13904-14484

13718-15017 subjects 14126-15173

^ step parameter for #succ
13904 [transfire gm] I would like to suggest that #succ (and #pred where applicable) be

^ What's the status of compiler/compiling on windows?
13908 [ggarra advan] I'm needing to embed ruby into some applications on windows.  As such, I
13913 [nobu ruby-la] Have you given a glance at win32/REAME.win32 file?

^ Re: [Spam] Re: What's the status of compiler/compiling on windows?
13914 [ggarra advan] Yes.  That just tells me how to get a bare-bones ruby install.
+ 13915 [nobu ruby-la] It's another story.
| 13921 [Daniel.Berge] There are still two problematic libraries in the stdlib - OpenSSL and =
+ 13926 [luislavena g] T24gRGVjIDcsIDIwMDcgODoyMSBBTSwgR29uemFsbyBHYXJyYW11w7FvIDxnZ2FycmFAYWR2YW5j
  + 13941 [ggarra advan] Absolutely.  That's exactly the info I was looking at.  Someone should
  | 13945 [nobu ruby-la] Is "out of source" to build in different directories from the source direct=
  + 14038 [joe.swatosh ] I apologize for the late reply/threadjack, but I am curious about the
    14039 [luislavena g] 1) memory allocated by one runtime cannot be free'ed by another, that
    14040 [halostatue g] And the *reason* it's not a theoretical scenario is that Unix doesn't
    + 14046 [nobu ruby-la] 2) thread local variables like errno cannot be shared across
    | 14061 [joe.swatosh ] Gack.  Is this common?
    | 14064 [nobu ruby-la] errno is quite common, I believe.
    | 14110 [joe.swatosh ] Thanks Austin, Luis, and Nobu for taking me through it even though it
    | 14115 [halostatue g] No problem. I should make a blog post on this so that there's a
    + 14060 [joe.swatosh ] If everyone used ruby_xmalloc and ruby_xfree, would this part of the binary
      + 14062 [luislavena g] If memory plays well with my age, I think perl guys created a
      | 14073 [halostatue g] Yikes. That doesn't seem pretty.
      + 14072 [halostatue g] No, unfortunately. It would help when you're embedding Ruby in

^ Changes in C API of Ruby 1.9
13916 [gethemant gm] Trying to port couple of my extensions to Ruby1.9 and hitting on some
+ 13917 [matz ruby-la] use RSTRING_LEN() which is defined both in 1.8 and 1.9.
+ 13920 [Daniel.Berge] Ack. Please give me .length as a shortcut for as.heap.len.
  13951 [nobu ruby-la] To define length as as.heap.len?  You never use variables named

^ A wafer thin (doc) bug in Thread#priority=
13922 [Daniel.Berge] Ruby 1.8.6 and 1.9
13924 [brent mbari.] I vote that the documentation is broken
13942 [djberg96 gma] "thr.priority= integer   => integer

^ Default encoding not set properly
13928 [hramrach cen] SGVsbG8sIEkgdHJpZWQgcnVubmluZyAxLjkgaXJiLCBhbmQgZmlndXJlIG91dCB3aGVyZSB0aGUg
+ 13936 [matz ruby-la] It's not implemented it yet.
+ 13939 [duerst it.ao] First, can you please fix your mail software so that it sends
  13946 [hramrach cen] There was no line ending thus it did not send any, sorry.

^ Again - class variables and inheritance
13934 [ed.odanow wo] Good morning!
13935 [matz ruby-la] It's changed again.  We abandoned some incompatible changes that may

^ IRHG -- GC Collection
13948 [ceo hawthorn] Sorry to be away so long -- But had to make money for a while!
+ 13949 [matz ruby-la] Yes, if you mean Ruby objects and memory structures managed by them.
+ 13950 [brent mbari.] I don't believe anything would prevent you from wrapping any

^ Re: [Spam] Re: Changes in C API of Ruby 1.9
13952 [ggarra advan] I'll admit I have not looked at the 1.9 source in some time.
+ 13953 [nobu ruby-la] It's not related to alignment, but for embedding the contents in RSTRING.
+ 13954 [matz ruby-la] Instead, they consume much less memory, resulting less garbage

^ [Mac OS X 10.4.11] ruby1.9-trunk make failed
13955 [yvon.thorava] on Mac OS X 10.4.11 the ruby1.9-trunk (downloaded yesterday) make failed
13956 [laurent.sans] Building trunk requires Bison 2.3. It seems that you run an old
13957 [yvon.thorava] OK, fine, thanks for this quick answer !!!

^ CVE-2007-5770
13958 [rbrown gento] suggested that r13656 [2] is a patch to fix it.
13959 [shyouhei rub] Yes, it does affect 1.8 series.  Revision 13567 is the fix for 1.8
13960 [rbrown gento] Oh, sorry I missed that revision, I think it was because I was looking

^ Error while bulding Ruby 1.9 from snapshot 2007/12/09
13961 [ed.odanow wo] This is a multi-part message in MIME format.
+ 13962 [ed.odanow wo] Sorry - I've forgotten environment information. I tried to build
+ 13965 [matz ruby-la] You don't have <pwd.h> on your machine.  You have to install proper
+ 13966 [nobu ruby-la] xt/pty=20
| + 13968 [kimura.koich] I have tried to build the snapshot  with VC++ 7.1 and have same problem.
| + 13973 [ed.odanow wo] This is a multi-part message in MIME format.
|   13976 [nobu ruby-la] First, ext/pty/extconf.rb should skip compiling on Windows.
+ 13987 [ed.odanow wo] Using the snapshot from 2007/12/10 everything works well. There
  13988 [ed.odanow wo] This is a multi-part message in MIME format.

^ redefineable not operator
13969 [david davidf] Matz,
+ 13970 [matz ruby-la] We wanted to allow some libraries that build some kind of syntax tree
+ 13971 [murphy rubyc] why such heavy syntax changes, two weaks before the release of Ruby
  + 13972 [matz ruby-la] As I said before we try our best to stabilize features before
  | + 13977 [transfire gm] Are you feeling okay, Matz?
  | | 13978 [matz ruby-la] Fine, thank you.
  | | 13984 [transfire gm] Very good news then :-)
  | + 14007 [murphy rubyc] sorry, I don't get this example.
  |   14011 [matz ruby-la] I wish I could express more fluently.  Why don't you speak Japanese? ;-)
  |   + 14013 [murphy rubyc] This is a multi-part message in MIME format.
  |   | + 14016 [david davidf] I believe that Matz has only just allowed ! to be redefined.  It has
  |   | | 14019 [matz ruby-la] !@ and ~@ are aliases to ! and ~ respectively.  Since they are
  |   | | 14024 [gwtmp01 mac.] => {:"!"=>2}
  |   | | + 14025 [billk cts.co] Whoa.  I just tried it myself.  I must say I'm inclined to
  |   | | | 14026 [matz ruby-la] Try :"~@" on 1.8.
  |   | | | 14028 [billk cts.co] Thanks, matz.
  |   | | + 14027 [matz ruby-la] This case, it's more than having two names.  Those two should act
  |   | | | + 14030 [gwtmp01 mac.] In looking through the trunk code I see that the string->symbol
  |   | | | + 14063 [murphy rubyc] - why does Symbol.all_symbols return :"!" twice?
  |   | | + 14029 [dave pragpro] It's an interesting example of a leaky abstraction. On the one hand,
  |   | |   14031 [gwtmp01 mac.] This actually isn't a new situation.  Ruby 1.8 has
  |   | |   14032 [dave pragpro] OK, so the move to extend this behavior is wrong.. :)
  |   | |   + 14033 [matz ruby-la] OK, what do you think the best idea to "fix" this?  It's more
  |   | |   | 14037 [gwtmp01 mac.] I'm trying to understand the string interning.  It seems like
  |   | |   | + 14041 [matthias wae] Python 3k already specifies UTF-8 as the default source file
  |   | |   | | + 14044 [halostatue g] I've just been too busy to write a blog post about the problem of
  |   | |   | | + 14045 [luislavena g] T24gRGVjIDEyLCAyMDA3IDc6NDEgUE0sIE1hdHRoaWFzIFfDpGNodGVyIDxtYXR0aGlhc0B3YWVj
  |   | |   | + 14051 [duerst it.ao] Here is my general understanding from discussions with Matz
  |   | |   |   14053 [gwtmp01 mac.] I still haven't digested all the symbol/encoding interactions
  |   | |   + 14057 [charles.nutt] I would go so far as to say the behavior in any form is wrong. I believe
  |   | |     14066 [duerst it.ao] Yes. It looks to me as if the number of ruby-specific cases are
  |   | + 14020 [matz ruby-la] We have almost finished the feature design, then we need to fix the
  |   + 14014 [ed.odanow wo] Wasn't this technique described by Jim Weirich on RubyConf 2007 (see
  + 14018 [murphy rubyc] class Fixnum
    14035 [ed.odanow wo] Well, I want to do some written up thinking and dreaming in

^ Compliment to rb_thread_blocking_region() ?
13989 [tony clickca] The rb_thread_blocking_region() appears to allow releasing the GIL and
13992 [mental rydia] There is no public API to re-enter the interpreter at this time.  Besides
+ 13993 [mental rydia] And, of course I mean event_base_loop()/event_base_dispatch(). :)
+ 13994 [tony clickca] How would it be different from what rb_funcall() does already?
  13995 [mental rydia] It's kind of the same problem.  It isn't safe to call rb_funcall() directly
  13996 [tony clickca] Okay, that should work fine, and neatly sidesteps any weird messing around

^ aliasing methods that don't exist
13990 [drbrain segm] class X; alias_method :a, :a; end
13991 [Daniel.Berge] Hehe, I'd say so. I'm pretty sure it's this line in eval.c/rb_alias

^ Output of "ruby --help" looks not correct for Ruby 1.9
13997 [ed.odanow wo] Good morning!
14006 [jeremie.pier] MjAwNy8xMi8xMSwgV29sZmdhbmcgTsOhZGFzaS1Eb25uZXIgPGVkLm9kYW5vd0B3b25hZG8uZGU+
14008 [ed.odanow wo] I don't know what parts of the --help output are affected, but I think

^ Array#pack and String#unpack documentation
13998 [radar2002 gm] In working with Array#pack and String#unpack today, I found
14003 [matz ruby-la] OK, I will merge you patch to the trunk.  Thank you.
14005 [gwtmp01 mac.] You're welcome.
14009 [matz ruby-la] pack/unpack does not handle M17N at all.
+ 14012 [gwtmp01 mac.] Good to know.
| 14022 [matz ruby-la] Yes.
| 14023 [gwtmp01 mac.] Thank you for all these clarifications.  Other than the source are
+ 14015 [ed.odanow wo] Really? - What's about directive "U" for "UTF-8"?
  14021 [matz ruby-la] It works on UTF-8 byte sequences, but it does not change the encoding

^ [1.9] Is this right behavior?
14000 [ggarra advan] #!/usr/bin/env ruby19
+ 14001 [matz ruby-la] Expanded hash is allowed only at the end of arguments.
+ 14002 [dblack rubyp] This message is in MIME format.  The first part should be readable text,

^ Return value of Kernel#p
14010 [murphy rubyc] # Ruby 1.8
14036 [mame tsg.ne.] Exactly.  I suggested the behavior for the purpose at [ruby-dev:29736].

^ Re: Welcome to our (ruby-core ML) You are added automatically
14017 [davidxy123 g] help

^ Encoding and test::unit
14034 [duerst it.ao] [I'm posting this to ruby-core rather than ruby-dev,

^ Fix e2mmap.rb for 1.9
14042 [drbrain segm] E2MM.Raise complains about $! being read-only now, and E2MM is used by
+ 14043 [dave pragpro] FWIW, the only reason RDoc used it is because I ripped off all that
| 14047 [djberg96 gma] Speaking of libraries that should be removed from the standard library...
+ 14078 [wilsonb gmai] At RubyConf, matz indicated that e2mmap would be removed from the next release.

^ RDoc + irb (Was: Fix e2mmap.rb for 1.9)
14049 [drbrain segm] Yeah, with Ryan's ruby_parser project, and Koichi's agreement to
14052 [dave pragpro] So you'll be making RDoc MRI-specific?
+ 14055 [drbrain segm] No, ruby_parser could be imported to deal with parsing instead of irb
+ 14056 [charles.nutt] Good question.
  + 14058 [dave pragpro] Actually, I wasn't talking about removing the irb dependency--there
  | + 14059 [charles.nutt] Ahh, ok...I understand what you mean now.
  | + 14067 [duerst it.ao] While we are discussing RDoc, I'd like to take the chance and bring
  |   + 14068 [vjoel path.b] But you can link to the page displayed in the main frame. Right-click on
  |   | 14071 [duerst it.ao] Of course there are hacks around the problem, but that's not
  |   + 14074 [dave pragpro] I agree 100%. The intent with RDoc was to produce something quickly
  |     14075 [hgs dmu.ac.u] When I've looked into rdoc, hoping to be able to modify it to
  |     14080 [ryand-ruby z] agreed Hugh... which is why I'm planning on ripping it apart and
  |     + 14082 [dave pragpro] Excellent news!
  |     + 14096 [hgs dmu.ac.u] Specifically, I was trying to get the case where code is expanded
  |       14104 [ryand-ruby z] you should look at evan weaver's rdoc templates. they're frameless. I
  + 14065 [ryand-ruby z] AST structure may change over time, but a comment node in front of a
    14069 [charles.nutt] We've thought of hacking rdoc in JRuby to use our parser as well, so I
    14070 [drbrain segm] a) This will be tested, so it will be easy for you to make any fixes

^ [OT] Re: redefineable not operator
14050 [duerst it.ao] Could you be a bit more specific? XML has extremely clear rules
14054 [halostatue g] .NET can generate XML that it can't deserialize, because of UTF-8. Not

^ Ruby Core Performance Patch: Fixnum#gcd
14076 [ks kurtsteph] This is a multi-part message in MIME format.
14081 [ryand-ruby z] We sped up rubygems in a similar way by not using date. We used time

^ Short Question about Class Continuation in Ruby 1.9.n
14077 [ed.odanow wo] Good evening :-)

^ warning: variable $KCODE is no longer effective
14079 [drbrain segm] (I accidentally sent this to ruby-talk.)
14101 [matz ruby-la] Hmm, setting it to 'iso-8859-1' is OK, I guess.  If one want to change

^ multibyte characters in curses
14083 [hramrach cen] Perhaps it would be better to send this to core: This is a
14084 [hramrach cen] Sorry, forgot the patch
14122 [hramrach cen] The same patch applies to the curses extension in 1.9, and the patched
14129 [duerst it.ao] Ruby definitely will start to care about locales, but calling
14135 [hramrach cen] However, some libraries like ncurses rely on the locale to determine

^ [PATCH] Continuations on dead threads crash ruby during GC
14085 [brent mbari.] This patch is against Ruby 1.8.6-p110

^ attr_accessor for question methods (?)
14086 [transfire gm] Like to make a small suggestion for 1.9. It would be great if we could
+ 14087 [djberg96 gma] I proposed this a while back and it got shot down for what were, IMO,
+ 14090 [charles.nutt] Perhaps an additional method would be the way to do it?
  + 14091 [alex blackke] How about a single method that handles all possible accessor methods,
  | 14100 [transfire gm] It is an interesting thought. But really this is going well beyond the
  + 14094 [ggarra advan] +1 for this.  This creates a very clean and consistent API.  The other
  + 14098 [djberg96 gma] -1
  + 14099 [transfire gm] On Dec 14, 4:18 am, Charles Oliver Nutter <charles.nut...@sun.com>

^ Latest 1.8 trunk compile problem
14088 [ceo hawthorn] I just checked out the latest 1.8 trunk and
14089 [nobu ruby-la] Fixed a couple hours ago.

^ IRHG - 'lev' Meaning and relationship to 'mark_stack'
14092 [ceo hawthorn] define GC_LEVEL_MAX 250

^ IRHG - PUSH_FRAME?
14093 [ceo hawthorn] The following define statement is driving me crazy
+ 14095 [ggarra advan] You haven't been coding in C too long.  While splitting the do/while
+ 14097 [dooby d10.ka] C90 compilers won't allow a declaration (struct) amid code,

^ Small (hopefully) problem with RDoc (Ruby 1.9)
14102 [ed.odanow wo] Good morning!
14103 [ed.odanow wo] I made some tests, unfortunately the debugger crashed, so I
14105 [ed.odanow wo] I found a patch that solves the problem. Unfortunately I don't
14106 [ed.odanow wo] The problem occurs for Ruby 1.8 (I have oci-186-25) too.
14107 [ed.odanow wo] Unfortunately this is not enough for Ruby 1.8. RDoc1.8 crashes

^ RDoc (Ruby 1.8) and RDoc (Ruby 1.9) generate different number of files for same directory
14108 [ed.odanow wo] I build Ruby 1.9 from yesterday night snapshot in a subdirectory
+ 14116 [ed.odanow wo] This is a multi-part message in MIME format.
+ 14124 [ed.odanow wo] <partly OT>

^ What purpose does struct FRAME's tmp pointer serve?
14109 [brent mbari.] The struct FRAME looks in Ruby 1.8.6 like this:  (excerpted from env.h)
14112 [brent mbari.] Never mind.  I dug a little deeper and answered my own question.
14120 [Daniel.Berge] What was the answer?

^ What does volatile assignment of *pos in rb_thread_s_new() do?
14111 [brent mbari.] In the rb_thread_s_new() of eval.c, a volatile VALUE *pos is assigned

^ Re: [Spam] What does volatile assignment of *pos in rb_thread_s_new() do?
14113 [ggarra advan] It keeps the value in the stack, preventing it from being GC'ed during
+ 14114 [brent mbari.] Gonzalo,
+ 14473 [pbrannan atd] Wouldn't the same thing be accomplished by marking the member in

^ Request for inclusion: better platform handling?
14117 [luislavena g] This issue was discussed too many times also, same like Windows status for Ruby.

^ Enumerable#uniq, anybody?
14118 [duerst it.ao] I have just tried to use zip followed by uniq in a 1.9 context.
+ 14119 [murphy rubyc] the current implementation of uniq depends on the receiver being
| 14132 [duerst it.ao] I don't understand your argument. to_a means "convert to an Array",
| 14208 [murphy rubyc] I was under the impression that all Enumerator methods only traverse the
| 14255 [duerst it.ao] I think I know what you mean, but I think the right way to express this
+ 14121 [nobu ruby-la] HTH.

^ Some Ruby 1.9 loose ends to tie up
14123 [david davidf] Matz,
+ 14125 [mental rydia] We've implemented it in JRuby.
+ 14127 [jeremy bitsw] This is the only remaining incompatibility with Action View (part of
+ 14131 [charles.nutt] I sure hope not. In my opinion Proc.binding and friends break
| 14165 [jeremy bitsw] Just added!
| 14166 [charles.nutt] Well, that's unfortunate. :(
+ 14220 [matz ruby-la] Just done.
  14238 [charles.nutt] I would like to hear from Koichi how this affects various optimizations
  + 14241 [david davidf] __method__ returns the name of the current method as defined.
  | 14243 [ko1 atdot.ne] When do we need it?  Now callee information is thrown on method
  | 14244 [matz ruby-la] OK, __callee__ was implemented so just because it was easy on MRI.
  | 14300 [gwtmp01 mac.] There is a use case that would benefit from __callee__.  In
  + 14242 [ko1 atdot.ne] The reason why I commit Proc#binding is that some of applications
    14245 [charles.nutt] The only thing it gains you is that you don't have to explicitly pass in
    + 14246 [rocky.bernst] On Dec 21, 2007 8:21 PM, Charles Oliver Nutter <charles.nutter@sun.com>
    | 14250 [charles.nutt] Currently we do not optimize away call frame information in most cases,
    + 14470 [pbrannan atd] Charlie and I were talking about this at Rubyconf.  It seems silly to me
      14477 [charles.nutt] I don't know about macros, but the idea that passing a block opens up
      14483 [pbrannan atd] Or, if it were truly a keyword and not a method, it wouldn't be possible
      14484 [charles.nutt] Quite...or at the very least, it wouldn't invoke "binding" as we know it
threads.html
top