18907-19217 subjects 19385-19707

[Feature #249] wish list item: binding.set_local_variable
19142 [redmine@ru y] Issue #249 has been updated by Roger Pack.
19145 [meinrad.rech] ...

[Bug #461] Re: Hard question: Any ideas why this code leaks memory?
19143 [redmine@ru y] Issue #461 has been updated by Roger Pack.

[Feature #249] wish list item: binding.set_local_variable
19144 [redmine@ru y] Issue #249 has been updated by Roger Pack.

19146 [transfire@gm] Anyone else interested in creating a RINQ?
19147 [transfire@gm] Oops. I meant to post this to ruby-talk, not ruby-core.

[Bug:1.9] rubygems depend on test/unit/ui/console/testrunner
19168 [mame@ts .n .] Executing gem with --test may fail because of requiring
+ 19179 [ryand-ruby@z] I'll get Eric to fix this today.
+ 19191 [drbrain@se m] Is it also ok if I check in some windows RubyGems fixes at the same
  + 19311 [mame@ts .n .] I think Yugui or Matz has a duty to reply to this :-(
  + 19312 [yugui@yu ui ] What are the fixes?
    19355 [drbrain@se m] RubyGems now works on windows platforms (which don't have
    19476 [yugui@yu ui ] OK. commit them.

[Bug:1.9] some tests still seem to depend on test/unit
19169 [mame@ts .n .] $ ruby19 test/json/runner.rb
19178 [ryand-ruby@z] I'll track those down and fix em.

[Feature #619] improved performance of BASIC_OP_UNREDEFINED_P
19171 [redmine@ru y] Feature #619: improved performance of BASIC_OP_UNREDEFINED_P

[Bug #620] #dup'd module autoload broken
19181 [redmine@ru y] Bug #620: #dup'd module autoload broken

[Bug #620](Closed) #dup'd module autoload broken
19182 [redmine@ru y] Issue #620 has been updated by Nobuyoshi Nakada.

[Bug #345] 1.9 racc appears to seg fault
19192 [redmine@ru y] Issue #345 has been updated by Roger Pack.
19193 [ryand-ruby@z] please change the title of the bug. this doesn't have to do with racc,

[Ruby 1.9 - Bug] super in extension library and ruby script
19199 [nobu@ru y- a] OK, let's start a new ticket.

[Bug #622] merging nested hashes
19200 [redmine@ru y] Bug #622: merging nested hashes
19211 [martindemell] merge doesn't do a deep copy

performance of C function calls in 1.8 vs 1.9
19218 [pbrannan@at ] It seems it is more expensive to call a C function in 1.9 than it used
19220 [jameskilton@] Hmm, this is disheartening. I ran these tests on my machine, here are
19255 [michael.seli] ruby1.8 test.rb
+ 19256 [jameskilton@] Mike, I think you misunderstand what Paul is pointing out. Compare the
| 19259 [michael.seli] Unless I am completely off-beam here, my understanding of the original
+ 19277 [pbrannan@at ] Interesting.  I wonder why ruby method invocation is so much faster for
  + 19278 [jameskilton@] @Michael
  + 19281 [charles.nutt] The dvar vs lvar shouldn't make a difference, should it? In both cases,
  | + 19283 [ko1@at ot ne] It is a known issue.
  | | + 19286 [charles.nutt] In JRuby we do the same, generally speaking, but we also special-case
  | | + 19292 [michael.seli] expr.each { |*args|
  | |   19293 [ko1@at ot ne] I committed such an optimization last night :)   please check it.
  | |   19295 [michael.seli] That's fantastic - the performances of "for" and "each" are now quite
  | + 19285 [charles.nutt] The "fasterness" of 1.8.6 appears to be because the original bench did
  + 19291 [michael.seli] Only default ones - nothing special at all. I am using Ubuntu Linux 8.04

Module.freeze vs Object.freeze
19225 [curth@mi ro ] What's the difference between Module.freeze and Object.freeze?  They seem t=
+ 19236 [nobu@ru y- a] Module#freeze resolves the name of anonymous module.
+ 19245 [matz@ru y- a] It initializes internal name field before freezing it.  The following
  19246 [curth@mi ro ] Thanks, but I don't see any difference in 1.8.6 (patchlevel 111).  Is this =
  19248 [nobu@ru y- a] Because you do it within irb.
  19249 [Tomas.Matous] I don't see a different still.

Fwd: Build problems on OS X PPC
19231 [mike@st k. a] ...
19235 [nobu@ru y- a] I think r19700 fixed it.
19241 [mike@st k. a] I have just successfully compiled & tested r19725.  Thanks.

[Bug #622](Closed) merging nested hashes
19234 [redmine@ru y] Issue #622 has been updated by Nobuyoshi Nakada.

Regexp Order Matters in 1.9
19242 [james@gr yp ] $ ruby_dev -ve 'if /(?<a>.)/ =~ "abc"; puts a end'
19243 [akr@fs j. rg] /(?<a>.)/ =~ "abc" is a special syntax which is not overridable.
19244 [james@gr yp ] Interesting.  I didn't know that.
19247 [david@da id ] Furthermore, the regexp that includes the named capture must appear

default_internal encoding
19250 [dave@pr gp o] I'm documenting default_internal for the PickAxe, and have a couple of
+ 19254 [michael.seli] I am the guy that write the "default_internal" patch, most of which was
+ 19266 [matz@ru y- a] I think default_internal is a per process (or VM) state so that magic
| 19272 [michael.seli] I agree that default_internal is global to a ruby process, but it is
| 19273 [matz@ru y- a] I admit I once considered using magic comments, but unlike shehang
| 19274 [nobu@ru y- a] It's not good for Emacs; -*- encoding: euc-jp:utf-8 -*- causes
| 19276 [james@gr yp ] $ ruby
+ 19270 [duerst@it ao] I have followed the ruby-dev discussion fairly closely, but

[Bug #626] stop test-all at test/rdoc/test_rdoc_ri_driver.rb
19251 [redmine@ru y] Bug #626: stop test-all at test/rdoc/test_rdoc_ri_driver.rb

[Bug #627] Time#_dump and Time#_load
19252 [redmine@ru y] Bug #627: Time#_dump and Time#_load

base64.rb is back!
19253 [dave@pr gp o] I thought we'd removed it because it was a trivial wrapper around pack/
19287 [matz@ru y- a] It's newly added to support RFC 4648, which is slightly incompatible

default_internal for CSV (was Re:  Re: default_internal encoding)
19257 [james@gr yp ] Sorry to thread hijack, but Michael need your advice.
19263 [michael.seli] Yes, that seems right - the main thing is that the strings returned are in
+ 19269 [duerst@it ao] I haven't looked at your code in detail, but probably, yes.
| 19271 [michael.seli] My preference would actually be that default_internal is always set, so
+ 19280 [james@gr yp ] Great.  Change made.  Thanks.

test/unit and minitest
19258 [ryand-ruby@z] I've reverted test/unit to its original state (*). Further, I've

[Bug #627](Closed) Time#_dump and Time#_load
19260 [redmine@ru y] Issue #627 has been updated by Nobuyoshi Nakada.

Re: [ruby-cvs:26958] Ruby:r19740 (trunk): Added minitest 1.3.0
19261 [nobu@ru y- a] `svn mv' could keep the continuities and properties.

[Bug #628] Assigning to $0 truncates value
19262 [redmine@ru y] Bug #628: Assigning to $0 truncates value

[Bug #628](Rejected) Assigning to $0 truncates value
19264 [redmine@ru y] Issue #628 has been updated by Nobuyoshi Nakada.

[Bug #629] Test/Unit Seems Broken
19265 [redmine@ru y] Bug #629: Test/Unit Seems Broken
19267 [ryand-ruby@z] should be fixed. I'm bad at svn.

offline dev-meeting
19268 [ko1@at ot ne] We'll hold offline dev-meeting with matz, yugui-san (1.9.1 release

ruby.h defines HAVE_RUBY_NODE_H
19275 [pbrannan@at ] Now that node.h is no longer public, should this macro be removed?

[Bug #627] Time#_dump and Time#_load
19279 [redmine@ru y] Issue #627 has been updated by Wilson Bilkovich.
19290 [nobu@ru y- a] Probably.

[Bug #632] StringIO has an odd alias implementation [?]
19282 [redmine@ru y] Bug #632: StringIO has an odd alias implementation [?]
19288 [dblack@ru yp] static VALUE

[Feature #578] add method to disassemble Proc objects
19284 [redmine@ru y] Issue #578 has been updated by Roger Pack.

[Bug #633] dl segfaults on x86_64-linux systems
19289 [redmine@ru y] Bug #633: dl segfaults on x86_64-linux systems

[Bug #634] Time parsing works in 1.8 but not 1.9
19294 [redmine@ru y] Bug #634: Time parsing works in 1.8 but not 1.9

racc parser generator
19296 [aaron@te de ] I'm just curious...  Why is only the racc runtime distributed with ruby?
+ 19297 [ryand-ruby@z] agreed. especially since the package seems mostly unmaintained(*).
+ 19346 [aaron@te de ] *bump*
  19767 [ryand-ruby@z] As a reminder to the core team for when they get home and rest up.

[Feature #639] New String#encode_internal method
19298 [redmine@ru y] Feature #639: New String#encode_internal method
19303 [matz@ru y- a] I'd rather propose, default_internal to be the default target encoding
19306 [michael.seli] Yes, that is very neat.
19309 [michael.seli] I just tried String#encode with no parameter, but I was hoping to just be
19327 [matz@ru y- a] Indeed.  I will fix.
19338 [michael.seli] Thank you - it works very well now :)

[Feature #640] New Array#encoding_pack -> string method
19299 [redmine@ru y] Feature #640: New Array#encoding_pack -> string method

[Feature #619](Closed) improved performance of BASIC_OP_UNREDEFINED_P
19300 [redmine@ru y] Issue #619 has been updated by Koichi Sasada.

[Bug #641] GC patch to cache the most recent heap for is_pointer_to_heap
19301 [redmine@ru y] Bug #641: GC patch to cache the most recent heap for is_pointer_to_heap

1.8.6, make test-all errors on MinGW
19302 [luislavena@g] I started to update the recipes for One-Click installer to work with

1.9, encoding & win32 wide char support
19304 [lloyd@hi ai ] I've recently hacked 1.8.7 on win32 to use Wide char CRT & system
+ 19305 [matz@ru y- a] It's something we should do before the release, so we are VERY
| 19321 [lloyd@hi ai ] Great!  I'm glad this work has the potential of being useful.  I'll
| 20108 [billk@ct .c ] Does anyone have information as to the current status of
| 20109 [billk@ct .c ] Ugh.  Sorry, I mean of course: Unicode-savvy path handling
| 20110 [matz@ru y- a] Every path encoding is UTF-8 and converted to UTF-16 internally.  If
| 20111 [billk@ct .c ] Awesome.  This is excellent news!
+ 19308 [billk@ct .c ] VERY interested!!! :D

[Feature #639](Closed) New String#encode_internal method
19307 [redmine@ru y] Issue #639 has been updated by Yukihiro Matsumoto.

gem install with --test says a warning
19310 [mame@ts .n .] When I execute gem install with --test, it says "Gem::SourceIndex#search
19356 [drbrain@se m] Thank you, I will apply it to RubyGems and import it to trunk.

[Bug #642] __DIR__
19313 [redmine@ru y] Bug #642: __DIR__

[Bug #642](Rejected) __DIR__
19314 [redmine@ru y] Issue #642 has been updated by Shyouhei Urabe.
19316 [transfire@gm] Yes. Of course. My apologies for overlooking it.
19322 [matz@ru y- a] Of course there is.  Done.

[Feature #643] __DIR__
19315 [redmine@ru y] Feature #643: __DIR__
+ 19317 [ryand-ruby@z] yes.
| + 19318 [jimfreeze@gm] No, no, no Ryan. You have it all wrong.
| + 19319 [now@bi wi se] I realize that I'm responding to a joke, but wouldn't it perhaps be
| | + 19320 [jimfreeze@gm] Like a built it #split(/\//). Maybe call it explode. (The current
| | + 19328 [nobu@ru y- a] I like this idea too.
| + 19326 [duerst@it ao] Just some alternatives for everybody. What about
| + 19331 [transfire@gm] It's one thing to be sarcastic, it is another to be untruthful. If you
|   19347 [aaron@te de ] Yes Ryan.  Please stop the lies.  It hurts us all.
|   19349 [transfire@gm] On Oct 15, 11:00=A0am, Aaron Patterson <aa...@tenderlovemaking.com>
|   19398 [pit.capitain] Tom, I could understand if you left the community, but I would very
+ 19324 [phil@ha el .] If we don't want to add another toplevel constant, another way to handle
  19325 [dave@pr gp o] But with relative_require, suddenly most of those use cases evaporate,
  19330 [transfire@gm] =A0

[Bug #641] GC patch to cache the most recent heap for is_pointer_to_heap
19323 [redmine@ru y] Issue #641 has been updated by Roger Pack.

[Bug:1.9] tests for miniunit have bad chemistry with other tests
19329 [mame@ts .n .] The current tests for minitest seem to use at_exit, which lets the tests
19339 [ryand-ruby@z] Fixed

Can I confirm a change to source file encoding
19332 [dave@pr gp o] A month ago, if I had
19333 [matz@ru y- a] It's intentional.  Sorry for the last minute changes.
19334 [dave@pr gp o] Should I report the following as a bug?
19336 [matz@ru y- a] Yes, it's a bug.  Thank you.
19344 [matz@ru y- a] Although I thought it was a feature bug against consistency, it found

[Bug #633] dl segfaults on x86_64-linux systems
19335 [redmine@ru y] Issue #633 has been updated by Rados??aw Bu??at.

[Bug #647] Source encoding not applied to 7-bit regular expressions
19337 [redmine@ru y] Bug #647: Source encoding not applied to 7-bit regular expressions

[Bug #647] Source encoding not applied to 7-bit regular expressions
19340 [redmine@ru y] Issue #647 has been updated by Yui NARUSE.

ruby-mode for Emacs - :"should this work?"
19341 [augustlillea] ...

[Bug #649] Memory leak in a array assignment?
19342 [redmine@ru y] Bug #649: Memory leak in a array assignment?
21809 [brent@mb ri ] $ ruby18-mbari -v -e "a = Array.new(1000000){'a'};
21811 [radek.bulat@] I saw other topic about MBARI patches and wonder what is the status of
21921 [rogerdpack@g] They're being merged in slowly as Brent Roman has time to convert them
21931 [giuseppe.bil] I assume this means that they will see the light in some 1.9.x release
21936 [brent@mb ri ] Obviously, I cannot say when these patches will "see the light", but

Yet another block semantic/syntax question
19343 [dblack@ru yp] I just want to make sure I'm right that this is what happens, and that
+ 19359 [cohen.jeff@g] I hope that's not intentional :-)  Sounds like a bug to me.
+ 19367 [matz@ru y- a] It's a bug.  ;-<
  + 19368 [mame@ts .n .] $ ./ruby -e '
  + 19369 [dblack@ru yp] What is it supposed to do? Meaning: should Proc.new {|a,b=1| } behave
    19371 [matz@ru y- a] It's suppose to behave like lambda here.

[Bug #647](Rejected) Source encoding not applied to 7-bit regular expressions
19345 [redmine@ru y] Issue #647 has been updated by Yukihiro Matsumoto.

Net::HTTP form_data Array support
19348 [kleshwong@gm] ...

Net::HTTP.post_form bug : can't post form to correct uri which contains QueryString(QueryString part are lost) and revise
19350 [kleshwong@gm] When I tried to do post_form to some url like
19352 [chiology@gm ] You are trying to use GET-style query params instead of POSTing the
19363 [kleshwong@gm] Dear Matt
19364 [chiology@gm ] From my experience, it's simply easier to process requests that way,
19366 [kleshwong@gm] Thanks,
19393 [meta@po ox c] The specification for URIs is RFC3986, and the specification for HTTP
19395 [Tim.Bray@Su ] Quoting from the top of RFC3986: "Obsoletes: 2732, 2396, 1808 "
19510 [meta@po ox c] Like I said, the previous version.
+ 19516 [rm_rails@ch ] Surely Net::HTTP should follow the HTTP spec,
+ 19630 [duerst@it ao] RFC 2616 is currently under revision in the httpbis WG in the IETF.

[Bug:1.9] `rake gem' emits a warning
19351 [mame@ts .n .] rake seems to call Gem.manage_gems which is deprecated at
+ 19353 [jim.weirich@] The gempackagetask should probably be allowed to continue to work with
| 19354 [drbrain@se m] require 'rubygems/builder'
+ 19362 [jim.weirich@] Done.

timeout use in protocol.rb....why?
19357 [charles.nutt] Ok, I give up. Why is this (from net/protocol.rb)...
19360 [dave@pr gp o] Might it be a portability issue with Windows and select?
19361 [charles.nutt] if (WINDOWS) ...

[Bug #654] Net::SSH known_hosts.rb, pageant.rb
19358 [redmine@ru y] Bug #654: Net::SSH known_hosts.rb, pageant.rb

[Bug #655] Query part lost when using Net::HTTP.post_form function
19365 [redmine@ru y] Bug #655: Query part lost when using Net::HTTP.post_form function

RDoc bug fixes
19370 [drbrain@se m] Missing space in ri list output.
19477 [yugui@yu ui ] Thank you. Commit them.

assert_in_delta(1.4, 1.4, 0) failes in minitest
19372 [akr@fs j. rg] assert_in_delta(1.4, 1.4, 0) fails in minitest.

capture_io in minitest
19373 [akr@fs j. rg] capture_io changes $stdout.fileno.
+ 19467 [nobu@ru y- a] No chance to use reopen inside the block?
| 19469 [akr@fs j. rg] What's the advantage?
| 19472 [nobu@ru y- a] It could restore the originals even if STDOUT.reopen were used
| 19490 [ryand-ruby@z] It couldn't back in the day... that was what I was using. Matz gave me
| 19500 [nobu@ru y- a] To where?
| 19511 [ryand-ruby@z] to the code in my perforce repository... I'm still not done deciding
+ 19489 [ryand-ruby@z] I just got back from vacation so I'll look at this and the others you
  19495 [matz@ru y- a] Yeah, the great lesson that you cannot trust me always. ;-<
  19496 [james@gr yp ] I'm shattered.  :)

assert_raises(mod) in minitest
19374 [akr@fs j. rg] assert_raises(mod) in minitest fails even if the raised
19375 [kou@co mi ng] This change causes other incompatible.
19377 [akr@fs j. rg] Oops.  You are right.

19376 [akr@fs j. rg] csv test suite in test/csv/ is not tested by "make test-all" because
19380 [james@gr yp ] I renamed all tc_* files to test_*.  Hopefully that fixes this issue.

Constant names in 1.9
19378 [dave@pr gp o] When Ruby makes the tIDENTIFIER/tCONSTANT test, it looks to see if the  first letter is upper case using
19379 [matz@ru y- a] And its definition is
+ 19381 [dave@pr gp o] I think you're correct, because otherwise it would be card to have a
| 19382 [matz@ru y- a] I think that's a different story, although it is very interesting.
+ 19383 [Tim.Bray@Su ] I think this is correct.
  19401 [david@da id ] The consensus seems to be to leave the current rules as they are, and I
  19406 [duerst@it ao] Excluding the somewhat complicated issue of the Georgian script

multibyte case conversion (Re:  Re: Constant names in 1.9)
19384 [matz@ru y- a] Now I feel clearer now.  It's not too difficult to use Lu property
+ 19386 [Tim.Bray@Su ] I think Java is generally considered to get this reasonably correct.
+ 19387 [michael.seli] I believe that there are a few more issues eg: context sensitivity, a
  + 19392 [matz@ru y- a] I agree with you here.  I'd like to seek the line where we can make
  | 19394 [Tim.Bray@Su ] 1. Allow case-mapping for the double-wide ASCII compatibility
  + 19405 [duerst@it ao] The Greek Sigma may be seen as such a case, but actually, there may