14126-15173 subjects 14501-16805

IRHG - GC Memory Fragmentation?
14303 [ceo@ha th rn] While working on Chapter 05  and referencing various works
+ 14308 [ryand-ruby@z] Where is all this work going? I looked for it the other day and found
| 14309 [ceo@ha th rn] The latest version of IRHG is at   hawthorne-press.com --
| 14322 [ryand-ruby@z] 1) your site is 100% unusable in safari.
| 14347 [ceo@ha th rn] 1) - As far as I know 'safari' has no connection to my project
| + 14351 [halostatue@g] I use the Safari web browser (as many Mac users do). Your site does
| | 14387 [ceo@ha th rn] OK --- As I have been throughly embarrassed since my web page was not
| + 14356 [stephen.bann] We continue to need help in translation and techinical review
|   14359 [ceo@ha th rn] OK --- That is what I needed.
+ 14346 [duerst@it ao] Yes, on the ruby heaps, all objects are represented by 20 bytes,
| 14358 [ceo@ha th rn] Thank you,
+ 14651 [ks@ku ts ep ] Ruby objects, like Strings and Vectors, explicitly allocate
  + 14680 [ceo@ha th rn] Thnaks for the input !
  + 14682 [matz@ru y- a] I'd like to see it too.  When I linked Ruby with Bohem GC at very
    14684 [ks@ku ts ep ] Using BDW GC will require changes to finalization and probably Ruby's
    + 14686 [gwtmp01@ma .] If Fixnums had a two-bit tag of 00, how would you distinguish between
    + 14784 [matz@ru y- a] Why not?  Symbols are used often.  Making it immediate saves memory.

[PATCH] lib/open-uri.rb warning
14304 [drbrain@se m] This patch fixes a method redefined warning on open-uri.rb redefining
14306 [akr@fs j. rg] OK.

[PATCH] more 1.9 warning fixes
14307 [drbrain@se m] May I apply these patches?

[PATCH] Fix Symbol#to_proc
14310 [frederick.ch] I noticed that as of 14511 Symbol#to_proc isn' working as I would
14312 [nobu@ru y- a] Sorry.
+ 14313 [frederick.ch] charset=US-ASCII;
+ 14317 [matz@ru y- a] Seems old code remained.  Could you check in?

Problem with ri and gems in 1.9
14314 [dave@pr gp o] It looks as if the path to the installed gem documentation is no
14323 [ryand-ruby@z] Beat up on Eric directly. He was playing in this space last night/this
14343 [drbrain@se m] ri --help doesn't display the actual paths used, so still worked right

Return value for Class#name for unnamed classes
14320 [murphy@ru yc] $ ruby -e "p Class.new.name"
14324 [matz@ru y- a] Since unnamed-ness is a temporary state (it can be named afterwards),
14325 [murphy@ru yc] it's trivial, since nil.to_s == "".

Gem loading twice on Ruby 1.9?
14326 [murphy@ru yc] this problem with Rubygems 1.0 only appears in Ruby 1.9. I load some gem
14334 [drbrain@se m] I have a patch ready to check in, it was too late to do it competently
14341 [drbrain@se m] Committed with r14564.

[PATCH] Spell checking for README.EXT
14328 [shiba@ma l2 ] a simple one, and the meddling one.  Please take one which you like.
14331 [shiba@ma l2 ] Thanks matz, and here is more patch ;)
14332 [matz@ru y- a] Oops, I am awfully sorry.  Fixed now.

[PATCH] RDoc update for Array#at
14330 [shiba@ma l2 ] Thanks to brand new VM's specialized instructions, now Array#at seems

1.9 problem with function *args parameters.
14333 [ggarra@ad an] require 'fileutils'
+ 14337 [murphy@ru yc] faszinating! it should be a syntax error...no?
+ 14338 [murphy@ru yc] oh, I forgot: just surround your hash with { } if it is not the last
+ 14340 [matz@ru y- a] 1.9 only allows hash without braces at the very last of the
  14345 [ggarra@ad an] I thought I was doing that, but it is clearly not what the parser thinks

Many external symbols _without_ prefix in libruby object file
14335 [shiba@ma l2 ] Unfortunately, I found many new external symbols without prefix in
+ 14336 [duerst@it ao] I'll take care of these within an hour or so.
| 14339 [duerst@it ao] Done in revision 14561.    Regards,    Martin.
| 14361 [shiba@ma l2 ] I comfirmed this.  Thank you for your quick and great work.
+ 14344 [matz@ru y- a] Thank you.  I'd like to merge the patch, but it's corrupted.  Could
| 14362 [shiba@ma l2 ] Oops, sorry.  And I added #undef things in vm.c.
+ 14357 [ko1@at ot ne] IMO, it's easy to recognize ID insteand of single character.  If
| 14363 [shiba@ma l2 ] Hmm, I don't prefer it, sorry.
+ 14405 [shiba@ma l2 ] I'm still continuing the work.  Here is the patch for dtoa().
+ 14406 [shiba@ma l2 ] IMHO, externded function'name that begin with "is_" is not a good nor
  14465 [hramrach@ce ] Yes, I had some problem with libruby exporting symbols that clashed

build problem?
14348 [dave@pr gp o] ./enc/depend:6:in `open': No such file or directory - ./enc/trans
+ 14352 [rich@in oe h] I got the same thing.  If you mkdir that dir it builds.  I assume
+ 14354 [duerst@it ao] Nobu is preparing that directory for the transcoding data files

"gem server" on Ruby 1.9?
14350 [znmeb@ce ma ] I downloaded Ruby 1.9 from SVN. It built fine. I changed my path so it's
14353 [murphy@ru yc] gem19 server works for me on http://localhost:8808.
14355 [znmeb@ce ma ] Yeah ... I found out the problem ... bad KVM switch. :( It works fine if

RDoc: [FATAL] failed to allocate memory
14364 [duerst@it ao] With revision 14590, I suddenly get an error when I do "make install"
+ 14365 [suppakilla@g] Same here, I can't get the "make install" process to complete on a
| 14368 [ko1@at ot ne] Could you update trunk and retry it?
| + 14370 [duerst@it ao] The problem persists at r14597. If you ask us to update,
| + 14371 [suppakilla@g] It seems fixed now, make install completes successfully on debian sid
|   14372 [duerst@it ao] It's not fixed here (cygwin). It's not an issue of how much memory,
|   + 14373 [masaki.suket] In my cygwin environment, trunk version works fine.
|   | + 14382 [duerst@it ao] Thanks for checking! It seems that I'm not so lucky. I have tried
|   | + 14415 [duerst@it ao] I finally also managed to have things work again.
|   + 14374 [suppakilla@g] Unfortunately I don't have a cygwin environment set up on my windows
+ 14366 [drbrain@se m] What are your process limits?
  + 14369 [duerst@it ao] It's not an RDoc problem. I get the same [FATAL] error
  + 14375 [ed.odanow@wo] I build Ruby 1.9 daily from the daily snapshots on native Windows2000

replace csv.rb with fastercsv.rb
14367 [nakahiro@sa ] Sorry for long time no post.  My main business is too much fun in these
+ 14384 [james@gr yp ] I can try.  I'm ill today and on my way to a doctor.  I could do it
+ 14390 [james@gr yp ] Matz answered my time question, so I can probably get the change in if
  + 14391 [james@gr yp ] I just checked, by way of example, and Rails does use csv.rb.  On
  | 14626 [murphy@ru yc] def read_csv_fixture_files
  | 14646 [james@gr yp ] That should do it, yes.  Do you want to submit the patch to their
  | 14649 [murphy@ru yc] I think you should. They may be applying it right away. Having "Ruby 1.9
  | 14693 [james@gr yp ] I've added the ticket.  You may want to note that you have tested it,
  | 14733 [murphy@ru yc] Thank you. It's verified.
  + 14396 [nakahiro@sa ] Hope you get well soon.  Sorry for dropping you in it.
    + 14397 [dave@pr gp o] Are there any points you'd like me to make about it when I update the
    | + 14409 [james@gr yp ] Yes, I have some thoughts on this.  Let me have a look at what you
    | + 14692 [james@gr yp ] * It's the FasterCSV code base, renamed to CSV.  Many people are =20
    + 14408 [james@gr yp ] I wish we had more time to evolve the interface into something you do

Re: [tangent] IRHG - GC Memory Fragmentation?
14376 [ryand-ruby@z] thanks for answering my q's... to follow up
+ 14377 [ryand-ruby@z] I prolly should have read the rest of the thread... tho my #2 wasn't
+ 14388 [ceo@ha th rn] It should be fixed now --- If you have any more troubles let me know.
  14417 [ryand-ruby@z] still says "Ultility"
  14424 [ceo@ha th rn] Ryan,

module_eval change between ruby1.8 & ruby 1.9. Is it intentional ?
14378 [frederick.ch] Consider this code
+ 14379 [ko1@at ot ne] This is a bug.  If you write a block in method do_stuff_to_foo
| + 14381 [frederick.ch] Cool. Thanks!
| + 14653 [murphy@ru yc] Any progress? It seems to be the last thing that makes Rails tests fail.
|   14719 [ko1@at ot ne] release.
+ 14380 [murphy@ru yc] module Foo

Crash in ruby 1.9
14385 [frederick.ch] I've found the following (extracted & abstacted from part of rails)
14386 [ed.odanow@wo] Same occurs with the actual snapshot from yesterday evening (Windows, MinGW/MSYS)...

pre-release note for the christmas release.
14389 [matz@ru y- a] I think I have almost done the basic design & implementation work for
+ 14393 [charles.nutt] Can we expect no more features or incompatible changes between 1.9.0 and
| 14394 [matz@ru y- a] Yes, basically, except for the possibility that we have to fix
+ 14401 [alterego@sd ] Thanks Matz, You're my hero. Have a merry Christmas :)

Fwd: replace csv.rb with fastercsv.rb
14392 [james@gr yp ] (Sorry for the duplicate post, I want to make sure NaHi sees this.)

Segmentation fault in 1.9 trunk
14395 [d.bussink@gm] When running the Rails tests, one of the tests failed with the

Timeout.timeout expires early - breaks RubyGems
14398 [drbrain@se m] Sometimes Timeout.timeout expires early.  I see this with `gem

Re: [PATCH] Timeout.timeout expires early - breaks RubyGems
14399 [drbrain@se m] Sometimes it seems that the timeout thread is woken early.  This patch
14400 [drbrain@se m] Sorry, this is the correct patch.
14402 [matz@ru y- a] Could you commit?
14403 [dave@pr gp o] Do we know why the thread wakes early? I ask because I'm thinking
14404 [ko1@at ot ne] Timer event wakeup thread.  This timer event is used for thread
14414 [drbrain@se m] Thank you, it is fixed now.

[PATCH] make native_* functions static in thread_*.[ch]
14407 [shiba@ma l2 ] I made native_{mutex,cond}_* functions static, and this is the patch.

Rolling map_with_index
14410 [james@gr yp ] While hunting bugs caused by changes in Ruby tonight, I was surprised =20=
14412 [matz@ru y- a] No, thank you for the report before the release.

test-all on Debian/GNU etch
14411 [ko1@at ot ne] This is current test-all result on my laptop (Debian/GNU Linux etch

result of test-all on FreeBSD 6.2
14413 [ko1@at ot ne] This is a result of test-all on 6.2-RELEASE-p4 on VMware.

1.9.0 status as of 12/25 19:00
14416 [matz@ru y- a] As someone may expect earlier, the release have not happened yet as of

Base64 not there makes Rails 2.0.2 fail to load in 1.9.0
14418 [rich@in oe h] Just re-built latest svn of 1.9.0 and base64.rb is removed.  Its
14419 [frederick.ch] edge rails already has :-)

Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0)
14420 [drbrain@se m] I don't think a legacy gem should be created.  This doesn't encourage
+ 14423 [rich@in oe h] Thanks Eric, I have been using array pack for base 64 encoding since I
+ 14427 [znmeb@ce ma ] Yeah ... I'm with you on this one, having just discovered that two gems,
  + 14429 [gethemant@gm] Yes lots, Almost ALL C extensions would require some kinda tinkering.
  | 14430 [znmeb@ce ma ] 1. Isn't there a pure Ruby mysql gem?
  | + 14432 [ggarra@ad an] Are you using the same ruby scripting language versions I am?  :)
  | | 14459 [charles.nutt] Which test is this? I'd like to try it in JRuby.
  | | 14482 [ggarra@ad an] for mandelbrot.  It is somewhat akin to the app_mandelbrot of ruby1.9
  | | 14493 [drbrain@se m] RubyGems is not loaded in 1.9 until you use it (some method on Gem).  =20=
  | + 14438 [james@gr yp ] I think this was true even before all these great new interpreters =20
  | | 14467 [halostatue@g] This is one reason I never implemented any of the libraries that I've
  | + 14439 [flori@ni e. ] ...
  | | 14451 [d.bussink@gm] submitted it to the MySQL project on Rubyforge. It's a good place to
  | | 14457 [flori@ni e. ] I found it suspicious, too, that I should be the only one, that has ever
  | + 14460 [charles.nutt] I've been advocating this for JRuby for newcomers looking to write
  |   + 14461 [frederick.ch] Is that going to complicate things vastly? Instead of having the 2
  |   | + 14462 [charles.nutt] Like Matz, I'm not in the business of limiting what people can do.
  |   | | 14464 [rocky.bernst] ...
  |   | + 14468 [znmeb@ce ma ] *Three* platforms (MRI 1.8, KRI 1.9 and jRuby 1.8) and *more* on the way
  |   + 14471 [znmeb@ce ma ] Well ... I think flexibility is a good thing in general. Almost every
  + 14431 [dave@pr gp o] My experience is that a whole bunch of gems are broken. Some fail
  | + 14433 [gethemant@gm] I am working to make couple of gems (C extensions) compatible with
  | | + 14434 [wcfarrington] To me, the logical choice is to take the same approach one does with
  | | + 14435 [filipe@ic wa] #ifndef RSTRING_PTR
  | + 14436 [znmeb@ce ma ] Well ...
  | | 14437 [dave@pr gp o] I personally think that bad PR from stuff not working is probably
  | + 14444 [james@gr yp ] Sadly, this is spot on.
  | | 14695 [james@gr yp ] Please feel free to add your own observations.
  | + 14446 [drbrain@se m] There is a required_ruby_version attribute in Gem::Specification.
  |   14452 [dave@pr gp o] and I'm suggesting making it default to <= 1.8.6. That way, we'll be
  |   + 14453 [vincent.isam] Making it default to < 1.9.0 would probably be better as I think 1.8.7
  |   | 14454 [dave@pr gp o] Yes. Good point.
  |   + 14492 [drbrain@se m] I think it would be better to add a warning if required_ruby_version
  |     14494 [dave@pr gp o] What about all the untouched gems on rubyforge, though. If folks don't
  |     14503 [rich@in oe h] I think it would be good to warn in 1.9 if the gem DOES NOT have the
  |     + 14505 [charles.nutt] I raised this question on the RubyGems list and didn't get the answer I
  |     | + 14508 [luislavena@g] I missed that message (or maybe the subject was different).
  |     | | 14526 [james@gr yp ] If you call zip() anywhere, your code is probably broken now.  If you
  |     | + 14513 [znmeb@ce ma ] Ouch ... there are already two different versions of gems with C
  |     | | 14514 [luislavena@g] Ruby mswin32 will work on Win64, using WOW (Windows On Windows) 32bits
  |     | | + 14530 [james@gr yp ] array_1.zip(array_2).to_a
  |     | | + 14675 [rogerpack200] +1 for mingw :)
  |     | + 14528 [duerst@it ao] For M17N, it really depends on what you are doing. Many things
  |     + 14532 [drbrain@se m] Here is a summary of the values of required_ruby_version for all the
  |       14546 [dave@pr gp o] From my work on the PickAxe3, and based on a small sample (of about
  + 14488 [gregory.t.br] Ruport and PDF::Writer will need to be updated.
    14491 [halostatue@g] Color should not have any issues; Trans-Simple may, but again should

Ruby 1.9.0-0 benchmarks
14421 [znmeb@ce ma ] ...
+ 14428 [ko1@at ot ne] Cool!
+ 14485 [ggarra@ad an] And here's my summary.

Paths, environment variables, etc. for building C extensions with Ruby 1.9?
14422 [znmeb@ce ma ] ...
+ 14425 [frederick.ch] Well as far as hpricot goes, you used to be able to get at the length
+ 14426 [gethemant@gm] This came over and over again on ruby-core i suppose. Hpricot is using
+ 14486 [alex@pr ss r] One thing that I couldn't find to build C extensions with in the

Odd situation with mingw, readline and readline tests.
14440 [luislavena@g] Posted about this a few weeks back on ruby-talk, but guess noone saw
14442 [nobu@ru y- a] I suspect readline doesn't consider it can run with redirected.
14443 [luislavena@g] Thank you for your answer Nobu, I'm looking into this right now and it
14456 [luislavena@g] Readline 5.0 binaries from gnuwin32: eat CPU and hang on the same line.

IDBG - Fixed idbg_manual.html
14441 [ceo@ha th rn] In addition to cleaning up the HTML code in my website as requested

14445 [amaurya@br c] ...

PATCH: fix socket support for 1.9 on IRIX
14447 [andrew@hi ac] I know I'm probably one of like 5 people still using IRIX, but theres no
14489 [matz@ru y- a] It is merged in the trunk.  Thank you.

Multiple ways to do a thing
14448 [jarpcs@gm il] There are multiple ways to do a thing ruby, while some other languages
+ 14449 [ryand-ruby@z] It makes Matz (and many of us) happier. (Matz, correct me if I'm
+ 14481 [murphy@ru yc] it is a bit self-explanatory if you just accept that there is no "right
  14527 [duerst@it ao] As for / versus /.../, that one got me caught badly in more than an

Gem versioning not quite right?
14455 [dave@pr gp o] dave[RUBY3/Book 9:00:56] gem list
14469 [rich@in oe h] Dave,
14475 [rich@in oe h] This is now fixed in RubyGems trunk and once eric does his magic it

Re: C++ Functors and Ruby extensions
14487 [jameskilton@] ...
+ 14495 [ggarra@ad an] Which is partially not true.  SWIG supports nested classes just fine but
| 14496 [ggarra@ad an] Actually, just looking at the swig internals, the typedef is unneeded
+ 14506 [pbrannan@at ] Your code doesn't use dynamically-generated function pointers; it uses

Status of Event Driven Network Programming Libraries in 1.9
14497 [gethemant@gm] Anyone doing serious event driven network programming on *nix will
+ 14498 [mumismo@gm i] ...
+ 14499 [garbagecat10] ...
+ 14509 [pbrannan@at ] I think this is a good idea.  It would be cool if the user could plug in his
  14510 [gethemant@gm] It depends upon the implementation. If it has to go in standard

[1.9] rb_catch()
14500 [decoux@mo lo] Well, surely I'm making something stupid but here the  problem
14511 [ceo@ha th rn] How did you produce the control-frame
14571 [decoux@mo lo] Just crash ruby and it will do it for you :-)