13491-13985

^ autoconf (in 1.8 at least) and feature defs vs. macros
13491 [lists-ruby s] I'm fiddling around trying to get Ruby 1.8.6-p110 to build on a Stratus VOS
13493 [hgs dmu.ac.u] I will be happy to give such a patch a spin on sun-sparc-solaris2.9

^ RñÇ. :  Re: Proposal: Array#walker
13497 [tad.bochan b] Whats wrong with

^ state of threads in 1.9
13498 [mumismo gmai] Are Threads mapped to threads on the underlying operating system in
+ 13505 [znmeb cesmai] You mean "changelog" in the sense of differences between 1.8 and 1.9
| 13507 [mumismo gmai] - A new interpreted (previously YARV)  2.5x faster on averaged
+ 13506 [luislavena g] Native Threads.
+ 13516 [david davidf] As others have responded, Ruby 1.9 uses native threads.  But you still
  13518 [Daniel.Berge] Can't we have both native *and* green threads? There are advantages and
  13519 [billk cts.co] No multi-core support... ever?  I had presumed native threads were
  + 13520 [luislavena g] They didn't say that :-P
  | 13549 [hramrach cen] If the only fear is that extensions won't be thread-safe why not let
  + 13526 [drbrain segm] No multi-core support yet.
    + 13545 [sylvain.joye] Would not it be possible to release the global lock manually in C
    | 13563 [mental rydia] There is rb_thread_blocking_region()...
    | 13574 [gethemant gm] Wasn't there a talk of possibility of allowing release of global lock
    + 13547 [mumismo gmai] It is really bad news for me. My application is really heavy threaded
    | + 13554 [znmeb cesmai] Do you have any detail/references on that? I'm very interested in the
    | | 13555 [mumismo gmai] This is the best thing I have found
    | | 13556 [znmeb cesmai] That's fantastic!! Thanks!!
    | + 13576 [drbrain segm] If this is something you really want to have happen, the best way to
    | + 13582 [charles.nutt] You may wish to try JRuby, which allows concurrent native threading. In
    |   13593 [murphy rubyc] do you have some code example for this? I'm not sure if I understand the
    |   + 13643 [charles.nutt] Run any heavily multi-threaded code on a multi-processor system; you'll
    |   + 13651 [mental rydia] It's an implementation detail.  From a programmer's standpoint, the main
    + 13557 [mumismo gmai] But a high level language is far different from a kernel.
    | 13562 [mental rydia] Unfortunately, Ruby is rife with shared, mutable objects: classes
    + 13579 [billk cts.co] Hmm.  In 1.8 we currently have multiple C++ threads each talking to
      13580 [ko1 atdot.ne] You can do it on 1.9.  See rb_thread_blocking_region() in thread.c.
      13583 [billk cts.co] Nice!  Thanks!!!! :D
      13584 [ko1 atdot.ne] Note that 1.9 thread interface is not tested, so we need more
      13590 [billk cts.co] OK, I will begin writing a test program - thanks.  :)
      13592 [ko1 atdot.ne] Future work.  Near future.

^ Proc#hash returns different values for same body
13513 [ed.odanow wo] I recognized a behaviour of Proc#hash that does not follow the description.
+ 13514 [david davidf] The hash method works consistently with ==, but both have misleading
| 13517 [ed.odanow wo] Is read the discussion in http://www.ruby-forum.com/topic/110696
+ 13515 [ed.odanow wo] I made another comparison with Ruby 1.8...
  13534 [matz ruby-la] by 1.9=20
  13568 [rick.denatal] Ruby 1.9

^ test/unit and miniunit
13528 [ryand-ruby z] When is the 1.9 freeze?
+ 13529 [ryand-ruby z] I should add, and since it is 10% of the code, it turns out to be
+ 13533 [transfire gm] sounds good. but what happens to all the other runners? might miniunit
| + 13535 [cohen.jeff g] Normally I would agree, but since miniunit sounds like it's 100%
| + 13541 [drbrain segm] You install test::unit classic and use the runners.
| | 13550 [transfire gm] will the require be the same -- require 'test/unit' ? If so might
| | 13573 [drbrain segm] If miniunit replaces the existing version, it would be.
| | 13615 [Daniel.Berge] I've used Tim Pease's "turn" runner (it is a runner, isn't it?). I think
| + 13585 [ryand-ruby z] I entirely disagree. Losing capabilities is exactly what we need right
|   + 13587 [transfire gm] You misunderstand me. I don't mean that the capabilities are included
|   | 13603 [james graypr] Specs are crazy popular right now.  It seems really great for Ruby to
|   | + 13604 [halostatue g] I'm not sold on specs, personally, but why not test/spec instead or
|   | + 13606 [transfire gm] On Nov 16, 7:58 am, James Edward Gray II <ja...@grayproductions.net>
|   | + 13609 [luislavena g] This sounds like the not-so-popular business practices Microsoft or
|   + 13602 [james graypr] Yeah, I think I have to agree with that.
+ 13586 [akr fsij.org] I like miniunit.
  13588 [ryand-ruby z] I will take care of all of that and ensure that I've got the same

^ [PATCH] mswin32-vc6 segmentation fault due ruby_in_eval wrong definition
13536 [luislavena g] Working in collaboration with Aman Gupta, we were trying to make
+ 13552 [Daniel.Berge] =20
+ 13558 [nobu ruby-la] Why do you need to access it?  To create AST without executing?
  13559 [luislavena g] Nobu, thank you for answering.
  13560 [nobu ruby-la] Many UNIX variant systems don't have that kind of asymmetry.
  13561 [luislavena g] The joys of using Windows I guess ;-)

^ Iconv#iconv returning wrong object
13542 [dirk.traulse] Iconv.iconv(to, from, *strs)
13543 [nobu ruby-la] may want to write about Iconv.iconv?
13589 [dirk.traulse] Definitely not!
13591 [nobu ruby-la] Sorry, confused was me.
13594 [dirk.traulse] OK.
13610 [nobu ruby-la] It breaks backward compatibility, and it doesn't make much
13697 [dirk.traulse] Because it is a lot easier to learn/remember the methods if they behave

^ Thoughts about Array#compact!, Array#flatten!, Array#reject!, String#strip!, String#capitalize!, String#gsub!, etc.
13564 [ed.odanow wo] Good evening all together!
+ 13565 [david davidf] Matz has added Object.tap to Ruby 1.9 which is intended for use in
| + 13567 [murphy rubyc] even better, we can do
| + 13571 [ed.odanow wo] I've forgotten the "tap". This is better than to find new method names.
+ 13575 [now bitwi.se] On Nov 15, 2007 8:14 PM, Wolfgang N=E1dasi-Donner <ed.odanow@wonado.de> wro=
  + 13577 [ed.odanow wo] It's a typical situation in using map/filter/reduce paradigm from functional
  | 13595 [now bitwi.se] On Nov 15, 2007 11:38 PM, Wolfgang N=E1dasi-Donner <ed.odanow@wonado.de> wr=
  | 13596 [ed.odanow wo] Well - I said that I might have problems to express in English what I mean :-)
  + 13578 [mneumann nte] "I am a 20MB string".tap {|i| i.downcase!}.split(/\s+/).tap {|i|

^ Ignoring generated files
13566 [murphy rubyc] could somebody please add bmlog-* to the svn:ignore of trunk?

^ Ruby 1.9 code name?
13569 [murphy rubyc] to answer myself: a colleage just reminded me that tap-bang-tap-bang

^ Embedding Ruby 1.9 in multithreaded app
13570 [mneumann nte] I've a multithreaded (pthread) application that uses one Ruby interpreter. Is it
13572 [mneumann nte] Oops, I missed that other thread about threads :)

^ Array#split_at
13597 [murphy rubyc] This is a multi-part message in MIME format.
13608 [nobu ruby-la] What's this?
13620 [murphy rubyc] This is a multi-part message in MIME format.

^ wondering about #tap (was: Re: Thoughts about Array#compact!, Array#flatten!, Array#reject!, String#strip!, String#capitalize!, String#gsub!, etc.)
13598 [dblack rubyp] It's interesting that almost every example of #tap ends up saying that
+ 13605 [transfire gm] It possible to define tap slightly differently by returning a functor
| 13612 [dblack rubyp] str.downcase!
| + 13621 [transfire gm] Well, I basically agree. But tap is in the language now, right?
| + 13623 [now bitwi.se] str.downcase.split...
|   + 13626 [dblack rubyp] I'm assuming we're talking about a case where someone wants (for
|   | + 13639 [ed.odanow wo] I'm somehow surprised about this long discussion, based on my short question,
|   | | 13641 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
|   | + 13700 [now bitwi.se] Sorry.  That should have been
|   + 13628 [transfire gm] dont you mean?
|     13702 [now bitwi.se] I was way too sloppy.  It should have been
+ 13624 [now bitwi.se] def cleverly_return_an_array_after_building_it
| + 13625 [ryand-ruby z] No. I don't think we do see the point is valid.
| | 13699 [now bitwi.se] No.  /You/ don't see the point as being valid.
| + 13627 [dblack rubyp] But my point is that *all* examples seem to be horrible, so a horrible
| | 13701 [now bitwi.se] OK.  What follows is the scanner.rb file from my ruby-lisp library.
| + 13631 [james graypr] def build_array
|   13703 [now bitwi.se] This only works for Arrays, not Objects in general.
+ 13630 [martindemell] Maybe more along the lines of "#tap is an extremely useful debugging
| + 13633 [james graypr] Yeah, I feel like that's it's primary usage.  We have p() and similar
| | 13655 [martindemell] In fact, I wonder how much of the usage of tap a #p method that was
| + 13642 [lists-ruby s] I'm guessing you all saw Evan Weaver's rant last night?
+ 13632 [david davidf] The use case that occurs to me (I'm not sure that it will convince you
  13634 [dblack rubyp] I wouldn't use the term "fluent" to describe this new and, to my eye,
  13636 [rick.denatal] I've got to agree with D.A.B. here.  No offense D.F. but to my eye
  13637 [murphy rubyc] I like the code. and I like languages that allow all different
  + 13638 [rick.denatal] de gustibus!
  + 13640 [ed.odanow wo] I agree. It should be up to the user to decide what stile he likes or not. Ruby
    13644 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
    + 13645 [transfire gm] some_expression.
    | 13659 [halostatue g] Meh.
    | + 13660 [transfire gm] Don't think I even seen that before.
    | + 13661 [james graypr] if =85 {
    |   + 13662 [halostatue g] Well everyone KNOWS which is the One True Way.... ;)
    |   + 13666 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
    |     13667 [ed.odanow wo] That's clear. It means that Ruby added the rule "if a
    |     13668 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
    |     13669 [ed.odanow wo] I think it was somewhere in the large amount of posts
    |     13670 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
    + 13647 [ed.odanow wo] I see only one problem, as I named some time ogo...
      + 13649 [dblack rubyp] This message is in MIME format.  The first part should be readable text,
      + 13656 [alex blackke] Surely that's fixable for irb, though?  It's got the _ variable to play
        13677 [lists-ruby s] Isn't it possible to have some

^ Re: [PATCH] CGI::Session::PStore partitioned directories
13600 [akr fsij.org] What's happen if someone creates a symlink from /tmp/0 to
13601 [maurice.aubr] What's a safe approach?
13663 [drbrain segm] You'd need to securely create a directory in /tmp to hold your
13731 [maurice.aubr] Would it be acceptable to default the depth to 0 then, and

^ railroad problem
13607 [srinivasan.a] This is a multi-part message in MIME format.
13611 [luislavena g] T24gTm92IDE2LCAyMDA3IDEwOjQyIEFNLCAgPHNyaW5pdmFzYW4uYW5hbmRAYWNjZW50dXJlLmNv

^ Enumerable#zip vs. Enumerable#everything_else
13613 [dave pragpro] Enumerable.zip now returns an Enumerator, rather than an array.
13616 [matz ruby-la] Enumerable#map etc. now returns Enumerator too, if no block is given.

^ Suggestion for native thread tests
13614 [eustaquioran] I'm checking 1.9 and I'm curious for some test which shows me that there
+ 13650 [ko1 atdot.ne] I can't understand what do you mean.  Where is a FIFO?  On my
| + 13654 [eustaquioran] On Linux it can be created with
| + 13657 [eustaquioran] On Linux it can be created with
|   13658 [eustaquioran] Sorry my messages are going twice - maybe some weird stuff on the
+ 13671 [charles.nutt] Under JRuby it appears to continue running, and the results are
  13672 [eustaquioran] Yeah, the threads should run and show the results intermingled even if
  13674 [eustaquioran] On Nov 17, 2007 2:25 PM, Eust=E1quio Rangel <eustaquiorangel@gmail.com> wro=
  13678 [ko1 atdot.ne] I fixed this issue on my laptop.  I'll commit after refactoring.
  13683 [charles.nutt] My understanding of Ruby IO is that the example given ought to be
  13692 [ko1 atdot.ne] Yes.  We must release global VM lock (GVL).  In this case, I fixed
  13734 [eustaquioran] I made an code update here and it's running just fine here now,

^ Some Core Questions
13617 [saladin.mund] This is a multi-part message in MIME format.
13629 [fxrubyguy gm] These questions are probably better suited for the ruby-talk list,

^ segfault in ostruct with 1.8.6, where to get help?
13618 [aktxyz gmail] I am using alot of ostructs.  Everything was fine with the 1.8.5 I was
13646 [rogerpack200] run it in gdb, see if it gives you a better backtrace (?)
13648 [Daniel.Berge] ruby 1.8.6 (2007-09-24 patchlevel 111) [i386-mswin32]
13653 [aktxyz gmail] Yeah, I ran some solo tests like that also.  I finally replaced

^ Change in signal handling
13619 [dave pragpro] I've noticed differences in the handling of SIGCLD in 1.9. For

^ Suggestion for native thread tests
13622 [eustaquioran] I'm checking 1.9 and I made a test with two threads reading two files

^ Keyword arguments "spec"
13635 [rick.denatal] It seems that there are two aspects to adding keyword arguments in Ruby.
13673 [matz ruby-la] Keyword argument "spec" is very fragile.  You cannot trust what I said

^ pty seems to be broken
13652 [dave pragpro] This might be related to the general signal handling issues I seem to

^ Observe variable set and variable change
13664 [saladin.mund] This is a multi-part message in MIME format.
13665 [celtic sairy] You'll probably get many more responses (and very quickly :)) if you ask

^ Class Code?
13675 [ed.odanow wo] Good evening!

^ Failing to compile trunk under Ubuntu
13676 [B.Candler po] I tried compiling trunk yesterday and today, on two different Ubuntu 6.06
+ 13679 [znmeb cesmai] What version of GCC do you have? What revision number of trunk is this?
+ 13680 [B.Candler po] $ gcc -v
| 13682 [znmeb cesmai] $ bison --version
+ 13681 [david davidf] I had this same error message until I upgraded bison from 2.0 to something

^ Problems with \M-x in utf-8 encoded strings
13685 [ed.odanow wo] I recognized a problem when using special character modifying sequences
+ 13686 [murphy rubyc] And each_char shouldn't throw an index error...no?
| + 13687 [ed.odanow wo] This always happens in Ruby 1.9 if the encoding expects more bytes, which are
| + 13694 [duerst it.ao] It would be nice if each method that handles characters throws
+ 13693 [duerst it.ao] Good catch!
  + 13698 [ed.odanow wo] I think it is difficult to define what \M-a or \C-ãâ¥ã should mean.
  + 13717 [ed.odanow wo] I found the following in http://svn.ruby-lang.org/repos/ruby/trunk/ChangeLog

^ base64.c vs. base64.rb
13688 [transfire gm] for TMail. I looked at them and they do basically the same exact
+ 13689 [mblondel rub] Base64 has been marked to be unbundled in the email "What libraries to
| 13691 [transfire gm] Really? Why's that?
| 13720 [sroberts uni] Its supported by Array#pack.
| 13726 [murphy rubyc] yes, but
| 13728 [murphy rubyc] even
+ 13690 [transfire gm] Base64.encode64('a')
  13719 [sroberts uni] Kindof. Base64 is defined in the MIME specs, so you are "supposed" to

^ Build failure trying to use rb_define_alias on rb_mKernel
13704 [Daniel.Berge] =20
13705 [nobu ruby-la] Fixed it yesterday, I guess.  What revision do you build?
13706 [Daniel.Berge] Sorry, forgot to mention. This was against 1.8.6-p111.
13707 [nobu ruby-la] * eval.c (rb_alias): do not call hook functions until initialization

^ Minimal embedded Ruby 1.9 application
13708 [mneumann nte] I'd like to embed Ruby 1.9 into my application. But I have some problems with it

^ Change in system() behaviour
13709 [dave pragpro] In 1.8,  system("badcmd")  returned false.
+ 13710 [ryand-ruby z] I think I've asked this before (or something VERY similar) and got the
| 13711 [ed.odanow wo] Here ist the short answer: "yes" (http://www.ruby-forum.com/topic/126588)
| 13712 [dave pragpro] Shame...  It makes a simple interface a lot uglier...
| + 13713 [ed.odanow wo] I have no idea why, because "false" is a good return value, if
| + 13714 [james graypr] if system(=85) rescue false
| + 13715 [ed.odanow wo] Raises SystemCallError if the command couldn't execute
|   13716 [dave pragpro] Wow That's strange. I have other code in the book that relies on =20
+ 13732 [matz ruby-la] * command execution failed (non zero exit status)
  + 13733 [drbrain segm] I would prefer the [ruby-core:13715] method of returning nil over an
  + 13736 [dave pragpro] I can see the reason for distinguishing the three, but I think that
  | 13985 [ara.t.howard] i know i'm in the minority in saying i support true/false/exception,
  + 13890 [lists bertra] A thing that didn't change as far as I see is the quite obscure decision