2894-3184 subjects 3293-3547

^ 1.8.2: Segfault / Addition
3098 [elven elven.] Sorry, i was being overly stupid; i pasted the wrong code snippet.
+ 3099 [phubuh phubu] Weird.  'nil if begin end' is enough to trigger the bug, by the way.
+ 3100 [phubuh phubu] The culprit is that `begin end' creates a begin NODE whose nd_body is

^ gdbm abort - OSX
3102 [dave pragpro] Just before I start another debugging session, has anyone seen this, or
3103 [nobu.nokada ] Recompile gdbm.so.
3104 [dave pragpro] Both it and Ruby were clean compiles: I installed a fresh gdbm that
3105 [nobu.nokada ] The reason of `unknown type 0x22' is that ruby after Mar 4
3106 [dave pragpro] Thank you - I should have looked for that. The interesting thing is
3108 [nobu.nokada ] Well, it could occur also when ruby 1.9 loads an extension for
3111 [dave pragpro] That was my question :)

^ -L'%s'  problem with mingw32
3107 [quixoticsyco] The mingw32 compiler (3.3.1) does not accept -L'/lib/path'.
3109 [nobu.nokada ] I've never seen that trouble, what do you mean by "the mingw32
3110 [quixoticsyco] I use cygwin but with a pure mingw32 toolchain from mingw.org.  I do

^ rss library
3112 [dave pragpro] I was browsing the new rss library code and noticed that it has
3120 [kou cozmixng] In <7329EAA0-C983-11D8-B645-000A95676A62@pragprog.com>
3122 [dave pragpro] The biggest difference I see if the spacing: the other library files
+ 3124 [kou cozmixng] In <D9AEA7AE-CD6D-11D8-9AE0-000A95676A62@pragprog.com>
+ 3126 [nobu.nokada ] Not only rss, see lib/*ml.
  3128 [dave pragpro] I hadn't noticed that until now.

^ Problem in RSS library, or problem in my blog :)
3113 [dave pragpro] I've been trying to use the new RSS library to parse a number of
3121 [kou cozmixng] In <D0A4C1A2-CB15-11D8-B645-000A95676A62@pragprog.com>
3123 [dave pragpro] But my blog is 0.9, and the RSS library still can't read it. If I
3125 [kou cozmixng] In <651875D8-CD6E-11D8-9AE0-000A95676A62@pragprog.com>
3129 [dave pragpro] If I understood your earlier message correctly, your library cannot
3143 [kou cozmixng] In <4357B35D-CDC1-11D8-9AE0-000A95676A62@pragprog.com>
3144 [dave pragpro] That's wonderful - thanks. I'll change my example in the PickAxe

^ assert_timeout for Test::Unit::Assertions
3116 [kirya85 mail] I write subj. Patch in attachment.
3117 [nobu.nokada ] Your implementation leaves the thread alive, and changes

^ SOAP maintainer
3118 [dross code-e] Who maintains SOAP4R? would it be possible to get implementation notes? The
+ 3119 [dross code-e] I would be delighted to help with the SOAP4R as well, after I finish up
+ 3127 [nahi keynaut] I'm the maintainer.  Here's the code snipet to create untyped SOAP

^ Base64.encode and Base64.decode
3130 [gsinclair so] A little while ago, the 'base64' methods were sensibly moved into

^ [yaml] reducing Hash in Array
3131 [nobu.nokada ] Moin,

^ Reporting RI-documentation corrections ?
3132 [holmberg iar] What is the preferred way to report corrections
3133 [dave pragpro] Probably an e-mail here.
+ 3134 [gsinclair so] Me too...
+ 3135 [halostatue g] Speaking of ri documentation, is there anywhere that documents the
  + 3137 [dave pragpro] I'm working on it. In the meantime, the Ruby source isn't a bad guide.
  + 3140 [gsinclair so] Gavin
    3145 [mailinglists] Hmm using Array.flatten as an example method is maybe not very clever. You
    3147 [dave pragpro] added it to the description, but should probably add it to the top as
    3149 [holmberg iar] Nice to know that there is a source where one can look for info
    3153 [dave pragpro] It really doesn't matter, but it would probably be nice to be

^ [doc-patch] Wrong rdoc formatting in {array,pack}.c
3136 [holmberg iar] Here is another small documentation patch.
3162 [ocean m2.ccs] I've committed [ruby-core:3132] and [ruby-core:3136].

^ [doc] Inconsistent indentation of rdoc entries ?
3138 [holmberg iar] The embedded doc entries in for example array.c are not consistently
3139 [dave pragpro] I agree absolutely
3142 [holmberg iar] I understand.

^ rexml/validation/validationexception is missing.
3141 [nobu.nokada ] `make test-all' fails since validationexception.rb is missing.
3146 [ser germane-] Grr.  Yes.  CVS really needs a decent "stat" command.
+ 3148 [nobu.nokada ] What's about HEAD?
| 3151 [ser germane-] charset="iso-8859-1"
+ 3152 [ruby-lists l] cvs up should show

^ [PATCH] Array#index with block
3150 [g_ogata optu] there was discussion on ruby-talk[1] about letting Array#index (and

^ [PATCH] Nonblocking socket connect - Win32 - 181
3154 [jean-francoi] charset="us-ascii"

^ Re: UPDATE - [PATCH] Nonblocking socket connect - Win32 - 181
3155 [jean-francoi] charset="US-ASCII"
3157 [usa garbagec] Thank you!
3182 [jean-francoi] What's the status of this patch?

^ Installation of ri when installing 1.8.2
3158 [ruby jamesbr] The current CVS for 1.8.2 does not install ri unless you either run
3160 [ser germane-] charset="iso-8859-1"

^ OT: Request for a Japanese Speaker
3159 [dave pragpro] Sorry to intrude for something off-topic, but I've been trying to get

^ CVS Head with mingw on Windows
3161 [eule space.c] I posted on ruby-lang earlier, but got no reply.

^ Looking for reviewers
3163 [dave pragpro] I'm getting towards the end of the Second Edition of the Programming
+ 3164 [sujithkumar ] I am planning to learn Ruby. If you would like comments when a total newbie
+ 3165 [nobu.nokada ] Which libraries?
| 3166 [dave pragpro] Everything in lib/ and ext/
+ 3169 [evan falling] Dave,

^ [doc] Inconsistent "call-seq" usage etc.
3167 [holmberg iar] Looking through the existing RDOC documentation I found another
3168 [dave pragpro] That's because I had the information available for the C stuff, but
3175 [holmberg iar] I realize now that there are several differences in the HTML
+ 3176 [dave pragpro] Call-seq for Ruby methods was added this morning :)
+ 3186 [florgro gmai] Actually I would also welcome a way to document methods that aren't
  3187 [dave pragpro] You can already do this in 'C' using 'Document-method: and
  3188 [florgro gmai] Yup, that would be very great.

^ [doc-patch] Another rdoc formatting error in array.c
3170 [holmberg iar] Thanks.
3172 [ocean m2.ccs] (2004/07/11 02:09)
3174 [holmberg iar] In general I think the changes are in the right direction (but I'll
3178 [ocean m2.ccs] Yes, that's I intended. I understand Array.each means Arrays's class method.
3183 [ocean m2.ccs] Sorry, I lied. This is not fixed yet. After fix this, I'll commit current working.

^ binding a URL to a label in RDoc
3171 [ian caliban.] This is a port of Perl's Foo::Bar[http://www.foo.com/bar].
3173 [dave pragpro] I just backported a feature from RubLog: as of the latest CVS you can
+ 3177 [ian caliban.] Thanks a lot. That's great.
+ 3179 [ian caliban.] I've tried this out and it now works for multi-word labels separated on
  3181 [dave pragpro] Try now: I previously tested it with {Perl's Foo:Bar}, but not Perl's
  3211 [ian caliban.] That works nicely now. There's still one minor issue with labels,

^ Contributors: ready for 1.8.2 preview1?
3180 [matz ruby-la] Or do you have anything left for preview1?

^ Text folding in YAML
3185 [don na.rim.o] 1.8.1 and CVS 1.8.2.  However, no problem occurred using Ruby 1.9.0.

^ [Bug] Ruby 1.8.1 and 1.9: interactions between extensions: opengl and imlib2
3189 [schur cpsc.u] I am trying to use the opengl and imlib2 extensions with Ruby 1.8.1 and 1.9.

^ Bug in OpenGL module (Re: [Bug] Ruby 1.8.1 and 1.9: interactions between extensions: opengl and imlib2)
3190 [nobu.nokada ] It's a bug in rbogl, imlib2 isn't concerned with the SEGV
3191 [schur cpsc.u] Many thanks for the quick reply and patch.

^ Documentation fix for String#ljust rjust center
3192 [mneumann nte] The pad string argument of these methods is undocumented.
+ 3194 [florgro gmai] I've written a small tool which does this for me. It works reasonable
+ 3198 [dave pragpro] In the book source, where this all came from, the values were generated

^ [rdoc] client side imagemap patch (rdoc/diagram.rb)
3193 [mneumann nte] Konquerer does not like image maps with coordinates where x1 > x2 or

^ Patch file.c (dev_minor dev_major)
3195 [mneumann nte] appended is a patch that adds the dev_major and dev_minor methods to the
3196 [matz ruby-la] It seems fine.  Can you prepare ChangeLog line for this addition?
3197 [mneumann nte] new methods File::Stat#dev_major and #dev_minor.

^ Trying to understand \G
3199 [dave pragpro] I'm being silly again, but I can't get \G to work with String.index. If
3200 [matz ruby-la] '\G' only works for repeating match methods, String#gsub, String#scan,
+ 3201 [dave pragpro] str = "abcdefg"
| 3202 [matz ruby-la] \G matches at the end of point of the last match.  The next match
+ 3203 [nobu.nokada ] Or when starting position is given.
  + 3204 [matz ruby-la] Done (in HEAD).
  | 3205 [nobu.nokada ] Regexp#=~ doesn't accept nil now.
  | 3206 [matz ruby-la] Thank you, and thank you.
  + 3227 [dave pragpro] I thought I understood this, but when I tried to document it, I found
    3231 [matz ruby-la] \G forces <rest> to match immediately after the last match (or
    3236 [dooby d10.ka] Ah, "last" meaning "most recent" (contrasted with "final").
    3239 [dave pragpro] words = 'ant bee cat, dog elf fox'

^ lib/.document for 1.8.2
3207 [akira ruby-l] <URL:http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=252871>.

^ [BUG] Ruby doesn't ignore Byte Order Mark when run with -Ku
3208 [florgro gmai] Moin!
3215 [matz ruby-la] It makes sense.  Let me think about it for a while.

^ Unit tests in the CVS repository
3209 [ser germane-] charset="utf-8"
3214 [matz ruby-la] You can put a new directory under the /test with test suites inside.

^ Ruby Specification
3210 [drossruby ya] Request. Can someone create a ruby specification? I

^ Readable/Writeable modules
0265 [mneumann nte] Appended is the patch for RCR 265.

^ Typo and grammar/style fixes for ext/win32ole/win32ole.c
3213 [jos catnook.] The attached patch attempts to create a more consistent style for error
3256 [jos catnook.] Has anybody had a chance to review this patch yet? All the changes are
3257 [matz ruby-la] The patch is OK for me, but I still want to wait for the maintainer
3258 [jos catnook.] Agreed. Thank you for the feedback, Matz.

^ Re: Incident Analysis of the intrusion on helium.ruby-lang.org May 2004
3216 [ser germane-] charset="utf-8"
3221 [shugo ruby-l] Sorry, ruby-committers ML was configured to permit mails from members
3222 [ser germane-] charset="iso-8859-1"

^ Re: Incident Analysis of the intrusion on helium.ruby-lang.org May 004
3217 [dooby d10.ka] A pleasant journey amidst the Periodic Table of the Elements.
3218 [ser germane-] charset="iso-8859-1"
3220 [dooby d10.ka] That's exactly how I read you and it's easy to believe

^ yaml module feature question about String Tag
3219 [nahi keynaut] _why, is this expected behavior?
3223 [ruby-core wh] This appears to be correct translation.  From Section 2.3 of
3224 [nahi keynaut] Thanks.  I searched the spec to find only Section 4.3.5.  I had to look

^ Transportation Request
3225 [jwebb gol.co] Taiwan was featured in last night's news report on TV. While in Taiwan I

^ Incident Analysis of the intrusion on helium.ruby-lang.org
3226 [shugo ruby-l] As already reported, helium.ruby-lang.org, which is one of the servers

^ Core support for Gems, and namespace
3228 [luke madstop] I'm getting to the point where my ruby libraries are sophisticated enough
+ 3229 [dave pragpro] The answer to the second part of your question is versioning. The
| 3232 [luke madstop] I don't understand how versioning solves the problem of being forced to
| 3233 [gsinclair so] A good approach might be the following.  Let's assume your application
| 3235 [luke madstop] That's exactly the behaviour I want to avoid.  It seems that that is
| 3237 [rich infoeth] Luke,
| 3241 [luke madstop] [SNIPped discussion of load path complexity]
| 3244 [drbrain segm] =20
| 3254 [luke madstop] Ah.  I hadn't thought about the fact that definition of a class also
+ 3230 [halostatue g] Here's what I do: I package both as a gem and as a .tar.gz file. I
  3234 [luke madstop] [SNIP]
  3238 [halostatue g] There isn't one -- yet. An alternative solution is also being worked
  + 3242 [luke madstop] Given how well both FreeBSD and Gentoo has done with this, I have high
  + 3243 [gsinclair so] [long and interesting post snipped because there's not one particular
    + 3248 [halostatue g] Well, there's a few problems.
    | 3250 [gsinclair so] That's mostly false.  The binary stub loads the most recently
    | 3253 [halostatue g] Um, you just made my point for me on both counts. The convenience
    | 3255 [gsinclair so] OK, that's a fair way of looking at it.  I just figured that it's a
    + 3249 [batsman.geo ] You are implying that the options are
      3251 [gsinclair so] I certainly didn't mean to imply that.  I was just fishing for
      3252 [batsman.geo ] RPA is meant to make life easier for the upstream developer. He'd require

^ Is this a commit miss?
3240 [ksibilev bel] * eval.c (rb_eval): copy on write for argument local variable
3246 [matz ruby-la] I've tried some big change last night at HEAD.  I think I fixed it.

^ cgi.rb possible memory leak
3245 [ksibilev bel] We are almost ready to deploy our commercial website completely written
3247 [matz ruby-la] CGI_PARAMS and CGI_COOKIES are obsolete.  So we can (and should)

^ [PATCH] Infinite recursion in rdoc
3259 [batsman.geo ] While generating Borges' rdoc documentation, rdoc began to demand in
3260 [dave pragpro] Thanks for digging down and finding that. I'll apply the patch shortly.

^ [Security] security hole was found on PuTTY
3261 [ocean m2.ccs] PuTTY team anounced serious security hole was found.

^ Issue with Dir.glob documentation
3262 [halostatue g] I noticed this because I'm working on some stuff and needed to do an

^ Issue with Dir.glob documentation
3263 [halostatue g] I noticed this because I'm working on some stuff and needed to do an
3264 [ocean m2.ccs] (2004/08/09 00:00)
3265 [halostatue g] I am using File.join *only* because C comments are borken. I would

^ 1.8.2 preview bug?
3266 [chad chadfow] I haven't had time to really investigate this, but I've boiled it down
3267 [dooby d10.ka] I think this is what should happen (and _does_ on Windows).
3268 [ocean m2.ccs] (2004/08/10 11:49)
3278 [chad chadfow] # (2004/08/10 11:49)
+ 3279 [batsman.geo ] IIRC something like that was reported in the past, and the answer (from
+ 3281 [dblack wobbl] The latter, I think :-)

^ YAML problem, possibly?
3269 [hgs dmu.ac.u] I obtained a largish lump of shallow XML, succcessfully read it with
+ 3270 [hgs dmu.ac.u] I used D last night on one machine, but it blows up on this one in
| 3271 [ser germane-] charset="iso-8859-1"
| 3272 [hgs dmu.ac.u] I remembered over lunch that I did fixup some bugs in the XML in
+ 3274 [ruby-core wh] With a complex tree like a REXML::Document, dumping to YAML can be
  3276 [hgs dmu.ac.u] Thank you.

^ Latest warnings from 1.8.2 p2
3273 [djberg96 yah] Ruby 1.8.2 preview 2

^ rubytests in Ruby 1.8.2?
3275 [djberg96 yah] Will rubytests (the project currently on RubyForge)
3277 [chad chadfow] # Hi,
3285 [matz ruby-la] We don't want to ship THREE test suites in the distribution.  I am not
3289 [chad chadfow] # Hi,

^ [syck] non-ascii byte and allocator
3280 [nobu.nokada ] Since July 4, syck fails to load yaml which contains non-ascii

^ Re: syck.so and AIX
3282 [halostatue g] dbx /usr/local/bin/ruby

^ (none)
3283 [jose-alegria] charset="iso-8859-1"

^ s=(1..255).to_a.pack("C*"); p(YAML.load(s.to_yaml)==s) #=> false
3286 [akira ruby-l] s=(1..255).to_a.pack("C*")

^ [BUG] IO#reopen crashes
3287 [florgro gmai] Moin!
3288 [decoux moulo] This is ALLOC(OpenFile) without a MEMZERO() which is the problem.

^ select warns about argcount, map doesn't
3291 [dblack wobbl] $ ./ruby -v

^ [PATCH] dir.c --- Dir.chdir error handling
3292 [holmberg iar] I think I have found a bug in Dir.chdir.
3350 [matz ruby-la] Your patch will be merged.  Thank you.
3351 [ocean m2.ccs] I think memory `cwd' leaks when dir_chdir raised exception. (chdir_restore has same issue)
3352 [matz ruby-la] Thank you.  Can you commit the fix?  I think 'volatile' before
3353 [nobu.nokada ] Rather, shouldn't rb_ensure guarantee the argument?
3354 [matz ruby-la] Hmm, you mean like this?
3356 [nobu.nokada ] Well, I meant just about preventing a work object from GC;
3369 [ocean m2.ccs] Sorry for late posting. Typhoon striked me.....
3372 [nobu.nokada ] In that case, v should be stored somewhere in the function and
3374 [ocean m2.ccs] (2004/09/10 10:22)
+ 3375 [ocean m2.ccs] (2004/09/10 23:05)
+ 3376 [nobu.nokada ] It results in rb_call0(), which stores the receiver at
  3378 [ocean m2.ccs] (2004/09/11 00:01)
  + 3380 [ocean m2.ccs] I realized this is thread stuff, so should
  | 3382 [nobu.nokada ] Since current ruby threads never switch while pure C code is
  + 3383 [nobu.nokada ] Seems fine, but this wouldn't be needed.
    3384 [ocean m2.ccs] (2004/09/13 00:40)
    + 3385 [matz ruby-la] I'm not sure about your concern.  If the string from rb_str_new()
    | 3386 [ocean m2.ccs] Well, what I concern is it's difficult to confirm that ruby object is really protected.
    | + 3387 [decoux moulo] Well, not really : in [ruby-dev:19854] there is a problem because `args'
    | | 3392 [ocean m2.ccs] I hope I understood you.....
    | | 3393 [matz ruby-la] Right.
    | | 3394 [ocean m2.ccs] No, I meant compiler optimization "inline expansion". So, please imagine
    | | 3395 [decoux moulo] Sincerely, I really don't understand : [ruby-dev:19854] is a case of
    | | 3399 [ocean m2.ccs] OK, so why are these functions not volatiled like [ruby-dev:19854] or Bignum#&, Bignum#|?
    | | + 3402 [matz ruby-la] rb_intern() never allocates memory before calling strdup().  No other
    | | | 3410 [ocean m2.ccs] memory allocation? Why memory allocation?
    | | | 3411 [ocean m2.ccs] Until now, my concern was just theory, but I could reproduce the problem
    | | | 3412 [matz ruby-la] The problem under artificial circumstance may not be a real problem.
    | | | 3414 [ocean m2.ccs] What do you mean 'artificial'? This happens only if rb_str_intern is inlined
    | | | 3415 [matz ruby-la] You'd show us the possibility, that's OK.  But it's a very small
    | | + 3403 [decoux moulo] It's a little more complex : I can give you an example with 1.8.0
    | |   3404 [matz ruby-la] Interesting.  I'm happy to see the bug fixed.  But I'm not sure what
    | |   3405 [decoux moulo] Well, this bug is really, really special and you must look at the
    | |   3409 [matz ruby-la] <snip the explanation>
    | |   3416 [decoux moulo] There are 2 problems
    | |   + 3417 [matz ruby-la] <snip the explanation>
    | |   | 3418 [decoux moulo] Well, this is with the old 1.8.0 and with the example that I've given
    | |   + 3435 [matz ruby-la] A question: I have removed frame->argv from CVS HEAD recently.  Does
    | |     3436 [decoux moulo] not fatally : the problem exist each time that a pointer is on the stack,
    | |     3438 [matz ruby-la] Then after removing frame->argv, there should be no problem, right?
    | |     3439 [decoux moulo] yes,
    | |     3448 [matz ruby-la] Perhaps "stupid" in French means "brilliant" ;-)  Since Guy is one of the
    | + 3388 [matz ruby-la] As Guy stated in [ruby-talk:03387], the VALUEs must be referenced from
    + 3389 [nobu.nokada ] The latter.  VALUEs are not pointers, so difference in
      3390 [ocean m2.ccs] (2004/09/14 01:07)
      3400 [ocean m2.ccs] VisualC++ warned this in C mode.