2057-2339 subjects 2467-3290

[patch] CGI#server_port returns 0
2215 [d.borodaenko] @@ -923,7 +923,7 @@

ruby aborts in data-handling applications
2216 [xsdg@fr en d] I reported a similar bug about 2 or 3 months ago.  The problem seemed to go
+ 2217 [matz@ru y- a] I really really want to help you, but without error reproducing
| 2277 [xsdg@fr en d] I've finally found a combination that will reliably segfault.  Since my first post on this topic, I've switched the backend from BDB to GDBM.
| 2278 [nobu.nokada@] What about with CVS HEAD?
| 2279 [xsdg@fr en d] $ruby1.9 -I /usr/lib/ruby/1.8/i386-linux/ ./testcase.rb=20
| 2280 [nobu.nokada@] It should not work.
| 2281 [xsdg@fr en d] It works just fine with a data file not long enough to trigger the bug (I
| + 2282 [nobu.nokada@] Use matched version libraries.
| + 2283 [decoux@mo lo] Why do you need this ?
|   2284 [xsdg@fr en d] didn't realize that gdbm was included with ruby (didn't have the -dev lib
+ 2218 [ruby-lists@l] maybe testing with valgrind would help?, else the best
  2221 [decoux@mo lo] probably not,

Fwd: [ruby-cvs] ruby: * file.c (test_wr, test_ww): New functions implementing new
2225 [dave@pr gp o] dave[ruby/ruby 8:32:41] rdoc --ri --al file.c
+ 2229 [jfh@ci e. fl] Hear! Hear!
+ 2244 [ian@ca ib n.] I'm confused. I did include documentation for the above methods. Can you

Re: [ruby-cvs] ruby: * file.c (test_wr, test_ww): New functions implementing new
2231 [dave@pr gp o] It does: when you run rdoc it warns about each such one. It doesn't
2245 [dave@pr gp o] My mistake - sorry for the berating... :)
2247 [ian@ca ib n.] Ah, good to know; thanks. Since you have CVS access, can you fix that to
2248 [matz@ru y- a] Done.

YAML_Unit_Tests failed
2251 [nahi@ke na t] The test of the CVS HEAD fails.
2255 [ruby-core@wh] Thank you, Nahi.  Sorry, was a thoughtless omission on my part.  Lazy
2257 [gsinclair@so] What about the 1.8 branch?  Is that relevant here?
2258 [nahi@ke na t] Thanks.
2270 [ruby-core@wh] I would like to see this change merged.  I will see that it is done and

(forw) (forw) rdoc/ri - makefile
2253 [skaar@ao .n ] Mime-Version: 1.0

[PATCH] ChangeLog temporal anomaly
2260 [sunshine@su ] Here is a patch which corrects a temporal anomaly in the ChangeLog.  One of

Slow memory leak with 1.8.1, rb_yield
2266 [djberge@qw s] I've only tested this so far on Windows XP with VC++ 7.
+ 2267 [djberge@qw s] Oops, you don't need the windows.h file.
+ 2268 [matz@ru y- a] I don't call it a leak, as long as memory is reclaimed by the next GC
  2269 [djberge@qw s] I guess I need to do better research on when GC automatically occurs and

A suggestion for libraries such as base64.rb
2285 [dave@pr gp o] Some of the older libraries simply insert stuff into the top-level
+ 2286 [djberge@qw s] Apparently this RCR dropped by the wayside. :(
| 2287 [nathaniel@ta] Someone should resubmit it with the more informative structure of the
| 2288 [djberge@qw s] I was surprised it wasn't implemented in lieu of
+ 2289 [nobu.nokada@] I guess it should be warned, rather than just aliased.
  2290 [dave@pr gp o] Can we commit this?
  2291 [matz@ru y- a] Yes.
  2292 [gsinclair@so] Should this go into 1.8 branch as well?

CVS Accounts
2293 [BCoish@Dy ax] Sorry to intrude, but can someone please tell me how to sign up for a
2294 [matz@ru y- a] Usually patches are posted here, and we commit them for you.  If you
2295 [BCoish@Dy ax] Thanks for the timely response, it's appreciated.
2296 [matz@ru y- a] Unified diff (diff -u) with brief explanation is preferred.  In short,
2297 [BCoish@Dy ax] Ok, Thanks.

Problem with Exceptions and Threads
2298 [pit@ca it in] recently I had a problem writing unit tests using timeout.rb and other
+ 2299 [dooby@d1 .k ] [snip code]
+ 2300 [matz@ru y- a] When the main thread terminates, the other threads are terminated
  2304 [pit@ca it in] DOH! Thanks Daz and Matz

pp.rb patch to optionally sort hash keys
2301 [lists@za a. ] ...
2302 [akr@m1 n. rg] I don't merge the patch because a global option causes trouble with
2303 [lists@za a. ] Thanks for your response. Yes, I was worried about dealing with threads.

[syck] Time#usec round trip problem
2305 [aamine@lo er] I have found that the syck library does not dump
2307 [aamine@lo er] Sorry, this patch is useless.
2309 [ruby-core@wh] Thanks for the patch, Minero.  I am adding this patch, along with a
2316 [aamine@lo er] Thank you very much.  It completely works now.

[syck] YAML.dump("a".."z")
2306 [aamine@lo er] ~ % ruby -v -ryaml -e 'puts YAML.dump("a".."z")'
2310 [ruby-core@wh] This is improved as well.  I have added tests as well to confirm.
2311 [nobu.nokada@] Seems to work.

CGI::Session FileStore patch
2308 [emschwar@fc ] I tend to fork off long-running processes from CGI scripts.  The

[BUG] Find.find("xyz") will yield "xyz" even if no such file exists
2317 [gsinclair@so] The subject says it all, really.  I presume this is a bug, not a
+ 2318 [vjoel@PA H. ] $ ruby-1.9.0 -v -r find -e 'Find.find("xyz"){|f|p f}'
+ 2319 [matz@ru y- a] Thank you for reporting it.

Problems in mathn, rational, complex, matrix
2320 [gsinclair@so] I received a message from Richard Graham mentioning a problem in the
2321 [matju@sy pa ] You have to do Matrix[[7.0,6.0],[3.0,9.0]], or the equivalent thing with
2322 [gsinclair@so] I'm way behind you in the mathematics (snipped), but I remember taking
+ 2323 [austin@ha os] irb(main):001:0> require 'mathn'
+ 2324 [matju@sy pa ] A number like 1.0 is mathematically an Integer too. However, in a language

Ruby 2 Status
2327 [BCoish@Dy ax] I saw a post to comp.lang.ruby asking what the status of Ruby 2 is

[PATCH] CGI::Session.rb
2328 [emschwar@fc ] I'm sorry, I found a typo in my previous version of the patch.  Here
2907 [matz@ru y- a] Sorry for being late.

rb_io_puts is strange
2335 [ehughes@bl e] i just tracked down some incorrect behavior (missing newlines in the output)
2337 [nobu.nokada@] It's the feature.
2340 [ehughes@bl e] [re: puts behaving differently depending on the final character of the
+ 2341 [matz@ru y- a] "bad" is a strong word.  How bad do you think?
| 2342 [ehughes@bl e] * It has the same name as C's puts, but doesn't have the same
| + 2343 [jlst@ge ty g] At the risk of sounding old and tired I feel compelled to point out
| + 2344 [matz@ru y- a] I understand what you are saying.  But still, I don't feel they are as
+ 2345 [matju@sy pa ] class IO
  2366 [ehughes@bl e] [i've been trying not to respond, but, just for the record...]

REXML Bug in IOSource's initialize?
2336 [kapheine@hy ] I posted something about this to ruby-bugs but various people seem to

Patch for socket.c: control reverse lookup for every instance
2346 [uehli@bl ew ] ...
2347 [nobu.nokada@] Your patch makes impossible to control it in singleton method
+ 2348 [matz@ru y- a] Commit the patch, please.
+ 2351 [uehli@bl ew ] Yes, I wanted to add that today, thank you very much for your corrections.

RDoc should recognise https:// links
2349 [ian@ca ib n.] ...
2350 [dave@th ma e] Cheers

[PATCH] Persistent IRB history
2352 [uehli@bl ew ] ...
2354 [nobu.nokada@] Sounds nice, but I guess it would be better if configurable.
+ 2356 [uehli@bl ew ] Yead, that sounds good.
+ 2358 [matz@ru y- a] Commit this patch, please.

RCR Mailing List
2353 [transami@ru ] This post is primarily directed to Matz, but is posted here for everyone to
2355 [chad@ch df w] #
2359 [transami@ru ] I do have a mailing list, and for the three persons who use it, it has worked

[PATCH] Use the BasicSocket#do_not_reverse_lookup flag in Webrick
2357 [uehli@bl ew ] With this patch you can make use of the new per-socket do_not_reverse_lookup flag in Webrick.
2360 [kawaji@ho pe] I also submitted this pacth.
+ 2361 [uehli@bl ew ] So I'm not the only one... ;-)
| 2363 [kawaji@ho pe] I have not noticed it.
+ 2362 [nobu.nokada@] Perhaps, you miss Thomas'es patch.  It uses new per-instance

Maybe bug
2364 [sdmitry@lr .] ...
2365 [matz@ru y- a] Don't symlink extensions.

Standard libraries
2367 [dave@th ma e] I'd like to propose two more criteria.
+ 2368 [ser@ge ma e-] charset="iso-8859-1"
| + 2369 [dave@th ma e] one of the three option parsing libraries, perhaps, or delegate.rb. My
| | 2371 [ser@ge ma e-] charset="iso-8859-1"
| + 2372 [gsinclair@so] There's ruby/test for unit tests.  From my POV, a README or something
+ 2370 [chad@ch df w] # >
+ 2373 [matz@ru y- a] I didn't see any smiley here, how much are you serious?
  + 2374 [gsinclair@so] Sounds good.  Is there sense in starting a discussion about removing some
  | 2375 [matz@ru y- a] Yes.
  | 2377 [gsinclair@so] OK, what about 'cgi-lib', 'getopts', 'importenv', and 'parsearg'?  Here
  | 2382 [matz@ru y- a] Take steps to remove them.  But first we need to warn deprecation for
  + 2376 [nahi@ke na t] RDoc style is a must?  I don't like documentation buried within
  | + 2378 [gsinclair@so] RDoc is the standard way to document standard library files.  It's only
  | | + 2379 [nahi@ke na t] Anyway, duplication of something I don't like to see/maintain.
  | | | 2381 [gsinclair@so] I'm still not sure what you think is being duplicated.
  | | | + 2386 [nahi@ke na t] For example, a documentation of current CSV.open is as follows.
  | | | | 2392 [batsman.geo@] Isn't this precisely the point of using RDoc? Helping (forcing?) the
  | | | | 2393 [gsinclair@so] Besides, interfaces don't change often, after the software is a little
  | | | | 2401 [ser@ge ma e-] charset="iso-8859-1"
  | | | | 2405 [gsinclair@so] Good points, but I don't see the light overall.
  | | | | + 2410 [ser@ge ma e-] charset="iso-8859-1"
  | | | | | 2412 [dave@th ma e] Quite often though you can get away with some common sense parameter
  | | | | | 2414 [ser@ge ma e-] charset="iso-8859-1"
  | | | | | 2415 [dave@th ma e] I'd personally document the method to say
  | | | | | 2420 [ser@ge ma e-] charset="iso-8859-1"
  | | | | | 2424 [dave@th ma e] Linking to specific methods would be fairly straightforward if you
  | | | | + 2425 [nahi@ke na t] Kind of agree-ed.  Yes, there's no other way.  Regards to
  | | | |   + 2426 [dave@th ma e] I don't think all of it is. I might document it as
  | | | |   | 2428 [nahi@ke na t] Sorry, I found wrong DESCRIPTION.  Convert to string?
  | | | |   | 2436 [ruby-lists@l] nod.
  | | | |   + 2427 [gsinclair@so] Of course I allow you :)  I'm curious what your reasons are.
  | | | |     2429 [nahi@ke na t] Thanks.
  | | | |     + 2430 [gsinclair@so] I agree.  I just like a little method documentation to get me started.  If
  | | | |     + 2431 [dooby@d1 .k ] ... In reply to: "Gavin Sinclair"
  | | | |     + 2432 [ser@ge ma e-] charset="iso-8859-1"
  | | | |     | + 2435 [nahi@ke na t] We seems to live in different world.
  | | | |     | + 2437 [jlst@ge ty g] NaHi's soap4r is a good example.  It's a big library.  In most general
  | | | |     |   2439 [dave@th ma e] What I've done is create a README in RDoc format, and then use
  | | | |     |   + 2440 [jlst@ge ty g] I hear you.  It's great.  I'm just saying it'd be great if we could
  | | | |     |   | 2441 [dave@th ma e] Unless there's an exact match...
  | | | |     |   + 2442 [gsinclair@so] I generally include the README in the documentation as itself, rather
  | | | |     |   + 2457 [ian@ca ib n.] When I tried to include an RDoc document into another RDoc document, the
  | | | |     |     2458 [dave@th ma e] No - that's always been the same behavior. Is it possible that there's
  | | | |     |     2462 [ian@ca ib n.] I think that must have been the case, as I can no longer reproduce the
  | | | |     + 2434 [eivind@Fr eB] My experience partially differs.  Yes, the user should be able to read
  | | | |       2447 [nahi@ke na t] Agreed.  I think so.
  | | | + 2394 [dblack@wo bl] When I first committed scanf.rb, Matz asked me not to include the unit
  | | |   2396 [matz@ru y- a] It's changed.  At the time I asked, we didn't have unit test included
  | | + 2380 [nahi@ke na t] In Japanese 1.4/1.6 document, "standard library" is distinct
  | |   2385 [gsinclair@so] [NaHi:]
  | |   2391 [nahi@ke na t] To know which library has Ruby quality.  "The Ruby Way" and
  | + 2383 [matz@ru y- a] I strongly recommend RDoc style documentation, but you can ask
  | + 2384 [dave@th ma e] I think one style is a good idea: whether it's RDoc or something else I
  |   2387 [nahi@ke na t] I don't like inline documentation, not RDoc formatting rule.
  + 2388 [ruby-core@wh] This sounds like an ultimatum.  How long do I have before the next
  | + 2389 [gsinclair@so] I'll certainly work on it and will work with you.  It's something I had in
  | + 2390 [matz@ru y- a] Usually I declare two or three weeks prior to the release.  So you
  | + 2400 [jim@fr ez .o] I've never browsed the YAML source...but I have never needed to
  + 2421 [jlst@ge ty g] I think we should do matz.clone and then have netlab hire that person
    2433 [ser@ge ma e-] charset="iso-8859-1"

Comments on base64.rb
2395 [gsinclair@so] While adding documentation to base64.rb and backporting it to 1.8, I
2411 [djberge@qw s] <good observations snipped>
2416 [austin@ha os] Actually, I'd like to see Base64 under an Encode:: toplevel module. This

PATCH: deprecate cgi-lib, getopts, importenv, parsearg from standard library
2397 [gsinclair@so] ===================================================================
2398 [emiel@il fo ] Allow me to make a minor suggestion. :) I always find it helpful when an
2399 [gsinclair@so] Good suggestion.
+ 2402 [matz@ru y- a] I recommend optparse to be alternative.
| 2403 [gsinclair@so] I couldn't agree more.  Should I commit with that adjustment?
| + 2404 [nahi@ke na t] Wait a few days.  Ruby development team need to rewrite
| | + 2407 [gsinclair@so] Matz said yes, but since this is hardly urgent, I'll wait a few days
| | + 2408 [matz@ru y- a] Sorry, perhaps the rewriting issue is slipped away from my brain.
| |   + 2409 [eban@os ri .] % grep getopts instruby.rb ext/extmk.rb lib/un.rb
| |   | 2413 [matz@ru y- a] Understood.  Sorry for troubles.
| |   | 2419 [nobu.nokada@] ropt <http://raa.ruby-lang.org/list.rhtml?name=ropt> seems nice
| |   + 2446 [nahi@ke na t] No need to apologies.  I had to explain that.  I'm sorry.
| + 2406 [matz@ru y- a] Yes, please.
+ 2417 [austin@ha os] Do we want to support both of these, as well? IMO, we should recommend "best
  2418 [gsinclair@so] Definitely agree, and the getoptlong recommendation has been dropped.  In

Re: [ruby-cvs] ruby: * lib/ftools.rb: documented
2422 [usa@ga ba ec] Gavin, didn't you forget committing lib/ftools.rb? :)
+ 2423 [usa@ga ba ec] BTW, I think lib/ftools.rb is deprecated.
| 2444 [gsinclair@so] I think it's a good idea.  Matz?
| 2445 [matz@ru y- a] Sounds nice.  But many libraries and installation tools are still
| 2448 [gsinclair@so] I agree its in fairly wide use.  Adding a warning now (presumably a
+ 2443 [gsinclair@so] Done now.  Commit date/time is different from Changelog.

ri data location
2438 [jlst@ge ty g] I propose that $datadir/ri be moved to $datadir/ruby/ri in the future.

make install not getting through rdoc phase
2449 [dblack@wo bl] I'm getting the following fatal error when trying to install the
+ 2450 [nobu.nokada@] What is output by followings?
| 2451 [dblack@wo bl] /home/dblack/ruby
| 2453 [nobu.nokada@] Of course, there is rdoc/ri/ri_paths.rb underneath
| 2454 [dblack@wo bl] It's not there, but now I'm starting completely from scratch (not
| 2459 [ser@ge ma e-] Inspired by David's success, I did the same thing, with the same result.
+ 2452 [ser@ge ma e-] No, I had the same problem, but I was working on another problem so I ignored

Re: [ruby-cvs] ruby/lib, ruby: * lib/generator.rb: corrected doc format
2455 [usa@ga ba ec] Gavin, please commit lib/rinda/rinda.rb and lib/rinda/tuplespace.rb.
2456 [gsinclair@so] Sigh, grumble, tear my hear out.  It sure is :)

PATCH: deprecate cgi-lib.rb getopts.rb importenv.rb parsearg.rb
2460 [gsinclair@so] I raised this recently and got the go-ahead, except that a few days
2461 [matz@ru y- a] Try committing them in CVS HEAD first.  You can back port them after

[PATCH] ri --help output typos
2463 [djkea2@mu ca] ===================================================================
2464 [dave@pr gp o] Many thanks - applied.

PATCH: OpenStruct#initialize to yield self
2465 [gsinclair@so] This is a common approach I use to object initialization; I don't know
+ 2466 [djberge@qw s] I vote "yay" on this one. :)
+ 2476 [matz@ru y- a] How much benefit do you get from this than
  + 2478 [a.bokovoy@sa] Atomicity? Ability to return nil as 'r' in case of failure of complex
  + 2480 [gsinclair@so] Readability is all I had in mind.  Makes the code look more
  + 2482 [vjoel@PA H. ] Someone mentioned before that it lets you construct an object that can
    + 2483 [matz@ru y- a] What is the benefit of "object that can be GC-ed as soon as initialize
    | 2484 [vjoel@PA H. ] Not much that I can see, unless all the work of the object is done in
    + 2485 [jlst@ge ty g] Neither of these cases would gc until the local went out of scope or
      2486 [vjoel@PA H. ] That's true. The point is that in the former case, r goes out of scope
      + 2487 [jlst@ge ty g] Yeah but 'record' still points to the struct just as 'r'
      | 2489 [vjoel@PA H. ] Nothing, in the case of OpenStruct, but another class's initialize
      | 2490 [jlst@ge ty g] Gotcha.
      + 2488 [nobu.nokada@] That r is the just created instance, isn't it?  I don't want it