12087-13354 subjects 12618-13282

12294 [david@da id ] Nobu has just checked in a String.force_encoding method, which alters
+ 12295 [matz@ru y- a] The original #encode was supposed to do
| 12297 [david@da id ] I'm glad to hear that you still expect encode to be the common method.
| 12299 [celtic@sa ry] If I may chime in here, some encoding names include non-symbol symbols,
+ 12298 [duerst@it ao] Sorry for doing half of the work in Japanese.
  12303 [gwtmp01@ma .] In my quick look at the source it seems that
  12304 [duerst@it ao] As said already, there will most probably be a shorter method name

Will 1.8.6 remain compiled with VC6?
12305 [luislavena@g] Since last year (in July if memory works) there was a discussion about
+ 12306 [halostatue@g] This isn't really a core Ruby question. This should be asked on
| + 12307 [luislavena@g] Austin, One-Click-Installer rely on official builds made by Ruby-Core
| | 12309 [halostatue@g] I don't actually agree here. Curt goes with the VC 6 build because it's
| | 12311 [luislavena@g] That was my point.
| | 12359 [ml.chibbs@gm] I know this not the right place to post this, but I'll start here
| | + 12362 [celtic@sa ry] I sent you some mail off-list regarding this. I realise you'll probably
| | + 12363 [luislavena@g] I mailed you a few months back regarding OCI... guess it get trapped
| |   12365 [ml.chibbs@gm] Maybe.  :-(
| + 12308 [znmeb@ce ma ] Yes, let's take this to Ruby-Talk so we can all participate. Most of the
| | + 12310 [luislavena@g] Maybe you're right, cross-posting this could get more answers from
| | + 12315 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
| |   12316 [luislavena@g] I was working on documenting the whole MinGW/MSYS experience, so far,
| |   12317 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
| |   12319 [nobu@ru y- a] Of course it does work.  If you were correct, how could I debug
| |   12320 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
| |   12322 [nobu@ru y- a] Though I haven't debugged ruby-prof, I've used many other
| + 12318 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
|   12380 [hramrach@ce ] What's the problem here? The whole point here is that you do want to
|   + 12387 [halostatue@g] The CRT 8 files (VS2005) are SxS assemblies, intended to be installed
|   | 12405 [hramrach@ce ] If the above is supposed to mean that vc8 runtime libraries cannot be
|   | + 12407 [luislavena@g] We could go with the MinGW path, which is compatible with msvcrt (and
|   | + 12408 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
|   + 12392 [luislavena@g] I see there is no problem using the C/C++ Runtime 2005 redistributable
|   + 12413 [ml.chibbs@gm] The problem is that you can't safely mix a Ruby built with
|     12415 [cfis@sa ag x] This is a cryptographically signed message in MIME format.
|     12640 [duerst@it ao] I tried that once, but it was dead slow (about three seconds
|     12660 [znmeb@ce ma ] Well ... I run it on a 1.3 GHz Athlon T-bird, so it must be something in
+ 12321 [jack@jn so t] If there is a move to a later VC, let me suggest using VC9. I just

Need Japanese Help - VRuby & new One-Click Ruby Installer with patch 110
12312 [curt.hibbs@g] ...
12313 [usa@ga ba ec] I check the problem, and contact nyasu-san (author of VRuby/SWin)
12314 [usa@ga ba ec] Hmm, how did you get the binary of ruby, Curt?
12366 [ml.chibbs@gm] That must be the problem, because I am using ruby-mswin32.
12367 [usa@ga ba ec] I'm planning to release it.

Dir.chdir patch for MS Windows
12328 [Daniel.Berge] Dir.chdir without any arguments will usually raise an error on MS
+ 12340 [luislavena@g] Daniel, that will fail on a Corporative environment.
| 12341 [Daniel.Berge] I'm in a corporate environment. Works here.
| 12343 [luislavena@g] \Users\Developer
+ 12347 [nobu@ru y- a] In 1.9, when ENV['HOME'] isn't set, it will be set using
  12348 [djberg96@gm ] Ok. Can we patch that into 1.8.x?

gc.c refactoring
12338 [hgs@dm .a .u] ...

patch to implement Array.permutation
12344 [david@da id ] ...
12345 [matz@ru y- a] Thank you.  I tried to implement it, but screwed up.  Please don't

a patch for Array.permutation and a generalization of Array.product
12346 [david@da id ] ...

ruby19 omits program name for missing files
12351 [drbrain@se m] $ ruby19 -v

system and Errno::ENOENT in 1.9
12352 [drbrain@se m] $ ruby -ve 'system "no_such_file"'
12661 [nobu@ru y- a] Yes.

[Patch] Bug 6400: Time#+ and Time#- not honoring subclass
12371 [ron.leisti@g] ...

Release compatibility/train
12372 [Prashant.Sri] ...
12373 [matz@ru y- a] Yes.
12374 [david@da id ] Also, if I understand it correctly, Ruby 1.9 will remain an odd version
12376 [Prashant.Sri] Matz and David,
12377 [matz@ru y- a] Yes, the versioning scheme has been changed since 1.9, to utilize the
12382 [charles.nutt] Is 1.9.1 now going to be considered a stable release?
12385 [matz@ru y- a] We sincerely hope so, although being not sure how big that hope is.
12388 [charles.nutt] So would the recommendation be that people running production
12389 [matz@ru y- a] Honestly speaking, 1.9.1 at the end of this year would not be
+ 12391 [ed.odanow@wo] If "not be production ready" means, there can be several bugs in it, this is
| 12423 [matz@ru y- a] This is the case.
+ 12393 [hgs@dm .a .u] It sounds like the more people who test it the better.  Is there a
| 12396 [ed.odanow@wo] I use usually snapshots from "http://dl.ambiweb.de/mirrors/ftp.ruby-lang.org/",
| 12403 [hgs@dm .a .u] ...
+ 12406 [dblack@ru yp] Am I right that 1.9 is mostly a development version toward 2.0?
  + 12410 [rick.denatal] If I understand what Matz said earlier on this thread, not really.
  | + 12416 [charles.nutt] Hmm, that's not what I got. I was still under the impression that 1.9.1
  | | + 12417 [james@gr yp ] I'll be pretty surprised if people don't start porting after a time
  | | + 12419 [rick.denatal] My impression came from Matz's response to you
  | | + 12427 [matz@ru y- a] Even though the primary focus of the 1.9.1 release is expose the new
  | |   12608 [ser@ge ma e-] charset="iso-8859-1"
  | |   12609 [matz@ru y- a] It doesn't, as long as it keeps compatibility to pre-feature-freeze.
  | |   12610 [matz@ru y- a] I am vaguely thinking that feature freeze would happen around the end
  | |   12611 [hgs@dm .a .u] Is my revised http compression patch under consideration?  It's gone
  | + 12426 [matz@ru y- a] Right, except that since 1.9 has some incompatibility from 1.8, there
  + 12425 [matz@ru y- a] As 1.8 was a development version toward 1.9 ;-)
    12430 [greg@cy er u] You're kidding right?
    12431 [matz@ru y- a] No kidding.  I consider myself as a good designer, not a great

some typos in doc/comments in repos/ruby/trunk
12375 [eugoss@gm il] ...
12390 [matz@ru y- a] Thank you.  I just did.

Re: [SUMMARY] gc.c -- possible logic error?
12378 [sylvain.joye] ...

discrepancy between String#oct and String#hex
12379 [ryand-ruby@z] There appears to be a "feature" in ruby (1.8) in String#oct. Namely,
12381 [matz@ru y- a] It's one of the things we have inherited from Perl, oct does interpret

Include Rake in Ruby 1.9
12383 [nakahiro@sa ] RubyGems is about to be merged.  Excellent.  And the next critical
+ 12395 [lukfugl@gm i] Will Rake be "fully" bundled (core), or "pre-installed-gem" bundled
| + 12397 [drbrain@se m] What difference does this make?
| | + 12398 [blowmage@gm ] ...
| | | 12402 [drbrain@se m] Then it doesn't matter which method is chosen, unless you are using
| | + 12399 [lukfugl@gm i] I'm mostly thinking of the discussion in the thread regarding
| |   12400 [luislavena@g] +1 for Jacob and Mike comments. Having rake as gem gives us the
| + 12409 [vjoel@pa h. ] Does "core" still mean that the lib is built into the interpreter and
|   + 12420 [tmorgan99@gm] ...
|   + 12422 [drbrain@se m] I think there's a mixture of terms here.  I took "core" to mean in
+ 12576 [nakahiro@sa ] Thank you very much for your swift response.  Sorry for long silence.
+ 12582 [charles.nutt] FWIW, on JRuby we install the Rake gem as part of our build/dist
+ 12696 [nakahiro@sa ] Sorry to keep you waiting.
  12723 [halostatue@g] That's a problem, since most people don't work with Rake as a library,
  + 12728 [drbrain@se m] This is a feature of gems.
  | 12739 [halostatue@g] Only works for a rake stub that uses RubyGems.
  + 12773 [nakahiro@sa ] When an user newly installs rake with gem command, 'rake' command will
    12778 [halostatue@g] RubyGems doesn't do system-specific handling that way. It will try to
    12785 [drbrain@se m] RubyGems does not make this check.  It will clobber an existing bin/
    12791 [halostatue@g] Ok. That solves the worry that I had, but introduces another worry for

tIDENTIFIER and is_local_id()
12401 [Thomas.Enebo] There are a few places where the parser (in trunk and 1.8.x) checks to

ruby-1.9 build Sun-sparc-solaris2.9
12404 [hgs@dm .a .u] Things went very smoothly.  I did
12421 [steven@lu os] I think this is just some dumb fallback feature of pkg-config.  E.g.

rdoc patch
12411 [david@da id ] ...
+ 12428 [drbrain@se m] Thanks.  Can you file a bug in the RDoc project?  It may get lost
+ 12429 [lukfugl@gm i] Do we really want to be catching and suppressing all Exceptions there?
  12432 [drbrain@se m] Probably, since RDoc is light in the tests department, I'd have to

12412 [ser@ge ma e-] Hey,
12565 [kou@co mi ng] I've fixed the tests of RSS Parser.
+ 12607 [ser@ge ma e-] charset="iso-8859-1"
+ 12633 [kou@co mi ng] No. I just only updated 1.9.

Performance Issues with nil, true, false as Hash Keys (1.8.6)
12414 [matthias@wa ] This is a followup of [ruby-talk:272146].
12424 [nobu@ru y- a] Perhaps, `if (IMMEDIATE_P(a))' equals it and would be faster a bit.

Patch for configure.in
12418 [rbrown@ge to] ...

Re: Compile error for HEAD
12529 [naruse@ai em] First I tried
12530 [naruse@ai em] Oops, i used old libruby.so.

Re: arbitrary Unicode characters in identifiers?
12531 [ed.odanow@wo] Even if not recommended, it is sometimes helpful (O.K. - it makes fun too ;-) )
12532 [ed.odanow@wo] Sorry - question went to wrong thread :-(

Quick Enumerable#sort_by internals question
12533 [Daniel.Berge] In enum.c there's this bit of code in the sort_by_i function.
12534 [dooby@d1 .k ] $res = $k = nil # reduce warnings (daz)

Re: Why does def return nil?
12537 [zimbatm@or e] Classes don't return nil but the last used object. Try : class X;
12540 [zimbatm@or e] ...
+ 12543 [halostatue@g] No, because the Symbol is instantiated by def foo anyway.
+ 12544 [Daniel.Berge] I suppose it might also mean refactoring undef_method and remove_method.
+ 12559 [ryand-ruby@z] You're doing that simply by defining the method in the first place.
  12597 [zimbatm@or e] Hm, thanks for the insight (and Austin Ziegler too).
  12615 [brent@mb ri ] Ryan,

Ordered Hashes in 1.9?
12539 [mneumann@nt ] It really seems that 1.9 got ordered hashes by default. Can anyone
12542 [matz@ru y- a] Yes, it preserves order of keys inserted.
+ 12546 [mental@ry ia] What data structure is used to track the insertion order?  It seems like
| 12548 [lopx@ga et .] It's linked hash, every hash entry has *fore and *back fields.
+ 12549 [david@da id ] Is this an implementation detail, or is it part of the specification?
| + 12551 [vjoel@pa h. ] A related question: what does the following output?
| | 12553 [matz@ru y- a] Hash order does not affect equivalence.
| | 12554 [eric@se pe e] still in austin   busy  work
| | 12555 [eric@se pe e] oops  wrong email  please ignore
| + 12552 [matz@ru y- a] That's the point we have to discuss with other implementers before
| + 12557 [murphy@ru yc] If it becomes specification, we might want to add Array-style handling
|   12560 [eric.mahurin] I would guess that the implementation of the ordering is a linked
|   12561 [david.stokar] ...
+ 12556 [murphy@ru yc] Very nice!
  + 12558 [matz@ru y- a] Your observation is right.
  + 12563 [now@bi wi se] I'd rather have leaner classes than a #methods method that returned
    12567 [murphy@ru yc] Well, Hashes and all symbol tables use st_table, which is highly

Bug in win32ole
12547 [cwaters@ne w] I found a bug in win32ole when it is used to extract information from an

[PATCH] to let Ruby 1.9 to process signals while blocked in ext/readline
12550 [brent@mb ri ] ...
12562 [david.stokar] ...

Re: net/http.rb, patch, possible bug, test problem
12564 [hgs@dm .a .u] That's the way I found out how to do it -- i.e. it seems a popular

Revised net/http patch
12566 [hgs@dm .a .u] ...

$" and require
12568 [tmorgan99@gm] ...
+ 12569 [drbrain@se m] $ ruby19 -vrfileutils -e 'p $"'
+ 12570 [murphy@ru yc] irb(main):001:0> require 'set'
  + 12571 [murphy@ru yc] ...
  + 12572 [drbrain@se m] For backwards compatibility.  Its in the ruby-core archives.

File::Stat#grpowned? is wrong
12573 [lists@be tr ] user $ ls -lad /etc/asterisk /etc/cron.daily/

12574 [cyrus_dev@ho] ...

Marshal.dump doesn't dump string encoding
12575 [david@da id ] Matz,
12577 [matz@ru y- a] We know.  It's going to be dumped soon.

Possible memory leak in ruby-1.8.6-p110??
12578 [znmeb@ce ma ] I haven't had a chance to narrow this down in enough detail yet, but
12595 [znmeb@ce ma ] 1. It works fine on 32-bit.
12596 [hgs@dm .a .u] Is it worth trying -O2, -O1 ?
+ 12600 [tmorgan99@gm] ...
| 12601 [luislavena@g] There was a security fix (Net:HTTP with secure connections) and the
+ 12602 [znmeb@ce ma ] Eventually I might do that. I'll be testing the "-g" without "-pg" at
  12606 [hgs@dm .a .u] OK, thought it might be scripted and easy to run over night.
  12612 [znmeb@ce ma ] Well -- removing "-pg" left the issue in place, so it's something to do
  12613 [hgs@dm .a .u] gcc (GCC) 4.1.1

iconv enhancement in Ruby 1.9
12579 [eugoss@gm il] I noticed Ruby 1.9 had the class Iconv improved. In 1.9 it became
12580 [nobu@ru y- a] Once I thought to use a block for exception handling.
12583 [duerst@it ao] When I first had a look at the Iconv class in Ruby,
12585 [nobu@ru y- a] This idea is that Iconv#iconv and others yields error info on

ruby.exe exit code 255
12581 [dsears@gm il] A little background: I'm trying to get Ruby on Rails running under

Broken Ruby/ODBC due rbconfig COMMON_MACROS
12584 [luislavena@g] I'm trying to pin-point the problem we have compiling ruby-odbc 0.9995
12586 [usa@ga ba ec] We just hit a Platform SDK's bug.

Confusion about arities
12587 [charles.nutt] It seems like a number of methods have unexpected arities. For example,
12589 [drbrain@se m] The -1 arity exposes the argument checking implementation, which is
12591 [charles.nutt] Ideally we'd be able to use (and present) the actual arity to do
+ 12592 [ryand-ruby@z] ...
+ 12594 [drbrain@se m] The information is there, but I don't see a way to automatically make
  12603 [charles.nutt] Ok, I understand now.
  12614 [drbrain@se m] NODE_CREF could have nd_argc holding the calling convention as it
  12616 [charles.nutt] Yeah, seems like it would be a good thing to do. Only took me about

MatchData#select rdoc and arity incorrect
12588 [charles.nutt] ~/NetBeansProjects/jruby $ ri MatchData#select
+ 12590 [drbrain@se m] is out of date.
| 12593 [ryand-ruby@z] ...
+ 12604 [matz@ru y- a] The code was wrong.  Perhaps I was confusing MatchData to provide

[PATCH] adds __file__ and __line__ methods to Methods and Procs
12598 [brent@mb ri ] ...

private method dispatch changes dropped from 1.9?
12599 [david@da id ] I'm guessing that these mean that the plans to alter the method name
12605 [matz@ru y- a] Yes, it was too radical.

Question about heap_slots in gc.c
12617 [h.lai@ch ll ] I'm trying to modify the Ruby interpreter's garbage collector. At the
+ 12620 [hgs@dm .a .u] static int heap_slots = HEAP_MIN_SLOTS;
+ 12648 [sylvain.joye] What are you trying to do ? I did some experimental work on the current GC,
  12649 [h.lai@ch ll ] I'm trying to improve the Ruby GC's copy-on-write friendliness. You can
  12651 [matz@ru y- a] Moving marking bits to separate bitmap tables makes Ruby's GC more COW