the parens by
  | | + 4090 [mneumann@nt ] Note that this would prevent separating the parser from the runtime
  | + 4085 [matju@sy pa ] methods of Kernel sometimes enter in conflict with other things, and it
  |   4089 [florgro@gm i] Using single-letter names make the chances for conflict even more
  |   4097 [matju@sy pa ] limiting? how so? can you elaborate on this?
  |   4100 [florgro@gm i] I think an "i" suffix is more natural for imaginary numbers than "I",
  |   + 4102 [Peter.Vanbro] Well, the way I've implemented it the method call follows the usual rules
  |   | 4104 [florgro@gm i] Ah, but this does cause the look-up to be context sensitive. I'd prefer
  |   + 4148 [matju@sy pa ] Mathematica uses uppercase I for imaginary, and is by far the most used
  |     4151 [florgro@gm i] So maybe it is hard to come up with a case where Syntax would be
  |     4161 [matju@sy pa ] Yes, that's right. This has to be replaced by either B("0.5")+x or
  |     4163 [florgro@gm i] Hm, did you mean extend() vs. class << obj; include(); end? I think (or
  |     4165 [matju@sy pa ] WOOPS, yes, I did mean that. Well, the only difference is that they use
  + 4077 [florgro@gm i] Moin.

Garbage collection trouble
4033 [chneukirchen] recently, I've been trying to use parts of libxml2 with Ruby/DL.  At
+ 4034 [mneumann@nt ] And I hope, this has something to do with my continuation problems I
+ 4036 [decoux@mo lo] try it with
  + 4037 [ttate@tt ky ] It appears that source is free'ed by GC outside the method. So I think
  + 4038 [chneukirchen] No more crashes and loop so far!  You are my hero. ;-)
    4043 [decoux@mo lo] Well, when you work with 'dl', this is like when you write an extension,
    + 4045 [ruby-lists@l] i'm right in saying that i haven't messed up
    | + 4046 [decoux@mo lo] Well, in this case if you call GC.disable it must work. But this is not a
    | | 4047 [ruby-lists@l] it still crashes even with GC.disable
    | | 4048 [decoux@mo lo] svg% tail -12 pull.rb
    | + 4051 [chneukirchen] With GC.disable, it works here.  (That's why I put GC.start into the
    + 4049 [chneukirchen] This makes a lot of sense in the after-thought, but I would have
      4050 [decoux@mo lo] It can't do it, for many reasons

[PATCH] To make "Thread.critical" sections "trap" safe
4039 [brent@mb ri ] I recently was involved in trying to get a multi-threaded ruby motion

Extensions, Internal
4040 [juergen.mang] I'm currently writing an extension. Suppose two functions, to get/set a
4041 [matz@ru y- a] I'm afraid I don't understand what you want.  What do you expect for a
4042 [juergen.mang] myObj.text.gsub!(/a/,"b") will not alter the string returned by #text,
4044 [decoux@mo lo] What you want is a delegator. No need to modify ruby.

YAML::Syck::Loader#transfer dumps core again
4052 [akr@m1 n. rg] Ruby dumps core as follows.

Thread rendezvous with timeouts
4053 [brent@mb ri ] We needed an event scheduler to manage our multi-threaded motion control

Latest CVS HEAD ChangeLog has conflict marker
4055 [ville.mattil] Here are the details,
4058 [ville.mattil] Same goes with 1.8.2 branch
4059 [Dave@Pr gP o] Very sorry - my fault

Hash#update docs
4060 [cmills@fr es] 1562,1563c1562,1566
4064 [eule@sp ce c] (In response to news:3E59898C-5A83-11D9-80A4-000A95A27A10@freeshell.org by

curious about 1.8.2 change - empty method returns nil
4061 [ryand-ruby@z] * parse.y (reduce_nodes): empty body should return nil.
4062 [matz@ru y- a] def foo(a=5)
4063 [ryand-ruby@z] Please forgive me. I may have missed some mail that went by in the
4066 [matz@ru y- a] def foo(&b)

Surprise in Time#sec
4065 [steven.jenki] $ cat t1.rb
+ 4083 [steven.jenki] ...
| 4087 [matz@ru y- a] I'm listening.  But I need time.
| + 4092 [steven.jenki] OK. Just wanted to make sure it hadn't gone off into the void.
| + 4095 [batkins57@gm] Is that a pun? :)
+ 4088 [nobu.nokada@] Do you mean the Time object returned by Time#getlocal should
  4094 [steven.jenki] I don't know. If you remove the call to Time#sec entirely, the examples

[BUG] Segfault in Thread#initialize / caller
4067 [florgro@gm i] Moin!
+ 4068 [chneukirchen] Program received signal EXC_BAD_ACCESS, Could not access memory.
+ 4069 [decoux@mo lo] It just need to test if the thread is alive.
+ 4071 [nobu.nokada@] I think re-initialization of Thread would be too dangerous.
  + 4073 [decoux@mo lo] Well, personnaly I think that methods which use a thread *must* test if this
  | 4074 [decoux@mo lo] it's rb_thread_kill()
  + 4098 [matz@ru y- a] Can you commit the fix?

4076 [jamis_buck@b] I recently used Ruby/DL to create bindings to the SQLite3 embedded
+ 4096 [pbrannan@at ] I don't think this is really feasible.  The problem is that there is no
| 4099 [decoux@mo lo] Why do you want to create new C functions on the fly. It use an hash, just
| 4119 [pbrannan@at ] The original problem Jamis pointed out is that the number of callbacks
| + 4120 [decoux@mo lo] Look at callback.func and see the difference between
| | + 4121 [decoux@mo lo] p.s. : I've forgotten to say, if I'm right it exist a memory leak in
| | + 4125 [pbrannan@at ] I don't see how to make this happen in C.  It is not possible (portably)
| |   + 4128 [batkins57@gm] I'm no C expert, so this may be completely inapplicable, but what
| |   | 4139 [pbrannan@at ] Variadic functions allow a function to take multiple number of
| |   + 4133 [decoux@mo lo] Each time that you define an interface to a higher language, like ruby,
| |     4140 [pbrannan@at ] I agree.  A callback system that does not pass back user-supplied
| |     4141 [decoux@mo lo] Well, there is a problem : read callback.func
| |     4143 [pbrannan@at ] I do not understand you.  Or perhaps you do not understand me, or maybe
| |     4144 [decoux@mo lo] Well, from dl.c
| |     4145 [ttate@tt ky ] I also can't see what you are getting at.
| |     4147 [decoux@mo lo] Well, dl must pass to rb_dl_callback_func_0() [proc, proto]
| |     4149 [ttate@tt ky ] Does the data struct returned by rb_dlsym_new() mean the struct sym_data
| |     4155 [decoux@mo lo] When you pass `func' to qsort(), you access fatally to the struct sym_data
| |     + 4157 [nobu.nokada@] qsort() passes only the pointers to elements to be compared.
| |     | 4158 [decoux@mo lo] and ?
| |     | 4171 [decoux@mo lo] Well, I've no idea actually about this, but I want to know : it's
| |     | 4172 [nobu.nokada@] Do you mean, in rb_dlsym_call(), the case that "to_ptr"
| |     | 4173 [decoux@mo lo] I've applied the patch
| |     | 4174 [nobu.nokada@] rb_dlsym_call() didn't free temporary argument lists.
| |     | 4176 [decoux@mo lo] You have found.
| |     | 4177 [nobu.nokada@] alloca() has implicit limit for the size on many platforms, but
| |     + 4160 [ttate@tt ky ] I know.
| + 4134 [matju@sy pa ] There's this sort of code in GridFlow (found in RAA) to detect the CPU
|   4137 [florgro@gm i] Isn't N^X all about that? It seems to be quite a hot thing for
|   4142 [matju@sy pa ] What's this N^X thing? is that the hot new way to say UNIX? Or is it the
|   + 4150 [florgro@gm i] Sorry, that should have been either NX (no execute) or W^X (write xor
|   | 4164 [matju@sy pa ] Thanks. Both of those abbreviations are also new to me.
|   + 4168 [ruby-lists@l] nx is a fairly well advertised and reported on feature
|     4170 [matju@sy pa ] Ok, maybe i just don't read reviews about processors...
+ 4103 [ttate@tt ky ] Jamis,
  + 4105 [tom@in oe he] So folks can follow along if they're so inclined...
  + 4106 [jamis_buck@b] Thanks for taking the time to respond. I look forward to Ruby/DL2!
  + 4108 [florgro@gm i] Very cool. Thanks a lot in advance; I know it will be wonderful. Just
    4135 [ttate@tt ky ] I'd like to add that support, however I don't have 64bit machine.
    4138 [florgro@gm i] I think I could get you in contact with somebody who would be willing to