5820-6240 subjects 6284-9508

Re: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault
6009 [ralph.amissa] ...

Fwd: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault
6010 [ralph.amissa] ...
+ 6029 [decoux@mo lo] Where do came this ruby ?
| 6032 [ralph.amissa] ...
| 6033 [decoux@mo lo] You have not understood : I'm speaking about *ruby* not your package.
| + 6034 [ralph.amissa] ...
| + 6036 [ruby-core@wh] Thankyou, Guy.  This fixed mine.
|   6040 [ralph.amissa] ...
+ 6079 [decoux@mo lo] Someone can verified is this is still true
  6081 [decoux@mo lo] Well, I've forgotten to give the reason : with bison 2.0
  6082 [akr@m1 n. rg] Hmm.  alloca is not used by default since bison 2.0.
  6083 [decoux@mo lo] Weill it's seems this is this
  6110 [ralph.amissa] ...

[ ruby-Bugs-2523 ] $\ not escaped in IO#print documentation
6021 [noreply@ru y] Bugs item #2523, was opened at 2005-09-27 16:37

[ ruby-Bugs-2525 ] Undefined method error in net/http.rb
6031 [noreply@ru y] Bugs item #2525, was opened at 2005-09-27 08:38

make warning from 1.8.3
6038 [Daniel.Berge] ...
6042 [nobuyoshi.na] That target might have different meanings from GNU-make in

[ ruby-Bugs-2534 ] _'s not escaped in Object#send documentation
6053 [noreply@ru y] Bugs item #2534, was opened at 2005-09-29 00:36

[ ruby-Bugs-2535 ] YAML loading of quoted Symbols broken in 1.8.3
6057 [noreply@ru y] Bugs item #2535, was opened at 2005-09-28 11:50
6539 [aaron@sc ra ] Has anybody looked at this yet?  It doesn't look like this has been
6631 [aaron@sc ra ] I still haven't heard anything about this.  And it still doesn't seem to
6633 [ryand-ruby@z] SECONDED.
6637 [ruby-core@wh] I have a fix in my local Syck repo.  Please give me another day, my

Ads by gooooooooogle on Ruby Home Page...
6070 [halostatue@g] I'm not quite sure this is the right forum, but I just noticed a
6071 [benschumache] There is an option to turn off image ads. Maybe the administrator of
6072 [matz@ru y- a] Done.

Question about cgi.rb's read_multipart method and possible fix
6076 [rubyzbibd@ub] ...
+ 6129 [rubyzbibd@ub] Am I not being clear or is everyone too busy? :-)
+ 6130 [gianni@on 6i] I've had some big headaches when uploading files in a Rails app of mine.
| 6133 [rubyzbibd@ub] A quick answer is that I don't think it is the same problem, although
+ 6134 [rubyzbibd@ub] ...

irb segfault
6089 [dblack@wo bl] Reporting a 1.8.2-to-1.8.3 change in irb behavior, unearthed during an
6093 [decoux@mo lo] This is this in readline.c

Gems and repackaging, hopefully helpful
6095 [eivind@Fr eB] This has, as far as I can tell (checking mailing lists etc), been
+ 6096 [halostatue@g] And this is precisely why I've made the list of proposals that I've
+ 6103 [jim@we ri hh] I'm sorry, I didn't mean a tar file you just untar directly into the ruby
  + 6104 [hgs@dm .a .u] Excuse me jumping in on this point: I think there are a few things
  | 6106 [halostatue@g] No. I've already addressed why this won't be adequate, but briefly ...
  + 6105 [eivind@Fr eB] And this would make the tgz quite likely to be good, and the Gem to be possible
    6107 [dblack@wo bl] I respectfully suggest that the landscape of this discussion is
    6114 [eivind@Fr eB] I agree with half of this: It is a question of having gems abstract the layout
    6124 [dblack@wo bl] For some reason this revives in me the feeling that gems should just
    6125 [eivind@Fr eB] I've re-ordered for the important points first
    6126 [eivind@Fr eB] Oops, that was language that I felt the need to write in the moment and

Re: [sydney-devel] Threading Performance?
6108 [evanwebb@gm ] At present, no, there is not a patch against HEAD. I choose to do the

[PATCH] Fix a warning in Syck
6109 [tilman@co e-] /usr/lib/ruby/1.8/yaml.rb:133: warning: instance variable @input

Ruby 1.8.3: YAML.dump/load cannot handle Bignum
6115 [akira@ru y- ] <URL:http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=331050>
+ 6118 [laurent.sans] I did receive very similar reports from users of Alexandria
+ 6131 [ksruby@gm il] RCS file: /src/ruby/lib/yaml/rubytypes.rb,v
| 6132 [decoux@mo lo] moulon% ./ruby -ryaml -e 'puts YAML.load(YAML.dump(1234567890))'
| 6139 [ksruby@gm il] RCS file: /src/ruby/lib/yaml/rubytypes.rb,v
+ 6159 [akira@ru y- ] @@ -1142,6 +1142,9 @@

[OT] Re: gems is a language change, not a pkging system
6117 [sroberts@un ] David, Sean isn't comparing anyone in the discussion to Nazis.

Packaging BOF on Friday the 14th?
6119 [halostatue@g] (Crossposted to both ruby-core and rubygems-developers for the benefit
6127 [chad@ch df w] I agree that it would be nice to do, but as David pointed out, this

Re: [Rubygems-developers] Packaging BOF on Friday the 14th?
6120 [dblack@wo bl] We've already got a roundtable with Matz scheduled for that evening.
6122 [eivind@Fr eB] To make the work, I think it's very very important to make sure
6123 [dblack@wo bl] Please remember that Friday evening (up until about 9:00) is already

Re: Welcome to our (ruby-core ML)         You are added automatically
6121 [brb261@ya oo] ...

[ ruby-Bugs-2569 ] One-Click Ruby Installer 182-15 for Windows
6128 [noreply@ru y] Bugs item #2569, was opened at 2005-10-03 08:57

ObjectSpace.each_object, but not Symbols?
6135 [transfire@gm] {X=><Annotations a=1>}
+ 6136 [matz@ru y- a] ObjectSpace.each_object does not give you immediate objects such as
| 6145 [eustaquioran] For what he needs maybe the 1.8.3 Symbol.all_symbols
| 6146 [transfire@gm] Wow! I did not kown about #all_symbols. Then when I tired it... Oh my!
+ 6137 [nobu.nokada@] Symbols, Fixnums, true, false and nil are special objects,
  6138 [transfire@gm] Thanks.
  6150 [matju@ar en ] Why? I'd rather put those things in the class that the method belongs to.
  6153 [transfire@gm] Right. That works well. It's just that methods are generally referred
  6154 [matju@ar en ] Attaching methods to symbols isn't that bad, it's just not Ruby's way of

c++ Compileability and bugs in 1.9
6140 [christophe.p] ...
+ 6141 [matz@ru y- a] I'm pretty interested.  Let us see a patch.
+ 6142 [nobu.nokada@] I doubt that all of them need to be C++ compileable.

6143 [christophe.p] ...

Concerning c++
6144 [christophe.p] ...

Re: patch.tgz
6147 [nobu.nokada@] I don't think casting function pointers is safe.  This patch
+ 6148 [matju@ar en ] What makes you think so?
| 6152 [nobuyoshi.na] I'd have to say "portability safety (or strictness)."  Calling
| 6155 [matju@ar en ] However, the default cast is a reinterpret_cast at least on all pointers,
| 6156 [akr@m1 n. rg] I heard sizeof(long)!=sizeof(any*) on 64bit Windows.
| + 6157 [akr@m1 n. rg] Oops. LLP64.
| + 6158 [matju@ar en ] let's see... ruby.h, lines 90 to 100.
| | 6160 [michael.walt] This doesn't apply to C++ member function pointers, but you're
| | 6178 [matju@ar en ] Oh, right.
| + 6166 [halostatue@g] ...that's one thing I'd love to see happen.
+ 6149 [christophe.p] ...
+ 6190 [christophe.p] Will you be applying this patch (for safe function pointers) in 1.9 HEAD?
  6224 [matz@ru y- a] Will you please, Nobu?
  6225 [nobuyoshi.na] What about for 1.8?  I guess ANSI incompatible qsort() should
  6233 [matz@ru y- a] Yes.  Rename it ruby_sort() or whatever you like.
  6242 [nobuyoshi.na] Then, I'll just remove qsort() macro in util.h and use

NG link
6151 [christophe.p] ...

On NullClass or FalseClass#method_missing
6161 [transfire@gm] George Mischovitis' Nitro project uses OprnStruct-type containers
+ 6164 [dblack@wo bl] irb(main):001:0> def nil.method_missing(m); nil; end
| 6165 [transfire@gm] Right. That's wahy I thought
| 6168 [dblack@wo bl] Personally I'd rather not see singularities like false changed at all.
+ 6169 [nobu.nokada@] As far as I thik, NilClass is not important in Ruby, but rather
  6171 [transfire@gm] Yes, but isn't that a much more complex solution? Besides it can't but
  6179 [nobu.nokada@] I think it will require essential and big change to Ruby.

Concerning shared flag
6162 [christophe.p] ...
+ 6163 [decoux@mo lo] Before trying to modify ruby, why you don't try first to explain your
| 6167 [christophe.p] ...
| 6170 [decoux@mo lo] Sorry but I've not understood.
| 6172 [christophe.p] Given the fact that FL_SHARED is a general flag that exists on all objects,
| 6174 [decoux@mo lo] Can you give all the modifications that you want to do to ruby, to make
+ 6175 [eric_mahurin] Not sure if you are aware or not, there are some serious
  6176 [christophe.p] I have taken a different approach that works without changing the ruby
  6177 [eric_mahurin] In my patch, you'll notice that there is no "capa" field.

Re: another array patch - performance boosts all over the place
6173 [eric_mahurin] Has anybody evaluated this patch?  In addition to these cases
+ 6180 [matt.mower@g] Sadly I'm not qualified to evaluate your patch but it seems like
| 6181 [eric_mahurin] Thanks.  For now, it sounds like Evan Webb will integrate this
+ 6186 [nohmad@gm il] I've checked your patch, though I couldn't understand meaning of your

[ ruby-Bugs-2594 ] [PATCH] memset not clearing memory due to reversed arguments
6182 [noreply@ru y] Bugs item #2594, was opened at 2005-10-06 12:34
6185 [matz@ru y- a] This code happen to be removed yesterday.

[ ruby-Patches-2595 ] Cookies from Nokia devices may not be parsed correctly
6183 [noreply@ru y] Patches item #2595, was opened at 2005-10-06 23:08
6184 [matz@ru y- a] Accepted.

[ ruby-Bugs-2597 ] Ruby mail improvement
6187 [noreply@ru y] Bugs item #2597, was opened at 2005-10-07 12:32

yield and call not identical?
6188 [dblack@wo bl] I'd always thought that yield was just syntactic sugar for calling the
+ 6189 [hgs@dm .a .u] 356-360 in the Pickaxe II.  Particularly the bottom of 358.
+ 6191 [daniel.amela] I wonder how many people *do* fully understand the whole
| + 6193 [halostatue@g] No, I don't think it's just you. I don't think that your assessment of
| | 6194 [transfire@gm] class Function
| + 6195 [ruby@ja es r] But I might be.
+ 6192 [vjoel@pa h. ] b is a Proc, not a block, so it can be stored somewhere, and called

WIlderness: What is the purpose of T_ICLASS -
6196 [ruby-core@ha] While tracing the Objects generated by Object.C

WIlderness: Followup on T_ICLASS -
6197 [ruby-core@ha] While tracing the Objects generated by Object.C
6198 [nobu.nokada@] No.  Consider the case a class including multiple modules.

Kernel rdoc HTML file not being created when rdoc is run on 1.8.3
6199 [ruby@ja es r] When 1.8.3 came out, I grabbed the source and ran rdoc on it.  After
6206 [dougkearns@g] Yes. See [ruby-talk: 157245]
6211 [ruby@ja es r] Ah.  Thanks.
6318 [ocean@m2 cc ] I did binary search for commit time.

[ ruby-Bugs-2606 ] st.c calls ruby_xcalloc but frees with free()
6205 [noreply@ru y] Bugs item #2606, was opened at 2005-10-09 22:13

extend and super -- I cannot understand why this behavior
6213 [transfire@gm] module Q
6214 [matju@ar en ] class<<R;self;end is a subclass of class<<Q::H;self;end, right?
6215 [transfire@gm] Ah, I see. So it's the same @x in either method --I thinking super
6219 [matju@ar en ] What do you mean by hard-bound at compile time? In a super-call, the self
6220 [transfire@gm] Yes, in modules it's not so, but for classes it is.
6232 [halostatue@g] Why? Instance variables are in the class, not the module. So, when
6234 [transfire@gm] No good reason. Just got it twisted a bit in my mind when I was

function pointer usage
6226 [nobuyoshi.na] * array.c, enum.c, eval.c, util.c: safer function pointer usage.
+ 6227 [decoux@mo lo] Strange :-)
+ 6228 [nobuyoshi.na] Oops.

Keyword arguments in Rite
6235 [daniel.schie] I'm saying is absolute nonsense.
6236 [ruby-ml@ma i] document = make_document from: 'foo.txt', using: LaTeX
+ 6237 [daniel.schie] True, but that's also the only case I can think of :)
| 6238 [mailing-list] [a: :b is ugly]
+ 6243 [daniel.schie] Just one more question. I know you'll be able to write
  6244 [matz@ru y- a] Yes.
  + 6245 [skaes@gm .n ] I'm not quite up to date in this discussion, but, if
  | 6301 [daniel.schie] Exactly :)
  + 6246 [nobu.nokada@] Daniel writes about method definition, not calling.  So my
    6248 [matz@ru y- a] Oops, you're right.

Dir.chdir annoyance
6241 [gsinclair@gm] Anyone seen this warning?

[ ruby-Bugs-2613 ] building ruby 1.8.3 on Solaris
6247 [noreply@ru y] Bugs item #2613, was opened at 12/10/2005 15:56
6249 [nobu.nokada@] Not syscall.h?  Currently, syscall.h is checked in
6250 [nobu.nokada@] +union YYSTYPE;

RubyGems, upstream releases and idempotence of packaging
6251 [ mfp@ac .o g] [sorry for the very late reply; I left this message in +postponed and forgot
+ 6252 [hgs@dm .a .u] ...
| + 6255 [jim@we ri hh] I think there is some confusion on whether a .gem file is a distribution
| | 6256 [hgs@dm .a .u] However, if we apply what Donald Norman would call a Forcing
| + 6268 [gsinclair@gm] ".gem files should always be considered 'binary releases' in that
| | + 6270 [hgs@dm .a .u] Basically against making life more difficult for repackagers,
| | | 6271 [gsinclair@gm] I can't argue with the second one :)  The first depends on how well
| | + 6303 [ mfp@ac .o g] We're considering the following alternatives (the initial assumption is
| |   6307 [gsinclair@gm] Yes, valid assumption, but my point is about "repackaging" gems vs
| |   6310 [hgs@dm .a .u] ...
| + 6269 [hgs@dm .a .u] ...
| + 6273 [halostatue@g] You know, I *choose* (at this point) to not include the Rakefile or
|   6274 [hgs@dm .a .u] ...
+ 6253 [jim@we ri hh] Some very good nuggets of information is here.  It seems to me that phase

[ ruby-Support Requests-2615 ] Could not create openssl module during make
6254 [noreply@ru y] Support Requests item #2615, was opened at 2005-10-12 13:46

Re: ruby 1.8.3 (2005-09-21) [i486-linux] sisu segfault
6261 [ralph.amissa] ...
6263 [decoux@mo lo] What is this ?
+ 6264 [decoux@mo lo] Well it's u3.cnt, sorry
+ 6265 [nobuyoshi.na] If the size is indivisible by sizeof(YYSTYPE), it shouldn't be
  + 6266 [decoux@mo lo] Well,
  + 6267 [decoux@mo lo] Another way to see it
    6277 [ralph.amissa] ...
    6281 [decoux@mo lo] Hey, this is not me that you must thanks. Personnally I do nothing, I'm

Re: A concrete solution to RubyGems' repackageability problems
6262 [gsinclair@gm] Didn't you say the repackaging problems concerning RubyGems were about
6304 [ mfp@ac .o g] The most annoying problems are those that require patches to the source code.
+ 6306 [khaines@en g] Just a sidenote, but I packaged the recent Crypt::ISAAC release with Package
+ 6308 [gsinclair@gm] That sounds like a very positive step.  I like to see bits of RubyGems

require features on $" not absolute
6275 [transfire@gm] *  Ruby tries to load the library named _string_, returning
+ 6276 [transfire@gm] s/this is/this isn't/
+ 6278 [nobu.nokada@] Try it with 1.9.
  6279 [transfire@gm] Cool deal. I will whirl. Thanks.

Wilderness: Converting for Object Address ot Object ID
6280 [ruby-core@ha] I have written several functions to dump Object Trees.

Wilderness: Need Code to invoke ELTS_SHARED response
6282 [ruby-core@ha] Testing the My Object Dump and I am trying to cause creation
6283 [ mfp@ac .o g] v1 = "foo".clear
6288 [ruby-core@ha] Runinng 1.8.2 Produces
+ 6289 [eric_mahurin] I haven't looked at string.c, but if it is like array.c, you
+ 6290 [ mfp@ac .o g] $ ~/ruby1.9/bin/ruby -ve "p 'foo'.clear"
| 6291 [ruby-core@ha] OK - Thanks That did it!
+ 6300 [cmills@fr es] a = "ab"[1,1]
  6305 [ mfp@ac .o g] Not all substring #[] operations cause the underlying string to be shared (the
  6313 [ruby-core@ha] Actually, it does!
  6314 [decoux@mo lo] Look at pack.c (call to rb_str_associate())
  + 6316 [ruby-core@ha] Could you provide a Code Fragment to cause this feature to
  + 6317 [ruby-core@ha] I have looked a the code in pack.c --- This is the problem I am having
    6319 [decoux@mo lo] I can :-)