1030-1293 subjects 1419-1650

^ Re: Timezones in ChangeLog
1232 [matz ruby-la] _why's message has been blocked.  I repost the message.

^ [Patch] FTP.new with block
1237 [gsinclair so] I've modified lib/net/ftp.rb (not committed, obviously) so that
1238 [decoux moulo] Why all persons think that File::new can take a block ?
+ 1239 [matz ruby-la] interesting idea though.  Net::FTP.open will take block in the
+ 1240 [matju sympat] That's psychological: when people have a lot of things to learn, they
  + 1241 [batsman.geo ] What was the reason?
  + 1242 [matz ruby-la] Reasonable.
  | + 1243 [gsinclair so] It described my thought pattern pretty well.
  | + 1245 [dave pragpro] Perhaps this time you should invoke the principle of least confusion...
  |   1246 [gsinclair so] - disallow block for File.new (i.e. no change)
  |   1247 [matju sympat] ... but the logic is pretty dodgy.
  |   1248 [austin halos] I would recommend a change to this, based on Mathieu's objection -- which I
  + 1244 [dave pragpro] Perhaps it's an argument for having File.new take a block too...
    1250 [rz linux-m68] something more systematic is needed. So far only 2 widely used
    1251 [decoux moulo] svg% ruby -e 'p Array.new(2) { 12 }'
    1261 [rz linux-m68] $ ruby -e 'p Array.new(2) { 12 }'
    1262 [decoux moulo] Hash.new do 12 end
    1276 [rz linux-m68] After suffering a few problems like that where the documentation
    + 1277 [decoux moulo] Well, the problem in this case is in the documentation, not in ruby.
    | 1294 [rz linux-m68] [ sorry for replying so late ]
    + 1284 [matz ruby-la] As Guy stated, the point is how to tell the interpreter the fact the
      1295 [rz linux-m68] true.. I have noticed looking at the ruby implementation :(.
      1296 [matz ruby-la] Agreed.

^ File.write(path, data)?
1249 [gsinclair so] I am glad to see File.read(path) in Ruby 1.8.  But what about
1252 [matz ruby-la] I've never felt the need for "write at once".  Do you feel often?
+ 1253 [sean celsoft] Perhaps, instead of read/write functions reading/writing from paths, create
+ 1254 [austin halos] I actually use the pattern
+ 1255 [gsinclair so] I've implemented it myself on all (3) projects I've done recently.  That's

^ testunit, exit status and at_exit
1256 [dave pragpro] I'd really like TestUnit to be able to return an exit status when I run
1260 [ser germane-] Ditto, and I'll even provide a use case.
1263 [nathaniel ta] I agree, too, but as Dave pointed out, my hands are tied at this point. I
1266 [matz ruby-la] OK.  But you need to help me to clarify the problem.  What if there's
1267 [nathaniel ta] I was wondering if that would be a problem. For my situation, either of the
+ 1268 [ryand-ruby z] I expect exit to do just that. I would want the first one hit to exit
+ 1269 [matz ruby-la] They are executed reverse order they are defined.  I think there's a
  1270 [nathaniel ta] Oops, I had that backwards.
  1273 [matz ruby-la] The first one defined, which executed last.

^ [RFC] Add have_defined() and rework have_struct_member()
1257 [m.rokos sh.c] I'm trying to bring OpenSSL < 0.9.6 to OSSL2 module and while doing this
+ 1258 [nobu.nokada ] There is macro_defined? function.
| 1259 [m.rokos sh.c] Hmm, you're right.
| + 1264 [matz ruby-la] I'm sure someone will be confused between macro_defined? and
| + 1265 [nobu.nokada ] Is it necessary?  You can use that macro itself in preprocessor
|   1274 [m.rokos sh.c] Of cource it is NOT necessary. I was just blind.
+ 1271 [akr m17n.org] I think HAVE_STRUCT_MEMBER is better because it's consistent to

^ status of openssl 0.2 support for OpenSSL 0.9.6
1272 [a.bokovoy sa] Greetings!
1275 [m.rokos sh.c] Thank you Alex, this is what we need.

^ IRB in latest CVS doesn't run
1285 [dave pragpro] dave[ruby/ruby 8:52:53] irb
1286 [matz ruby-la] That means your install is incomplete.  Here's the workaround.  Let me
1289 [dave pragpro] The patch works fine. Thanks.

^ rb_define_const()
1287 [g intersect-] I would mention that this is my first post, but then I'd sound very
1288 [g intersect-] rb_warn("invalid name given for constant('%s')",name);

^ Malloc problem
1290 [dave pragpro] ref_c_class.tip >/dev/null
1291 [ryand-ruby z] this is on Mac OS X, yes? Try the tool

^ Fix for Bug 1058
1297 [walser scs.c] charset="us-ascii"
1298 [matz ruby-la] Thank you.  It's reported already from Akira Yamada (Debian Ruby
1299 [walser scs.c] Does that mean that Akira Yamada found the bug just with the backtrace?
1300 [matz ruby-la] He said he forwarded the message from BTS (I think from Debian's).
1301 [walser scs.c] Bug 1060 is the next one. Unfortunately I couldn't create an small

^ Bug 1060
1302 [walser scs.c] In function ev_const_get (eval.c:1489) I printed the iv_tbl type.
1303 [matz ruby-la] Check the value of klass by
1304 [walser scs.c] (gdb) bt
1305 [matz ruby-la] This means a class pointed from cbase is broken, in a way never seen
+ 1306 [walser scs.c] when I take out -O2 from the top Makefile, the segfault disapears...
| 1307 [walser scs.c] -O1 also segfaults. I will do the checks next week again with the original
| 1308 [matz ruby-la] Problems reported with -O higher than 0 on IA64.  I'm not sure whether
+ 1318 [walser scs.c] What exactly is here brocken?
  1319 [matz ruby-la] flags should not be 512.  This is an impossible value.  I don't know
  1320 [walser scs.c] May be, but may be also the optimization discloses a problem in the code.
  1321 [matz ruby-la] I know.  But without a machine at hand, what I can do is very

^ exceptions and such
1309 [matju sympat] 1. any way to call c++ destructors of stack-allocated objects in
1316 [matz ruby-la] Ruby is originally designed as an application after all.  There are
1356 [matju sympat] Are you actively working on Ruby 2.0 now? How different will the new API
1357 [matz ruby-la] No.  I still need to work on 1.8 for maintenance.  The new API will
1374 [matju sympat] By reentrant you mean signal-safe and such ?

^ adding NodeDump and ii
1310 [nobu.nokada ] I'd like NodeDump and ii to be contained in standard
+ 1311 [decoux moulo] not a good idea, at least for ii
| 1312 [nobu.nokada ] It's sorry.  May I ask the reason?
| 1313 [decoux moulo] If someone write a version of ruby like this
| 1314 [nobu.nokada ] It's the reason why I want the two to be standard.  They hardly
| 1315 [decoux moulo] if the AST is a blackbox, these modules can't exist :-)
+ 1317 [dave pragpro] I have to admit that I'm getting somewhat worried by the size of the

^ compile error on win32 platforms with ext/syck
1322 [usa osb.att.] There is enum definition of "WORD" in ext/syck/gram.h, but
1323 [ruby-core wh] I'm going to prefix these tokens to avoid further nameclash.  Sorry.  My
1324 [usa osb.att.] I've checked. Thank you!

^ Tainting in YAML
1325 [ruby-core wh] I'm following Marshal's tainting code in my YAML extension.  Basically,
1326 [matz ruby-la] Yes.
1327 [ruby-core wh] Ok.  Thanks, matz.

^ Extracting a parent class
1328 [mgarriss ear] Sorry about the newbie question but....
1333 [matju sympat] Like most other OO languages, the class of an object is a property of the

^ Cleanup patch for singleton.rb
1329 [chr_news gmx] This is a multi-part message in MIME format.

^ -Wall cleanup project
1330 [djberg96 yah] Here are my "patches" to Preview 7 to help cleanup the
1331 [matz ruby-la] * do not use old diff format. It's unreadable.  I prefer -u
1363 [ehughes blue] diff -uN
1367 [matz ruby-la] Oops, gcc 2.x warnings are like
1368 [ehughes blue] Sorry, I'll sort a patch out tonight. Where should I send patches? (I'll
+ 1369 [ehughes blue] Having failed to find the problem in today's CVS and checking the cvs
| 1371 [matz ruby-la] This simple one in particular.  Feel free to make patches and post.
+ 1370 [matz ruby-la] Post here.

^ Re: Rubicon for ruby 1.8
1332 [dave pragpro] (This was originally on ruby-dev, but I don't seem to be able to post there)

^ Another Test::Unit patch...
1334 [nathaniel ta] I still can't commit (perhaps things are locked down in anticipation of the
1335 [nathaniel ta] Oops... the indentation was messed up on that one :-/. Here it is again...

^ [patch] "_FILE_OFFSET_BITS redefined" warning on Solaris (II)
1336 [holmberg iar] A while ago I submitted a patch to remove compile time warnings on

^ [PATCH] missing.h declarations missing return types
1337 [lyle knology] This is a multi-part message in MIME format.

^ 1.8.0: possible socket problem with mswin32 builds
1338 [jos catnook.] ruby -rnet/http -e "p Net::HTTP.get('www.ruby-lang.org', '/en/')"
+ 1339 [chr_news gmx] I tested this with my ruby 1.8.0 (release) compilation  and
| 1340 [usa osb.att.] What version of compiler do you use?
| 1350 [chr_news gmx] I use VC 7.1 - b.t.w. VC 7.1 finally produces a usable binary
+ 1355 [jos catnook.] It turns out that this is a false alarm, caused by an as yet unknown issue

^ make test fails on my box
1341 [dave pragpro] not ok system 9 -- /tc/dl/ruby/ruby1.8/sample/test.rb:1622
+ 1347 [matz ruby-la] Agreed.
+ 1348 [nobu.nokada ] sample/test.rb:1622 seems blank.  Have you edited that file?
  1353 [dave pragpro] Sorry - yes - I added a tracing line to see which file was blowing it up.

^ SEGV in GC under Linux
1342 [dave pragpro] A while back I was getting double free()s reported on my MAC box when
1343 [matz ruby-la] Can you try the following patch?
+ 1344 [matz ruby-la] I removed lines too much.  Here's the whole patch.
+ 1345 [dave pragpro] You amaze me. That fixed it: both the segv and the double free()s are
  1346 [matz ruby-la] Wish I had fixed it before the release.

^ bug in shell.rb
1349 [ian caliban.] The Shell.rmdir method in shell.rb doesn't work, because it passes the
1351 [matz ruby-la] Thank you for the fix.

^ intialize_copy for classes
1352 [chr_news gmx] is this a bug or feature (1.8 release)?
1354 [matz ruby-la] Neither.  It's implementation incompleteness, which is to be resolved

^ CVS access
1358 [nathaniel ta] I used to have CVS access... while I can log in OK now, I get the following
1359 [matz ruby-la] Seems like ACL error.  Somehow your account was dropped from ACL list.
1361 [knu iDaemons] Done!
1366 [nathaniel ta] Great! Works like a charm.

^ I was doing pretty well until...
1360 [dave pragpro] So I'm documenting all the new stuff in 1.8, happily adding all the new
1362 [nagai ai.kyu] Please see [ruby-talk:76218].
1365 [dave pragpro] Thank you! That was exactly what I needed.

^ Broken REXML in Ruby 1.8
1364 [a.bokovoy sa] Greetings!

^ tk out-of-box on Mac OS X [PATCH]
1372 [ehughes blue] This message is in MIME format. Since your mail reader does not understand
1373 [brian coolna] Do you think you could post this to ruby-talk as well?  It would be good
1375 [ehughes blue] *I* can't, because I'm not a subscriber to ruby-talk (too high a volume).

^ anonymous cvs access
1376 [rich infoeth] I have been trying for last hour to get anonymous cvs acces to ruby (I
1377 [matz ruby-la] It works now.  I've heard that something was wrong with the router.

^ differences between Module and Class ?
1378 [matju sympat] I would like to know all the differences between Module and Class, except
1387 [matz ruby-la] I don't think so.  You've covered as much as I know.
1442 [matju sympat] Thank you very much. Now I have one-half of a future RCR and I'm trying to
+ 1443 [pit capitain] Mathieu, could you tell us what you are trying to achieve? I suppose
| 1448 [matju sympat] 1. unification of concepts that are very close to each other and for which
+ 1444 [decoux moulo] This was me, in ruby-fr
| 1451 [matju sympat] I left to you the chance to stay anonymous. Somehow you manage to reveal
| 1453 [decoux moulo] I always assume what I say.
| 1462 [matju sympat] Pablo Picasso
+ 1452 [matz ruby-la] (2a) breaks inheritance tree.  It's not acceptable.
  1469 [matju sympat] Class#new is moved to Object.new
  1470 [matz ruby-la] No, I didn't understand.  It now makes sense.
  + 1471 [chadfowler c] This is a multi-part message in MIME format.
  | 1474 [matju sympat] Chad, wasn't the module/class distinction a similarly "idealistic" move in
  | 1475 [chadfowler c] This is a multi-part message in MIME format.
  | 1479 [matju sympat] Several RCR I've submitted, and that were accepted, weren't backed by any
  + 1472 [matju sympat] module A; end
    + 1473 [ser germane-] I know that this is just a semantic difference, and lord knows I'm not John
    | 1476 [matju sympat] I don't call this "best practices", I call it "basic practices". The core
    | 1477 [ser germane-] An Array "can" Enumerate, and that's what allows it to expose methods in the
    | 1478 [matju sympat] I see that you have not replied to the part on "good OO design". I would
    + 1480 [matz ruby-la] Of course.  We have different viewpoint, and we don't share definition
      1481 [matju sympat] I thank you very much for your attention.
      1482 [matz ruby-la] I feel very sorry that I cannot explain "good OO design" well.

^ require vs. include
1379 [dinesh_list ] I am new to the language so please excuse me for this simple question.
1380 [matju sympat] require() is like load() except it does it only once per file. load() is

^ proc/block with return
1381 [matju sympat] it seems that the semantics of return inside of a proc/block have changed
+ 1382 [nathaniel ta] See [ruby-talk:77835] and the surrounding thread.
| 1391 [matju sympat] I don't know whether I'm blind, but in that thread I only see small hints
| 1393 [dave pragpro] No it isn't - I'm still struggling to get the updated PickAxe done: the
| 1397 [george.marro] Here's another return/proc/block oddity, though it may just be a side effect
| 1403 [matz ruby-la] It's odd indeed.  Let me check.
+ 1383 [matju sympat] irb(main):001:0> def f(&b) b.call; end; def g; f{return 42}+1; end; g
  + 1385 [matju sympat] Okay, I goofed up here, as I'm quite confused by differences I didn't know
  + 1388 [matz ruby-la] Hmm, I too feel smell of a bug.  Let me consider.
  | 1390 [matju sympat] One first mistake of mine was to assume I recalled how Ruby works wrt
  + 1392 [dblack super] irb(main):001:0> VERSION

^ libs/net/telnet.rb [PATCH]
1384 [ehughes blue] This message is in MIME format. Since your mail reader does not understand
1386 [gsinclair so] I have recently committed telnet.rb, the only change being
1389 [ehughes blue] Cool. That's more detailed than mine (and I like the JavaDoc style, too),
1431 [ehughes blue] This message is in MIME format. Since your mail reader does not understand

^ Std lib and updating PickAxe (was Re: proc/block with return)
1394 [gsinclair so] Out of interest, what do you (and others) consider the "standard library"
+ 1395 [ehughes blue] For what it's worth, that's one of my minor niggles with the pickaxe book
| 1396 [gsinclair so] My FWIW comment is: alphabetical order is solved by the index of the
+ 1398 [ser germane-] Agreed.  A while back, Dave, we talked about getting third-party APIs included
| 1404 [Dave PragPro] It wasn't so much that I was going to include other APIs in ri, but
+ 1399 [dave pragpro] If it comes with Ruby, I consider it standard.
  1401 [ruby jamesbr] Should we then expect to see an ri for 1.8 that includes, for example,

^ subclassing Structs
1400 [Eugene.Scrip] I'm trying to create class which behaves as struct (almost) and has some
1402 [matz ruby-la] That's because Struct class is a abstract factory for each Struct
1405 [Eugene.Scrip] My example didn't show that but I wanted to create struct-class with
1409 [matz ruby-la] Unlike other classes, Struct.new returns its subclass, not instance.
1425 [Eugene.Scrip] OK, thank you for explanations. Problem solved by changing def new2 to

^ _id2ref bug?
1406 [rpav users.s] While debugging some caching code, I've come across a segfault related
1407 [matz ruby-la] We can't say much without error reproducing code.  Can you disclose
+ 1408 [decoux moulo] It has a finalizer and it's in a "freeing page" : this case is possible ?
| 1410 [matz ruby-la] It is very possible.  It should be checked by "RBASIC(ptr)->klass == 0".
| 1411 [rpav users.s] Hrm, it doesn't seem to, I still get a segfault.  I checked, and the ID
+ 1412 [rpav users.s] The main problem is the codebase is pretty big.  I can probably come up
  1414 [matz ruby-la] I don't care about size.  I just care about whether I can reproduce

^ Re: _id2ref bug? (REPRODUCED, short)
1413 [rpav users.s] This is a multi-part message in MIME format.
1415 [matz ruby-la] Thank you.  This is what I need.  The following patch should fix the

^ Re: _id2ref bug? (another break)
1416 [rpav users.s] Thanks... fixes my example great.  Unfortunately it's still a problem in
1417 [nobu.nokada ] BUILTIN_TYPE 63 is T_NODE, which must not appear in Ruby.  Try
+ 1418 [rpav users.s] This fixes it.  Thanks.
+ 1424 [decoux moulo] * he make something strange in his module
  1426 [rpav users.s] The only modules I'm using (in addition to anything builtin) are MySQL
  1427 [decoux moulo] and this means that you have a bug in your source. You can retrieve a
  1428 [nobu.nokada ] Ryan is correct.
  1429 [decoux moulo] not for me. There is a difference between protecting ruby againt coredump
  1430 [nobu.nokada ] I agree that the protection should work.