3544-3837 subjects 3944-4245

^ [1.9] unpack("N")
3729 [decoux moulo] uln% /usr/local/bin/ruby -ve 'p "\000\000\000\001".unpack("N")'
3730 [gotoyuzo not] Could you try this patch?
3731 [decoux moulo] Work fine
3732 [matz ruby-la] Yuzo, can you commit the patch, please?

^ Core dump with latest Ruby from CVS
3733 [holmberg iar] I get a core dump (from "miniruby") when I try to build the latest
3734 [matz ruby-la] I couldn't reproduce the problem on Linux.  Does anybody have any
3735 [ryand-ruby z] on osx/darwin, process.c revision 1.119 causes a linker error in
3736 [akr m17n.org] Oops.  Please try following patch.
+ 3737 [kosako sofne] Won't you try this patch?
| + 3740 [holmberg iar] It got better with that patch (on Solaris).
| + 3741 [holmberg iar] *With* this patch I get the following crash on FreeBSD.
+ 3738 [ryand zenspi] Patch is good. Thanks.
| 3743 [akr m17n.org] I committed a bit refined version.
+ 3739 [ryand-ruby z] Patch is good. Thanks.

^ Latest CVS: each_line broken for frozen strings
3742 [matt lickey.] This just showed up in 1.9 CVS.
3745 [matz ruby-la] Fixed.  Thank you for the report.

^ patch for test/xmlrpc
3744 [usa garbagec] I found that I could not run test/xmlrpc/test_parser.rb on
3747 [mneumann nte] fixed. I removed "$LOAD_PATH.unshift '../../lib'", too.
3758 [nobu.nokada ] It becomes quiet by removing the constants before setting, but
3760 [mneumann nte] patch applied.

^ ERB suggestions
3746 [gsinclair so] Folks,
3748 [mneumann nte] Can you give an example, how your proposed non-cryptic code would look like?
3750 [gsinclair so] Instead of

^ Bugs/patches on the RubyForge Ruby project
3753 [tom infoethe] If you have commit on the Ruby CVS tree, please consider getting a

^ segmentation fault in Ruby
3754 [bujecas iol.] When I run a ruby program for a long time, it sometimes gives segmentation fault. I've tried the stable version 1.8.1 and all the development versions. When I compile the source code it gaves segmentation fault when running a program for a long time, but running with the ruby rpm package from fedora core 2, it runs perfectly for a long time without giving seg fault.  Is there some especial way to configure the compilation? Somebody knowns what I 'm talking about? :P

^ Regexp: compile error vs. exception
3755 [holmberg iar] Regexp.new("$*")
3757 [matz ruby-la] It's a bug.

^ Windows build problems (1.8.2preview3)
3756 [nathaniel ta] I'm having some very strange build problems on Windows, and thought
3759 [usa garbagec] This warning is meaningless. You can simply ignore it.
3761 [nathaniel ta] Good to know.
3763 [usa garbagec] OK, I (maybe) understand the situation.

^ x.include? speed observation
3762 [hgs dmu.ac.u] I wondered about implementing a fairly small set of  present/absent flags and

^ [BUG] Readline.completion_append_character in latest snapshot
3765 [vjoel PATH.B] In ruby 1.9.0 (2004-11-18) [i686-linux], something changed in readline,
3767 [shugo ruby-l] I added a test method to test/readline/test_readline.rb and fixed it:)

^ [ ruby-Patches-1074 ] Make READ_DATA_PENDING use __fpending() if available
3766 [noreply ruby] Patches item #1074, was opened at 2004-11-18 16:20

^ rdoc -1 tries to eat my files
3768 [ryand-ruby z] % rdoc -1
3776 [dave pragpro] Do you have any Ruby files in a subdirectory called doc? Alternatively,
3778 [ryand zenspi] No subdirs at all. The problem is reproducible with the latest released
3781 [dave pragpro] Eek - indeed it is.

^ Forwardable warnings
3769 [djberg96 yah] require "forwardable"

^ OT:  Book about compiler (parser) design
3770 [mailinglists] recusive parsers, but ruby needs a more complicated LR parser. Can
3772 [ruby jamesbr] Compiler Design in C, by Holub?
3775 [dross code-e] Compiler Design in C is a good book. I highly recommend it. The dragon

^ Bug in IO#gets("") in 1.9 ?
3771 [holmberg iar] Is the following change in IO#gets intentional, or is it a bug?
3779 [aamine lover] Following patch may help you.
3780 [matz ruby-la] Commit the change, please.

^ mkmf.rb
3773 [cmills frees] Would it make sense to have $LOCAL_LIBS as a dependency for the DLLIB

^ [PATCH] Nicer traceback at Interrupt
3774 [holmberg iar] When Ruby is interrupted by for example Ctrl-C, the traceback looks
3777 [matz ruby-la] No.  Thank you for the patch.

^ [ ruby-Bugs-1083 ] RDoc and requires
3782 [noreply ruby] Bugs item #1083, was opened at 2004-11-22 14:51

^ Rubyforge Ruby project enhancment
3783 [Daniel.Berge] I think it would be a good idea to create some categories and groups for
3794 [tom infoethe] This sounds good to me... unless anyone objects, I plan to add in these
3797 [tom infoethe] I'll leave categorizing the current bugs for someone who actually knows

^ [ ruby-Bugs-1084 ] The latest 1.8.2 cvs prints parse error when starting extension compiling
3784 [noreply ruby] Bugs item #1084, was opened at 2004-11-22 23:21

^ The latest 1.8.2 cvs prints parse error when starting extension compiling
3785 [matz ruby-la] Would anyone confirm the bug #1084 on the RubyForge bug tracker?
3787 [holmberg iar] ./ext/extmk.rb:27:in `require': ./lib/mkmf.rb:1043: can't find string "SRC" anywhere before EOF (SyntaxError)
+ 3788 [holmberg iar] I also got rid of the problem by the following code change.
+ 3789 [matz ruby-la] Can you do the additional check?
  3790 [holmberg iar] This message is in MIME format.  The first part should be readable text,
  3791 [decoux moulo] Apparently this is
  3792 [matz ruby-la] I see.  This was introduced to solve [ruby-code:3766] submitted by
  + 3793 [decoux moulo] Well, my big problem is that, in Solaris 8, __fpending() is described as
  | 3796 [matz ruby-la] Oh well, so I should cancel the patch.  If someone has any additional
  + 3795 [decoux moulo] Well, I don't understand the comment in [ruby-core:3766]

^ Please commit erb.rb
3786 [gsinclair so] Folks,
3802 [gsinclair so] Please?
3813 [matz ruby-la] Done.  Will be committed soon.

^ Possible Bug: Running thread blocks access to /proc/loadavg
3798 [creinig mayn] charset="iso-8859-15"
3803 [nobu.nokada ] Seems Linux kernel 2.6's issue, its loadavg doesn't implement
3804 [creinig mayn] charset="iso-8859-1"

^ autoconf >= 2.53 ?
3800 [holmberg iar] AC_PREREQ(2.50)
3801 [matz ruby-la] Which one?  CVS HEAD only? Or ruby_1_8 branch as well?
3805 [holmberg iar] I only had problems with HEAD.
3808 [matz ruby-la] Thank you for the information.

^ [ ruby-Bugs-1095 ] current ruby 1.8.2 cvs has inconsistent string handling (null char in string error)
3806 [noreply ruby] Bugs item #1095, was opened at 2004-11-24 16:48
3807 [nobu.nokada ] Buffer string is reused.
3811 [matz ruby-la] Can you commit the fix please?

^ [ ruby-Bugs-1107 ] Ruby does not compile on Red Hat Entreprise Linux 3.
3809 [noreply ruby] Bugs item #1107, was opened at 2004-11-24 15:25
3810 [nobu.nokada ] Could you show the result about rlim_t in config.log file?
3812 [holmberg iar] I think this is the problem with a too old autoconf (see the recent
3814 [holmberg iar] Maybe I should give some more details.
3816 [akr m17n.org] I found a problem in AC_CHECK_SIZEOF invocation.
3818 [matz ruby-la] Can you commit?

^ Re: ruby -y
3820 [florgro gmai] Hm, ruby -y doesn't seem to be documented in --help. Is this already
3822 [matz ruby-la] It's been available for long time, but it's not useful for most
3823 [florgro gmai] Hm, I am not yet sure how to use (couldn't google up documentation) it

3824 [dross code-e] Greetings everyone civil,

^ File::Stat#blksize for Win32
3831 [djberg96 yah] Can we add support for File::Stat#blksize for Win32?

^ CGI::Session invokes nonexistent method
3832 [ksibilev bel] In the current HEAD and in the ruby_1.8 branch, CGI::Session tries to
3834 [matz ruby-la] Right.  Thank you for the patch.

^ [ ruby-Bugs-1120 ] Segmentation fault with StringIO#puts and StringIO#rewind
3836 [noreply ruby] Bugs item #1120, was opened at 2004-11-27 08:37
3847 [nobu.nokada ] Thank you for reporting.  Fixed now.

^ RFE: Kernel#set_trace_func
3838 [florgro gmai] Moin!
+ 3839 [transami run] Would a stack be even better?
| 3840 [florgro gmai] Yes, but I'm trying to keep the changes as small as possible for
+ 3854 [pbrannan atd] This would introduce an incompatibility.  set_trace_func currently

^ Creating a NoOp Node for 'end' statements
3841 [mailinglists] I think we should generate No-Operation statements at the position of

^ [BUG?] convert_type() uses rb_respond_to()
3842 [florgro gmai] Moin!
3843 [matz ruby-la] I think so.  My opinion is that "respond_to?" of proxy object should
3844 [florgro gmai] This sounds good -- so convert_type() should basically use
3845 [matz ruby-la] Wait, I have forgotten one factor: performance.
3849 [florgro gmai] Hm, so will this be changed so that Ruby uses duck-typing in this case?
3855 [matz ruby-la] Yes.  Already changed on CVS HEAD.

^ REXML::Functions.name == nil
3846 [ryand-ruby z] I found this oddity while running ParseTree over every .rb file in the
3851 [florgro gmai] I'm not sure about .id as that should not be used anymore anyway, (use
3857 [ryand zenspi] Until .id is removed from the vocabulary (and probably a long while
+ 3858 [halostatue g] I personally always use #__id__ when I need the object's ID, not #id
+ 3859 [florgro gmai] My point was that the workaround works generally. Use it instead of
  3864 [ser germane-] charset="iso-8859-1"

^ Bug with ri Object#singleton_method_added
3848 [halostatue g] singleton_method_undefined(symbol)
+ 3850 [florgro gmai] I think this is caused by both (and quite a lot more methods) using the
+ 3852 [dave pragpro] I don't see this problem here. Are you using a recent RDoc (the one
  3853 [halostatue g] I'm probably using an older version of Ruby (1.8.2 preview 2 or

^ [1.9] define_method / super
3856 [decoux moulo] uln% cat b.rb

^ Wrong test for xmlrpc
3860 [nobu.nokada ] test/xmlrpc/test_parser.rb has a wrong assumption about the
3865 [mneumann nte] I fixed that issue in CVS, by using YAML instead of #inspect.

^ [1.9, 1.8] super
3861 [decoux moulo] 2 points about super and 1.9
3872 [matz ruby-la] Yes.  If you modify given arguments in 1.9, it will affect the super's
+ 3874 [decoux moulo] I'm not really sure because it's easier to write for 1.8 :-(
| 3876 [matz ruby-la] Wow, can I merge this in 1.9?
| 3877 [decoux moulo] Well, actually I don't see a problem with this patch but, because I'm
| 3878 [matz ruby-la] Of course I will fix (if you don't have time), when we meet a
+ 3879 [shugo ruby-l] It seems that this fix broke something.
  3880 [decoux moulo] This is why, in the patch to 1.9, it use `tmp' rather than `args'
  3882 [shugo ruby-l] Thank you.

^ [PATCH] Marshal.dump'ing OpenStruct objects
3862 [batsman.geo ] I think this allows one to dump OpenStruct objects, and addresses the
3863 [matz ruby-la] I have recently restored the old behavior, but this one seems
3868 [nobu.nokada ] This change makes no sense as for the compatibility.
+ 3871 [matz ruby-la] Right.  Probably I just should add marshal_dump and marshal_load, as
+ 3873 [batsman.geo ] Sorry.
  3875 [transami run] Sorry? No, say "Awe Shucks!"

^ [ ruby-Bugs-1147 ] x86_64 segfault on stack overflow
3866 [noreply ruby] Bugs item #1147, was opened at 2004-12-02 21:34
3867 [cmills frees] I think this happens on ia64 too.

^ should fatal class be "Fatal"?
3869 [mneumann nte] a lower-case class name "fatal"... is this expected?
3870 [nobu.nokada ] It is not for use in ruby level.
3883 [florgro gmai] fatal = nil
3891 [matz ruby-la] I'd rather expect you to ignore something you have found when you

^ mkdir, mkdir_p in FileUtils and mode
3881 [flori nixe.p] if I get this correct, the methods in FileUtils are supposed to mirror
3887 [aamine lover] Yes.
3888 [flori nixe.p] I tried this on Linux, Dragonfly BSD and Mac OSX. The s-bit was ignored
3889 [aamine lover] OK, now I understand the problem.
+ 3890 [flori nixe.p] Great. Thanks a lot!
+ 3892 [s-urabe par.] Minero, your changes borke Ruby's bootsrtap sequence.  When you make a Ruby from the source now, you get

^ Browsing the CVS tree?
3884 [sera fhwang.] Is there any place where I can visually browse the Ruby CVS tree, the
3885 [dooby d10.ka] ____________________________________________________________________________
3886 [sera fhwang.] Wonderful. Thanks!

^ [ ruby-Bugs-1156 ] Infinite loop in gsub
3893 [noreply ruby] Bugs item #1156, was opened at 2004-12-04 20:53
+ 3894 [decoux moulo] not a bug, for me.
+ 3895 [dblack wobbl] I believe it isn't actually infinite, just very, very long.  It's also

^ bug in debug's help
3896 [ryand-ruby z] <506> cvs diff -du lib/debug.rb
+ 3897 [gsinclair so] Your patch implies that
+ 3898 [matz ruby-la] Hmm, It should be either
  3906 [ryand zenspi] <506> cvs diff -du lib/debug.rb

^ Rexml bug report
3899 [matz ruby-la] Here's a bug report from the NetBSD camp.
+ 3902 [ser germane-] Ok.
+ 3904 [ser germane-] charset="iso-8859-1"

^ [ ruby-Bugs-1164 ] The latest 1.8.2 cvs has uncompilable is_socket on unix
3900 [noreply ruby] Bugs item #1164, was opened at 2004-12-06 14:10
3901 [usa garbagec] not 1.8.2 cvs, but HEAD?

^ [ ruby-Bugs-1165 ] The latest cvs HEAD does not link on solaris
3903 [noreply ruby] Bugs item #1165, was opened at 2004-12-06 15:40
3905 [akr m17n.org] Thank you for thr report.  It is commited.

^ [patch] Obtaining mode information on an IO object
3907 [jos catnook.] The attached patch implements IO#mode. This method returns the mode the IO
3909 [nobu.nokada ] "mode" feels confusing as "stat.mode", how about "accmode" instead?
3910 [jos catnook.] I rather like the symmetry between File.mode and File::Stat.mode, but accmode
+ 3911 [nobu.nokada ] permission of a existing file, while yours is the mode which
| 3913 [jos catnook.] You are indeed correct. Confusing use of the word `mode'. `accmode' is the
+ 3925 [ruby jamesbr] What is "accmode" suppose to be indicate? "Access mode?"  If that's the
  + 3926 [jos catnook.] Not at all. `access_mode' is even better imo. To be honest, I didn't think
  + 3927 [nobu.nokada ] Yes, it came from O_ACCMODE.
    3931 [jos catnook.] Hm, then again that in turn is a good argument for sticking with `accmode', as
    3932 [ruby jamesbr] Well, that's the question.  Who is the target audience, such that this
    3933 [jos catnook.] Well, we are talking about a POSIX/OpenGroup interface here which as such is
    3934 [nobu.nokada ] Or, how about separating them so that accmode will return
    + 3935 [halostatue g] I like this.
    | 3937 [robert.mcgov] +1
    + 3938 [jos catnook.] Excellent idea. +1
    + 3955 [flori nixe.p] Will the modes change if IO#binmode is called on an existing IO object?
      3956 [jos catnook.] binmode is useful only in MS-DOS/Windows environments and has no effect on

^ tcltklib load error from tk.rb
3908 [wybo servaly] ruby 1.9.0 (2004-12-06) [i686-linux]
3912 [nagai ai.kyu] Possibly, yes. :-)

^ Pathname needs a makeover
3914 [Daniel.Berge] I've taken a look at the Pathname class.  In my humble opinion, it needs
+ 3915 [gsinclair so] I don't see a good reason for that.  You mention a practical reason
| 3916 [djberg96 yah] Hm...to me a pathname is just a string that happens to
| 3917 [gsinclair so] I'm not suggesting that it _should_ be based on Array, but that
| 3918 [djberg96 yah] I think maybe we see a pathname as different things.
+ 3919 [matz ruby-la] It's not.  Pathnames can be given to open() etc. in place of strings,
+ 3920 [akr m17n.org] First of all, Python people discuss various points for a class for pathname.
  3921 [Daniel.Berge] pathname.
  3954 [akr m17n.org] Such redefinition violates the string behavior.

^ [ ruby-Bugs-1173 ] Incorrect escaping in strings produced by String::inspect
3922 [noreply ruby] Bugs item #1173, was opened at 2004-12-08 17:35
+ 3923 [dblack wobbl] I don't think it's inspect or eval specific.
| 3924 [florgro gmai] I think the original poster wanted '#@a'.inspect to be "\#@a" instead of
| 3928 [dblack wobbl] This message is in MIME format.  The first part should be readable text,
| 3936 [florgro gmai] It does not raise a SyntaxError, but it does not produce '#@a' -- it
+ 3929 [nobu.nokada ] Does this patch help?
  3930 [matz ruby-la] Commit the patch, please.

^ [ ruby-Bugs-1184 ] "Invalid Argument" for Windows Sockets
3939 [noreply ruby] Bugs item #1184, was opened at 2004-12-09 13:34
3940 [usa garbagec] You sent illegal HTTP message to Google, so Google disconnect

^ [ ruby-Bugs-1186 ] String#sub置
3941 [noreply ruby] Bugs item #1186, was opened at 2004-12-10 23:11

^ Re: [ ruby-Bugs-1186 ] String#sub
3942 [decoux moulo] not really a bug, it's how ruby interpret \ in #sub

^ Error handling in file related methods?
3943 [holmberg iar] I wonder what the "philosophy" is regarding error handling in file
+ 4091 [holmberg iar] A while ago I asked the questions below, and got no answer at all.
+ 4093 [halostatue g] I don't know what the preferred action would be on this, Johan -- I