13904-14484 subjects 14303-14784

^ Curious warning with IO.popen on Windows
14126 [Daniel.Berge] Ruby 1.8.6 (one click)
14128 [nobu ruby-la] Why curious?
14136 [Daniel.Berge] But I haven't closed it until after the command has finished, have I?
14137 [billk cts.co] popen forks the process.  The command runs asynchronously in a separate

^ Another last-minute Ruby 1.9 question
14130 [david davidf] Here's another last-minute question, probably for Matz or Nobu...
14133 [duerst it.ao] We have had some discussions on the Japanese list. In terms of
14138 [david davidf] Thanks for the clarifications, Martin!
14156 [duerst it.ao] With "always", do you mean even if the encoding is fully supported?

^ IRHG - Trying to understand gc_mark()
14134 [ceo hawthorn] The first thing I need to know is what is 'lev' used for?

^ String#% bug on Windows?
14139 [Daniel.Berge] # test.rb
+ 14141 [gwtmp01 mac.] What do you get for "0_3_7_7".to_i on Windows?
| 14142 [Daniel.Berge] =20
+ 14148 [nobu ruby-la] What are the versions?  Kernel#Integer works more strictly
  + 14151 [djberg96 gma] Ah, you know what, it was 1.8.5 on the Linux box I tested on. When I
  + 14152 [gwtmp01 mac.] It doesn't seem unexpected to me.  I would expect the following
    14155 [nobu ruby-la] Yes, "unexpected" is an ArgumentError.
    14159 [gwtmp01 mac.] I misunderstood your comment.  Thanks for clarifying.

^ rest argument able to duplicate/replace local argument bug?
14140 [Thomas.Enebo] This is a bug right?
+ 14144 [gwtmp01 mac.] I get a duplicate argument error on 1.8.6 and 1.9.
| 14145 [Thomas.Enebo] ruby 1.8.6 (2007-03-13 patchlevel 0) [i686-darwin8.10.1]
| 14146 [gwtmp01 mac.] $ ruby -v
+ 14153 [nobu ruby-la] Seems a bug in 1.8.  Thank you.

^ Last-minute encoding proposal
14143 [david davidf] transcode.c vs. encodings supported by encoding.c, I suggested that if
14157 [duerst it.ao] Many thanks for some more great thoughts.
14160 [david davidf] My original version of this proposal involved three methods: encode, decode,

^ named captures assigning to local variables
14147 [david davidf] I've just been browsing ruby-dev. For an english speaker, it is kind of
14150 [akr fsij.org] See the additional document of Regexp#=~.
14158 [david davidf] Thank you for the clarification, akr.  I'm embarassed to say that it didn't
14161 [david davidf] If I may, have a proposal.  My apologies if this has already been
+ 14162 [vjoel path.b] Very nice. Since = is not an operator (and not a method), a certain
| + 14163 [akr fsij.org] =~ has some magic already.  It is not redefinable for regexp
| | + 14167 [dblack rubyp] Of course that could mean it's OK to have more magic, or it could mean
| | + 14169 [mame tsg.ne.] Sorry slightly off topic, but I wonder if it's ok to be able to
| + 14168 [mame tsg.ne.] I prefer it too.  Matz, what do you think?
+ 14170 [akr fsij.org] I think it is possible idea.
  + 14171 [dave pragpro] I think that's a good point, and I'd support the consistency of
  + 14172 [david davidf] How about making the return value an array of the captured strings, or nil
  | + 14173 [dave pragpro] I rather like the consistency we now have, that the value of an
  | + 14181 [akr fsij.org] Another idea is $~, which is a MatchData or nil.
  |   14183 [david davidf] That's simpler to implement than what I had done.  And in the simple
  |   14184 [david davidf] This is a multi-part message in MIME format.
  |   14201 [dblack rubyp] I'm still wondering about making such a big exception to the usual
  |   14222 [david davidf] Let me make clear that I proposed this assignment syntax as an
  |   + 14239 [charles.nutt] Here here! This seems like a bad time to be trying to jam a few features
  |   + 14252 [akr fsij.org] Not strong objection.
  |     14256 [gwtmp01 mac.] It isn't too hard to stuff matches into local vars
  |     + 14258 [charles.nutt] I agree here...the magic assignment doesn't seem to buy much over just
  |     + 14290 [akr fsij.org] Unfortunately it is not usable in condition.
  |       14293 [akr fsij.org] Oops.  one more line, not two.
  + 14174 [matz ruby-la] I think we can consider it as a kind of multiple assignment, which
    14180 [david davidf] This is a multi-part message in MIME format.

^ Experimental PATCH to improve thread performance
14149 [brent mbari.] This is a multi-part message in MIME format.
+ 14187 [brent mbari.] Today, I discovered that the patch I originally submitted on this mail thread
| 14194 [sylvain.joye] That is very interesting. I'll try that in January on a very
+ 14202 [charles.nutt] I'd love to know what broke in JRuby. I tried both benchmarks and
| 14257 [brent mbari.] I'd like to work with you on this after the holidays.
| + 14266 [charles.nutt] In JRuby, all threads executing Ruby code checkpoint at the same places
| | + 14268 [mumismo gmai] I don't think Matz has thought seriously about threads. I hope soon after
| | + 14274 [sylvain.joye] Is #raise unsafe if done at checkpoints ? (i.e. raising a thread which
| | | 14276 [charles.nutt] I'm not sure if you mean checkpoints in your code or checkpoints in e.g.
| | | 14450 [sylvain.joye] multi-threaded applications. *Of course* you don't want to raise
| | | 14463 [charles.nutt] Except that requiring users to use synchronization points in order to
| | | + 14466 [sylvain.joye] What I meant was in the same line of though than Brent's: only act upon
| | | | 14472 [charles.nutt] Certainly, the feature could be made safe. And in JRuby's internal
| | | | 14474 [pbrannan atd] Broken and dangerous is ok by itself, but Thread#raise is also
| | | | 14478 [charles.nutt] Heheh....well put :)
| | | + 14476 [pbrannan atd] I just had a thought.  What if Thread#raise could only raise exceptions
| | |   14480 [charles.nutt] Actually this is a pretty good thought. Already these events can only
| | + 14278 [gwtmp01 mac.] For those of you interested in general concurrency issues, you
| | + 14301 [brent mbari.] Charlie,
| |   14458 [charles.nutt] You are correct, I confused myself here. I tend to mix up kill and stop.
| |   14490 [brent mbari.] Charlie,
| |   14504 [charles.nutt] This isn't really the hole I meant, but I'll endulge you for the moment :)
| |   14570 [brent mbari.] I've spent some time reading your post in detail and
| |   14574 [charles.nutt] This still raises "unknown" or "unexpected" exceptions in the target
| |   14619 [brent mbari.] Charlie,
| |   14625 [charles.nutt] Ok, the key point here I must have missed is that threads would be
| |   14679 [brent mbari.] Charlie,
| |   14681 [charles.nutt] Maybe as an extension...are you looking to implement it in C? How will
| |   14683 [brent mbari.] I intend to implement this as another patch to 1.6.8 and
| |   14687 [charles.nutt] Yes, I don't expect it would be too difficult. The threading constructs
| + 14267 [charles.nutt] I'll have a look at this, and I'll hope to hear from you after the
| + 14269 [charles.nutt] The bug here is that for some reason "self" is null going into a call.
| + 14270 [charles.nutt] Oh, one other feature of JRuby might be interesting here, and
|   14272 [matz ruby-la] Koichi told me about the idea about thread pooling in YARV few weeks
|   14273 [charles.nutt] I'd be happy to discuss all details of JRuby optimizations with Koichi
+ 14311 [nobu ruby-la] 1.9 has the similar feature to it, rewinding the machine stack
  14316 [brent mbari.] Thanks for at least looking at this little hack.
  14808 [pbrannan atd] I've tried getting rid of the stack copying, but it's definitely
  14811 [brent mbari.] Paul,
  14823 [pbrannan atd] It has to do with keeping pointers/references on the heap to objects on
  14827 [nobu ruby-la] And those pointers are referred asynchronously?  Such objects
  14834 [pbrannan atd] Yes, they are referenced asynchronously.  At least, I think that's the

^ "*a = 1" has different results in 1.8 and 1.9
14154 [ed.odanow wo] Good morning!

^ Input related to signal usage in Ruby 1.9 implementation
14164 [chirag80bece] We have study the code of yarv (Ruby 1.9 implementation) and as per our
15173 [chirag80bece] We had given some inputs for the signal usage in Ruby 1.9 based on our study

^ [feature] xmlrpc/client ssl client auth
14175 [jerith jerit] This is a cryptographically signed message in MIME format.
14177 [james graypr] I'm not understanding.  Doesn't the client support SSL now?
14178 [jerith jerit] This is a cryptographically signed message in MIME format.
14179 [james graypr] Ah, gotcha.

^ [patch] Making net/https.rb more documentable
14176 [jerith jerit] This is a cryptographically signed message in MIME format.

^ Selector Namespaces on a file basis?
14182 [transfire gm] Would it be possible to do selector namespaces on a file basis? That
14271 [matz ruby-la] I admit File based namespace seems simpler and understandable than

^ io and default external encoding
14185 [david davidf] "\u20ac\n".  I'm running in a locale that gives me a default external

^ Rake 0.8.0 added to Ruby
14186 [jim.weirich ] Just a heads up here.  I've added Rake (version 0.8.0 ... the latest)
+ 14188 [transfire gm] Why don't we just go ahead and add Rails now too.
| + 14190 [luislavena g] Oh come on... Rake is the most generic 'make' for ruby.
| + 14191 [jim.weirich ] There is nothing new here.  The bundling of rake has been discussed
| + 14198 [murphy rubyc] Because Rails isn't stable at all (critical updates frequently), has a
+ 14189 [jeremy bitsw] Awesome. This was a pleasant surprise when I just svn upped!
+ 14205 [murphy rubyc] * why is Rake included as a stdlib and not as a gem? how do I update it
  + 14207 [charles.nutt] I'm curious about this too. We've been shipping JRuby with rspec and
  | 14209 [Daniel.Berge] Let's make the entire standard library gems then. Seriously.
  | 14212 [james.britt ] Exactly.
  + 14230 [drbrain segm] Because the core team asked for it.
    + 14232 [transfire gm] Not true.
    + 14249 [murphy rubyc] thank you for all the answers. I'm feeling better now.
      14305 [drbrain segm] Uninstalling rake might break the command-line tool.  I haven't tested

^ Unbundling (Was: Rake 0.8.0 added to Ruby)
14192 [drbrain segm] My impression of unbundling would be to allow ruby packagers to

^ invalid hex escape for '\x'
14195 [duerst it.ao] I just used '\x' in both 1.8 and 1.9. In 1.8, this is just a string
14196 [matz ruby-la] Which version (revision) did you use?  My Ruby gives me two characters.

^ Setting length of a RString
14197 [gethemant gm] I ported mysql C bindings to run on 1.9, and I had to replace
14199 [murphy rubyc] I'm not sure if libraries are supposed to set string length themselves -
14200 [gethemant gm] yes true.. I read the same in README.ext for example. I would perhaps

^ Re: Problem with case statements
14203 [gethemant gm] Looks like they deprecated ":" in case statements.
14204 [gethemant gm] My bad.. i meant intentional, but you got the point anyways!.
14206 [murphy rubyc] you can replace them with ";" or "then" to be Ruby 1.8 and 1.9
14211 [rick.denatal] I suspect that it's gone because of a conflict with the new keyword

^ Re: [Spam] Rake 0.8.0 added to Ruby
14210 [ggarra advan] Yuck.
+ 14213 [rick.denatal] On Dec 21, 2007 10:24 AM, Gonzalo Garramu=F1o <ggarra@advancedsl.com.ar> wr=
| 14215 [jim.weirich ] I don't mind packaging rake as a preinstalled gem.  In fact, it makes =20=
| + 14216 [rick.denatal] Hmmm, I thought that the main motivation for pulling rubygems into the
| | 14221 [lists-ruby s] Oh, I'm sure it is!  Everyone's just waiting for someone to go first :)
| + 14217 [charles.nutt] For what it's worth, I don't think there's any chance of us installing
| | + 14218 [dave pragpro] One advantage of having Rake pre-installed is that the other Gems
| | | + 14226 [rick.denatal] 1) What comes pre-installed with Ruby.
| | | | 14227 [meinrad.rech] In addition to that, I must say that a library bundled to a given Ruby
| | | | + 14229 [meinrad.rech] Actually, I mean >>implementation<< in stead of >>distribution<<.
| | | | + 14235 [charles.nutt] Perhaps this is an even better argument for having more libraries as
| | | |   14248 [matz ruby-la] This is the path we are going to take.  Rake is an exception, just
| | | |   14253 [luislavena g] That will cover also extconf.rb without mkmf? using mkrf by Alain
| | | |   14292 [nobu ruby-la] I don't think it is possible soon, since BASERUBY may not be
| | | + 14228 [steven lumos] $ gem install anything -y
| | |   14233 [ed.odanow wo] ??? As I understood there will be a feature freeze December 2007
| | + 14219 [ed.odanow wo] This should only be done if it is possible to include a gem directly into
| + 14231 [drbrain segm] AFAIK there isn't yet.
+ 14214 [mgreenly gma] SSBhbHNvIGFncmVlIHRoZSB0cmVuZCB0byBwdWxsIHRoaW5ncyBpbiB0byB0aGUgc3RhbmRhcmQg

^ Matrix can't be subclassed
14223 [transfire gm] # Creates an +n+ by +n+ identity matrix.
+ 14225 [gavin.kistne] File a bug, then file a patch. It's easy enough to fix, right?
+ 14264 [matz ruby-la] I made a small patch.  Could you check it?  If it's OK, I will check
  14277 [transfire gm] Don't commit the patch. Unfortunately there's a complication. Vector

^ Status of two issues
14224 [transfire gm] Will the #autoload require issue be addressed?
14261 [matz ruby-la] Could you submit URL or mail ID [ruby-core:xxxx] along with your
14263 [transfire gm] T.
14265 [matz ruby-la] Thank you.
+ 14275 [transfire gm] Thanks for looking at these, Matz.
| + 14291 [akr fsij.org] I intend that open-uri use duck typing.
| | 14295 [rick.denatal] Akira-san
| | 14383 [akr fsij.org] I see.  I modified open-uri.
| + 14294 [nobu ruby-la] The reason I hesitate to make autoload to call Kernel#require
+ 14349 [rich infoeth] The RubyGems integration in 1.9 merely adds all the highest version

^ Information: Ruby 1.9 snapshot from 21 th December cannot be compiled
14234 [ed.odanow wo] Some minutes ago I tried to compile the actual daily snapshot.
14236 [daniel boven] "rb_default_encoding" was changed to "rb_ascii_encoding". parse.y was

^ io.c patch for two encodings in file mode
14237 [david davidf] This is a multi-part message in MIME format.
+ 14240 [david davidf] This is a multi-part message in MIME format.
+ 14254 [duerst it.ao] Thanks for this patch. I think Matz was planning to work on this,
+ 14288 [akr fsij.org] The conversion should be done before searching a terminator.
  14318 [matz ruby-la] Very hackingly fixed, by converting RS to the external encoding.  It
  14319 [akr fsij.org] ISO-2022-JP is not required.
  14321 [matz ruby-la] I know.  That's why I called it a hack.

^ Ruby 1.8 fails to compile under Mac OS X 10.5 (Leopard)
14247 [murphy rubyc] gcc -g -O2 -pipe -fno-common  -DRUBY_EXPORT  -L.    main.o dmydln.o \
+ 14970 [rogerpack200] I know the mac port works--you might download it and see what changes
+ 14971 [laurent.sans] It's a well-known bug, try applying the following patch, then rebuilding.

^ Semantics of caller (was Re: Some Ruby 1.9 loose ends to tie up)
14251 [rocky.bernst] On Dec 21, 2007 8:56 PM, Charles Oliver Nutter <charles.nutter@sun.com>
14259 [charles.nutt] We statically, conservatively optimize away frames at present.
14260 [rocky.bernst] Thanks for the detailed explanation.  A suggestion  to those who omit call
14262 [rocky.bernst] I hope this isn't too tedious but a couple of more thoughts here.  The

^ [BUG] Bus Error: block_given? (trunk)
14279 [murphy rubyc] def foo
14281 [murphy rubyc] and the same with

^ Block-local variables - obsolete?
14280 [murphy rubyc] excuse me, but what was the purpose of the "|vars; local|" block syntax
14282 [charles.nutt] Only block parameters are now always local to the block; variables
14283 [murphy rubyc] is this really the difference? both produce the "shadowing outer local
14285 [ryand-ruby z] I _think_ what Charles was getting at is by forcing local you can't
14287 [rick.denatal] irb(main):001:0> m = 1
+ 14299 [murphy rubyc] a very useful warning.
+ 14544 [dblack rubyp] Hopefully anything that warns of a deprecation or change like that
  14545 [rick.denatal] I  absolutely agree with this.

^ Racc doesn't play well with yield on Ruby 1.9
14284 [murphy rubyc] another thing that's breaking Rails on Ruby 1.9 is Racc (Ruby parser

^ Why is Array not Comparable
14286 [duerst it.ao] Why does Array not include Comparable?
14289 [rick.denatal] Good question.
14360 [duerst it.ao] Good points, but they only make good reasons for why Array
14572 [murphy rubyc] I don't agree. with this logic, Array#sort would have to go then,
+ 14575 [dblack rubyp] Array#sort depends on the array's elements responding to <=>, not on
| 14652 [murphy rubyc] Just like Array#<=> :) That was my point.
+ 14577 [rick.denatal] Rick DeNatale

^ IDBG Tool Suite -- Added instructions for adding Ruby Methods.
14296 [ceo hawthorn] The "idbg_tool_suite" was designed to help add/remove debug code

^ Testing on Athlon64 X2
14297 [znmeb cesmai] I'm just about to start "final testing" of Ruby 1.9 on a Gentoo Linux
14298 [murphy rubyc] it seems to be still in flux (10 commits in the last 10 hours), so I
14327 [znmeb cesmai] Yeah ... sounds like an excellent idea since there's supposed to be a

^ [PATCH] URI Regexp warnings
14302 [drbrain segm] in lib/uri.  May I apply it?
+ 14315 [murphy rubyc] you can also remove the second parameter if it's "false", because that
+ 14329 [matz ruby-la] Yes, please.
  14342 [drbrain segm] I have committed [murphy]'s patch for URI.