2467-3290 subjects 2894-3184

lib/ruby/1.9/yaml.rb:193: [BUG] Segmentation fault
2697 [batsman.geo@] Happens in both 1.9.0 (2004-03-25) and ruby 1.8.1 (2003-12-25).
+ 2698 [matz@ru y- a] Too few information to debug.  Do you have any reproducing script?
+ 2705 [ruby-core@wh] batsman, what is happening with this bug?  If you're still getting this,
  2706 [batsman.geo@] I'm working on it -- the bug happens during some rather long unit tests
  + 2708 [xsdg@fr en d] ,=20
  + 2710 [matz@ru y- a] You don't have to isolate the problems as long as they are
    2729 [batsman.geo@] Please find the full sources at
    2730 [ruby-core@wh] Good, good.   Thank you.

Bug in pp library
2699 [dave@pr gp o] ruby -r pp -e 'PP.singleline_pp({ 1 => 1})'
2701 [akr@m1 n. rg] Fixed.  Thank you for the reports.
2702 [dave@pr gp o] No - thank you for such a rapid fix :)

Bug in PrettyPrint
2700 [dave@pr gp o] 'pp=PrettyPrint.new; pp.seplist([1,2,3]){}'

Proposed patch to add SSL support to net/pop.rb
2703 [daniel@ni ht] ...
2704 [daniel@ni ht] ...
2707 [daniel@ni ht] ...
2721 [aamine@lo er] I think that #start has taken too many parameters.
+ 2722 [daniel@ni ht] I agree that there are a lot of arguments to #start, but I think it is the
| 2723 [austin@ha os] ... }
| 2728 [daniel@ni ht] That might work.
| 2733 [austin@ha os] Perhaps a third method could be provided that could serve this dual-duty.
| 2736 [nobu.nokada@] end
| 2737 [daniel@ni ht] I like this solution best, but that's still a few more lines than the way it
| + 2738 [d.borodaenko] Because it's ugly?
| + 2739 [aamine@lo er] Because it is not a good design that a method have 7 parameters.
|   2744 [daniel@ni ht] I'll reimplement based on nobu's suggestion.
+ 2740 [eivind@Fr eB] would give something like the following (for 1.8 - in 2.0, we have extra
  2741 [aamine@lo er] 1.  Matz has already discouraged use of ":name => value"
  + 2742 [eivind@Fr eB] (NOTE: The below is just an opinion piece, and I trust you to make a
  | 2743 [austin@ha os] This has also passed me by. Does anyone have a reference for this
  + 2746 [hermit-jah@y] Does that mean that ":name => value" style arguments will deprecate? And
    2747 [aamine@lo er] m(**{a=>b, c=>d})

typos in lib/singleton.rb
2709 [ian@ca ib n.] Attached is a patch to fix some spelling errors in the CVS of
2711 [matz@ru y- a] Is it possible for you to turn off diff coloring?
2712 [ian@ca ib n.] Yes. Sorry about that.
2715 [matz@ru y- a] Thank you.

more spelling and grammar fixes
2713 [ian@ca ib n.] ...
2714 [kapheine@hy ] In a lot of the places where you changed "Fail" to "failed" you missed
2716 [matz@ru y- a] And two modifies for socket.c will not be accepted.
2717 [ian@ca ib n.] Oops. Sorry for the mistakes.
2718 [matz@ru y- a] Thank you.  Patches will be applied.
2734 [ian@ca ib n.] Do you think it is worth the time to produce a similar patch for 1.8.2?
2735 [matz@ru y- a] I already applied your patch to my 1.8 local copy as well.

one more round of exception message clean-up
2719 [ian@ca ib n.] Here is a patch against the latest CVS that cleans up the exception

Re: Obtaining size of Ruby objects in C
2720 [elathan@ph s] I can't understand why SWIG fails.

Issue with rb_time_new()
2724 [djberge@qw s] Ruby 1.8.1
+ 2725 [matz@ru y- a] rb_time_new() takes two arguments: rb_time_new(time_t sec, time_t usec)
+ 2726 [nobu.nokada@] Originally, rb_time_new() requires two time_t arguments, but is
  2727 [matz@ru y- a] I didn't notice about the <missing.h>.  Please commit.

Re [ruby-dev:23297] new function for matching path name (and case sensitivity depends on system)
2731 [ocean@m2 cc ] I proposed new function File.match? to match path name

vflow 0.1a
2732 [jm@tr ns ct ] For my first attempt at a ruby module I've taken on the development of

Reworked SSL for POP patch
2745 [daniel@ni ht] ...
2773 [daniel@ni ht] Anyone have a chance to look at this yet?
2780 [aamine@lo er] Sorry, I missed your mail.

Proposal: New Bignum
2748 [evan@fa li g] During some experiments with ruby cryptography, I found some problems with
+ 2749 [matz@ru y- a] I'm not sure whether replacing current bignum is a good idea.  The
| + 2750 [evan@fa li g] See comments below...
| | 2751 [matz@ru y- a] Interesting.  Do you think MBignum is _that_ fast (and efficient)?
| + 2753 [eivind@Fr eB] Why would this be a problem for cryptography?  (I'm probably being
|   2756 [evan@fa li g] I believe that Matz concern is centered around where autoconversion between
|   2757 [dave@pr gp o] An alternative, which is more compatible with the existing scheme,
|   2758 [evan@fa li g] Hm. Seems like this might be a bit too magical, but it would work. Perhaps
+ 2752 [feldt@ce ch ] I agree with this proposal. Python has fast bignum exponentiation in the
  + 2754 [matz@ru y- a] I am open.  Either is ok that
  + 2755 [evan@fa li g] See below.

Re: Make clean runs autoconf
2759 [ryand-ruby@z] There is a Makefile rule that through transitive closure is dependent
2760 [nobu.nokada@] It doesn't work for other makes.
2761 [ryand-ruby@z] Can we rename the "Makefile" rule to something like "rebuild-Makefile"?
2785 [nobu.nokada@] It will run everytime.

RDoc markup problem with \_abc_<i>def</i>
2762 [akr@m1 n. rg] I found that RDoc formats \_abc_<i>def</i> as _abc_d<em>ef</em>.
2763 [dave@pr gp o] Many thanks for the report. It should be fixed now.

RDoc :enddoc:
2764 [akr@m1 n. rg] I found that RDoc document some method after :enddoc:.  Is it
2765 [dave@pr gp o] No, it was a bug... :)
2768 [akr@m1 n. rg] Thank you.
2769 [dave@pr gp o] Sigh - a special case I didn't change. Should be fixed now.
2770 [akr@m1 n. rg] It is not fixed on my environment.  M::C is still documented as follows.
2772 [dave@pr gp o] The problem here is the RDoc currently interpreters :enddoc: as an
2775 [akr@m1 n. rg] I understand that RDoc maintains a documentation-enabling flag for

RDoc generates a dangling hyperlink with :stopdoc without :startdoc:
2766 [akr@m1 n. rg] I found that RDoc generate a dangling hyperlink.
2767 [dave@pr gp o] Hmm... Here I'm not sure what to do. I suspect that I really should
2771 [gsinclair@so] No, I think that's a good idea.

Typo in gc.c 'descarding' should be 'discarding' [PATCH]
2774 [drbrain@se m] On line 1722, 'descarding' should be 'discarding'.

CSV changes
2776 [dave@pr gp o] * remove Cell#to_str [ruby-dev:23332]
2777 [dave@pr gp o] "While I was documenting it recently, I have to admit I didn't
2778 [nahi@ke na t] These requirements are from my old still running application.

Library bundling (was Re: CSV changes)
2779 [gsinclair@so] I think there's some merit in that.  There are pros and cons, and it

CVS Question
2781 [BCoish@Dy ax] Just downloaded a copy of CVS from cvshome.org (windows binaries)

Problems with gsub, double quoted strings
2782 [boson@cy er ] Opening a fresh irb session (using Ruby 1.8.1 on OSX) I have the following
2784 [george.marro] In the examples above, $1 and $2 are evaluated before the call to gsub even

[PATCH] typos in *.c RDoc comments
2783 [djkea2@mu ca] This fixes some typos in the RDoc comments for *.c It's not very

Possible bug in init_copy or rb_gc_copy_finalizer?
2786 [ryand-ruby@z] #!/usr/bin/ruby -w
2791 [ryand-ruby@z] any word?
2792 [matz@ru y- a] "Will be merged".

[PATCH] typos in lib/* RDoc comments
2787 [djkea2@mu ca] This fixes some more typos in the RDoc comments for lib/*

Problems building ext/io/wait.c in 1.8 branch
2788 [gsinclair@so] I can't get io/wait installed.  The main problem is that it doesn't
2799 [gsinclair@so] Any comments on this?
2800 [decoux@mo lo] cd /home/gavin/Projects/ruby18/ext/io/wait
2801 [gsinclair@so] Hmmm... configure problems.  See below.
2802 [decoux@mo lo] The Makefiule is created only if it find FIONREAD. See the content of
2803 [gsinclair@so] Here's the only mention, in /usr/include/sys/termios.h, and it doesn't

STARTTLS support for net/smtp
2789 [daniel@ni ht] ...
2798 [aamine@lo er] Thank you very much.

line tracing events
2790 [Markus.Barch] I found that line trace events are handled quite differently between
2796 [mailinglists] I can't follow you here. I don't have any problems with a non stopping
2797 [Markus.Barch] Thanks! My debugger (RDT, http://rubyeclipse.sf.net) seems to work with

BUG: segfault (with code)
2793 [rpav@me hl .] rpav@central:lib% irb
2794 [nobu.nokada@] I guess aref_args should be a list node.
2795 [nobu.nokada@] And CVS HEAD also seems wrong, shouldn't be like the latter?
2804 [matz@ru y- a] I guess you're right.  Commit the fix please.

Bug 1318
2805 [steven.jenki] Any comments on
+ 2806 [decoux@mo lo] Not really sure, but seems a problem with REXML. Add a character and it
| + 2807 [ocean@m2 cc ] (2004/04/23 21:04)
| | 2808 [ser@ge ma e-] charset="iso-8859-1"
| + 2809 [steven.jenki] Yes. I can make the problem go away in several ways, none of which
|   2810 [decoux@mo lo] Well, if I've well understood the problem is with the buffer (@buffer)
+ 2811 [ser@ge ma e-] charset="iso-8859-1"
  2812 [steven.jenki] It passes on the test file I provided. My real test cases are still
  2813 [ser@ge ma e-] charset="iso-8859-1"

Tempfile strangeness in 1.9.0
2814 [steven.jenki] I didn't open a bug for this because it's from the CVS head, but it
2815 [matz@ru y- a] Tempfile is a duck quacks like an IO, so that adding a method to the
2816 [steven.jenki] Sorry, my message wasn't clear. I read in the Pickaxe about Tempfile and
2817 [matz@ru y- a] Use Tempfile.open instead, as you might notice already.
2818 [steven.jenki] Yes. I don't really need to use the block form, I was just debugging.

File.join oddity ?
2819 [holmberg@ia ] File.join("a", [["b"], "c"], "d")     # "a/b/c/d"
+ 2820 [djberge@qw s] Extra complexity?  Unintuitive?  This behavior doesn't really surprise
| 2822 [holmberg@ia ] ?>
| 2823 [decoux@mo lo] Well, try this
+ 2821 [decoux@mo lo] Well, I don't know the real reason, but File::join take an Array as

[yaml] exponentials and double quotes in Range (Re: [syck] YAML.dump("a".."z"))
2824 [nobu.nokada@] This seems fixed now, but other issues.

[PATCH] ruby-mode.el indentation fix
2825 [g_ogata@op u] if not_contrived?(x)
2826 [matz@ru y- a] Thank you.

[jim@freeze.org: FileUtils Bug in 1.8.0]
2827 [jim@fr ez .o] Does the current installation of Ruby have this problem?
2828 [jim@fr ez .o] Oops, this email got out early.

[PATCH] lib/test/unit/testcase.rb
2829 [batsman.geo@] Problem

Bug in FileUtils - apathy or bad email
2830 [jim@fr ez .o] I've posted on this three times and have yet to see a response.
+ 2831 [steven.jenki] $ ruby -v test.rb
| 2832 [WBrown@is ft] Jim,
| + 2835 [jim@fr ez .o] "fred" and "neighbor" are siblings. The link inside "fred" points
| + 2836 [ser@ge ma e-] charset="iso-8859-1"
+ 2837 [decoux@mo lo] Well, the problem is perhaps here
+ 2838 [aamine@lo er] I'm sorry, I was missed your mails...

tcktklib.c patch
2833 [ksibilev@be ] Since Exception#to_str method has been removed in favor of #message

rb_struct_new example needed
2834 [jm@tr ns ct ] If I may endulge  in the use of someone's time for a bit. I've narrowed
2839 [matz@ru y- a] No.  You have to define members first by rb_struct_new() to use
2841 [jm@tr ns ct ] Bashes head against keyboard. Doumo arigatou. That was it. It now runs

Changing Resolv::DNS
2840 [daniel@ni ht] I put out a RCR a while ago (176) that subclassed the Resolv::DNS class to
2842 [akr@m1 n. rg] I feel this initialization style is good enough.
2847 [daniel@ni ht] I like this.  Creates a more unified Resolv class.
2884 [daniel@ni ht] Any objections to getting this committed to the tree?
2885 [akr@m1 n. rg] It is already committed.

Another bug in fileutils.rb
2843 [holmberg@ia ] Many methods in FileUtils take either an array of arguments or a
2858 [aamine@lo er] Thank you very much.

[PATCH] ruby-mode.el indentation fix (another one)
2844 [g_ogata@op u] "#{x}" if x =~ /x/

RDoc: m(&block) {|x| ... }
2845 [akr@m1 n. rg] I found that RDoc generates a method headline which has both a block
2846 [dave@pr gp o] Good point: I've change this in the latest CVS.

Tempfile strangeness
2848 [flori@ni e. ] this throws an exception Errno::EACCESS under Windows
+ 2849 [flori@ni e. ] Thinking about it a little (ca. 1 minute) longer, i came to the
+ 2850 [matz@ru y- a] How about ignoring EACCESS?  The tempfile will be removed anyway,
  + 2852 [djberge@qw s] Another alternative is to wrap "closed? or close" in a "if
  | 2856 [matz@ru y- a] If you close the handle for the temporary file, you can no longer have
  + 2855 [flori@ni e. ] It is removed, if it is garbage collected, I assume. That would be a

ruby_connect's error handling
2851 [mdorey@bl ea] lipper.ens.fr
2854 [jlst@ge ty g] I can't address your discussion of multiple connect()'s in the presence

cgi.rb: option to omit HTTP header emission
2853 [jos@ca no k.] I'm trying to use cgi.rb to write HTML-only output. This patch adds a
2886 [jos@ca no k.] What needs to be done for this patch to become part of the official cgi.rb? If
2887 [quixoticsyco] I don't know the context of the problem you are solving, but if you don't
2888 [matz@ru y- a] * converting (Japanese) text encoding.
2890 [jos@ca no k.] My face is so red. I was missing the obvious. Thanks to both of you for

RDoc ignores second argument of include.
2857 [akr@m1 n. rg] I found that RDoc ignores second (and maybe subsequent) argument of
2859 [dave@pr gp o] Thank you for this report. It should be fixed in the latest CVS.

RDoc doesn't make a link for inner module inclusion.
2860 [akr@m1 n. rg] I found that RDoc doesn't make a link for inner module inclusion as

RDoc doesn't make a link for a included module
2861 [akr@m1 n. rg] I found another link problem.
2862 [dave@pr gp o] To fix this, I changed the way symbol lookup works in HTML generation.

RDoc: parent class link problem.
2863 [akr@m1 n. rg] I found that RDoc doesn't make a link for a parent class in following

memory leak when using continuations+weakrefs?
2864 [slepnev_v@ra] the following code eats up more and more memory as it works, pretty

Need help with NKF and encodings
2865 [dave@pr gp o] ...
2866 [nobu.nokada@] Like this?
2868 [dave@pr gp o] ...
2870 [matz@ru y- a] I can see my name in Japanese correctly.

[BUG] ruby/dl
2867 [quixoticsyco] # dltest.rb

YAML::Store: {":..." => true}
2869 [zn@mb .n ft ] % cat a.rb
2871 [nobu.nokada@] Seems syck's bug.

ruby-talk : 98651
2872 [decoux@mo lo] From the Changelog
2873 [matz@ru y- a] Oops, I forgot to restore $SAFE value before execution.  Thank you for

minor tweak to visibility parser in rdoc
2874 [ryand-ruby@z] I have a lot of code that opens itself up if it is being unit tested at
2910 [dave@pr gp o] A variant of your patch has been applied - thanks.
2914 [ryand@ze sp ] "def method arg" winds up rendering to something visually equivalent
2915 [dave@pr gp o] Oh, OK - I misunderstood.

Tempfile problem
2875 [flori@ni e. ] Tempfile seems to use DelegateClass instead of SimpleDelegator in Ruby
2880 [matz@ru y- a] This was caused by delegate.rb incompatibility.  It will be fixed

yaml last update
2876 [ksibilev@be ] $ ri local
+ 2877 [ksibilev@be ] $ irb
| 2882 [ruby-core@wh] Thankyou.  This is repaired.  I'm restructuring the flow of multiline
+ 2883 [ruby-core@wh] I can't seem to reproduce this.  Update from CVS, try again.  If it

Bug in open-uri under win32 (?)
2878 [batsman.geo@] batsman@tux-chan:/tmp$ cat sdfsdfsd.rb

[PATCH] tempfile.rb (was Re: Bug in open-uri under win32 (?))
2879 [batsman.geo@] The bug has been traced back to Tempfile, which doesn't use the binary
2881 [matz@ru y- a] CVS 1.8 open-uri calls binmode for tempfile, to solve this problem.

2889 [quixoticsyco] 0.0e+00
2891 [ocean@m2 cc ] Thank you for the report.

Re: [PATCH] tempfile.rb
2892 [matt@li ke .] P.S. I frequently use Tempfile as a way to place stuff in a file that

[Help] Problem with SOCKET API on OpenVMS
2893 [BCoish@Dy ax] TCP/IP socket routine(s) not functioning properly on OpenVMS.
2900 [matz@ru y- a] It's not OpenVMS specific problem.  Socket.gethostbyname returns
2902 [BCoish@Dy ax] Thanks for the speedy reply!
2903 [decoux@mo lo] ^^^^^                        ^^^^^
2904 [BCoish@Dy ax] Thank you for your assistance.  It was quite a stupid assumption