6473-6830

6284-9508 subjects 6721-7089

Patches for 'etc' and 'fcntl' in standard library, to add Rdoc documentation
6473 [meta@po ox c] ...
6480 [matz@ru y- a] Merged.  Thank you.

Program BUGs unless I exit with Kernel.exit!()
6482 [andrew@wa ro] This is 100% repeatable, but a big program, so no 'small snippet of code which

[Patch] ext/openssl/lib/net/https.rb: Remove warnings
6483 [florgro@gm i] ...

[ ruby-Bugs-2746 ] Syslog cannot log messages with %s in them
6497 [noreply@ru y] Bugs item #2746, was opened at 2005-10-29 13:17
6498 [eban@os ri .] % irb

"hot code replacement" debugger functionality
6507 [Markus.Barch] The java interpreter (as an example) allows hot code replacement. If you

characters (and small strings) in ruby 2.0
6508 [eric.mahurin] ...
6510 [mailing-list] So you want to have flyweights for character strings.  It sounds as if
6512 [transfire@gm] Symbols?
6517 [eric.mahurin] ...

Planning FasterCSV (was Re: csv.rb a start on refactoring.)
6511 [james@gr yp ] I've decided to create a FasterCSV library, based on the code we
+ 6513 [zdennis@mk e] I don't like this solution for the long term. It does seem *kludgey* and it seems like we're just
| 6514 [james@gr yp ] To me, it would be that it's a dramatic change.  It make not be 100%
+ 6516 [Ara.T.Howard] require 'csv'
  6518 [nakahiro@sa ] Interesting.  Do you have any idea how API should be?
  6519 [Ara.T.Howard] hmmm.  maybe somthing simple like
  6536 [nakahiro@sa ] Agreed.  What do you think about adding the following snippet to csv.rb?
  6540 [Ara.T.Howard] i think that would adequate.  two comments
  6543 [nakahiro@sa ] Agreed.  Please find the new implementation at
  6545 [Ara.T.Howard] nice.
  6554 [nakahiro@sa ] Sure.
  6556 [james@gr yp ] field1,feild2,"... a whole lot of data that never ends with another
  6564 [nakahiro@sa ] Right.  Hmm.  My csv.rb may need field-size limit to avoid exhausting

[REXML] encoding precedence
6520 [nobuyoshi.na] Currently, REXML prefers iconv module to convert encodings.
+ 6521 [ser@ge ma e-] charset="iso-8859-1"
| + 6523 [matz@ru y- a] Since supporting encodings of iconv varies system to system, using
| + 6533 [nobu.nokada@] The patch gives priority uconv still, but falls back to nkf.
|   6534 [ocean@m2 cc ] +1. This patch won't change anything where iconv or uconv are installed,
+ 6579 [ser@ge ma e-] charset="iso-8859-1"

FileUtils.cp_r - when things go wrong.
6522 [hgs@dm .a .u] ...

What Does rb_obj_call_init do?
6535 [zdennis@mk e] Ok, I am noticing the one difference between proc/lambda and Proc creation, and it's a call to
6537 [nobuyoshi.na] The function just calls #initialize method.
6538 [zdennis@mk e] Ah, sweet. Thanks nobu!

Using LoadLibrary from within an extension
6541 [david@ko nt ] I am having an issue where LoadLibrary works fine on my system in a normal C
6542 [dooby@d1 .k ] This dll must be in your Windows PATH or the current working dir.

-r using Kernel.require instead of rb_require() ?
6544 [sera@fh an .] I asked about this on ruby-talk yesterday: It appears that the -r

1.8.4 p1, warning roundup
6548 [Daniel.Berge] Solaris 10
6549 [hgs@dm .a .u] Which make are you using?  (Sun, GNU, ...)
6550 [Daniel.Berge] Sun.
6551 [hgs@dm .a .u] GNU make may cope better
6557 [hgs@dm .a .u] ...

Socket Documentation
6552 [zdennis@mk e] ...
6575 [gsinclair@gm] I've reviewed this and recommend it be committed.  Can someone please commit it?
6580 [zdennis@mk e] ...
6596 [zdennis@mk e] Not to be picky, as I dont know how long these things take, but has this
6600 [gsinclair@gm] Time to commit is variable.  Docs are regenerated when a few things
6606 [ocean@m2 cc ] Gavin, aren't you a commiter? There is you name in CVS log.
6610 [gsinclair@gm] Yeah, I was.  When the system changed after the cracks, committers
6611 [ocean@m2 cc ] OK, I will.
+ 6612 [gsinclair@gm] Thank you.  And I'll endeavour to regain CVS access ... once my exams
+ 6616 [zdennis@mk e] Thank you very much H. Yamamoto.

[ ruby-Patches-2769 ] Patch for Array#insert
6553 [noreply@ru y] Patches item #2769, was opened at 2005-11-02 21:30
6618 [matz@ru y- a] Strange.  I had no significant performance improvement on my machine

[OT] Re: Planning FasterCSV (was Re: csv.rb a start on refactoring.)
6555 [Ara.T.Howard] 0 points!

[RFC] Method of feeding input to regexp matching
6558 [mailing-list] I would very much like to be able to provide a Regexp object input from
+ 6560 [mailing-list] True.  One solution would be to keep data around as long as its needed.
| 6563 [dblack@wo bl] ...
| 6567 [mailing-list] .* needs more data until there is no more data (#read returns nil), then
| 6568 [dblack@wo bl] ...
| 6571 [mailing-list] Wow, I totally misread this paragraph.  But yes, that”Ēs precisely the
+ 6569 [eric.mahurin] ...

[PATCH] win32/win32.c rb_w32_select patch for polling stdin
6565 [mmiller@hi k] Recently I encountered a problem that I think others have experienced
6566 [nobuyoshi.na] It should have been improved with 1.9, I guess.
6570 [mmiller@hi k] I looked around on the site and may have missed it, but is there a

Stack trace consumes information. patch...
6572 [hgs@dm .a .u] ...
6573 [drbrain@se m] Or you could just rescue the exception and print out the backtrace.
6574 [hgs@dm .a .u] I'd missed that, thanks.  But I'd still like more of it in the
6576 [chneukirchen] Maybe always show a full backtrace with -d or -w given?

Minor array.c doc patch
6577 [djberg96@ya ] ...

[Patch] lib/shellwords.rb
6581 [kevin.jackso] ...
6582 [nobuyoshi.na] You'd intend just `case' without `line' here.
6583 [matz@ru y- a] Could you update?
6584 [nobu.nokada@] New version in [ruby-core:06582] depens on #scan and #[]
6585 [matz@ru y- a] In general, bundled libraries can depends on features available in the

[ ruby-Bugs-2786 ] Pathname.realpath fails on non-existant path
6586 [noreply@ru y] Bugs item #2786, was opened at 2005-11-07 19:21
6587 [akr@m1 n. rg] It is intended because POSIX.

Object#clone missing documentation
6588 [ruby-ml@ma i] It appears that Object#clone, unlike Object#dup, retains
6589 [matz@ru y- a] It says "<code>clone</code> is used to duplicate an object, including
+ 6590 [ruby-ml@ma i] 'Internal state' probably gets interpreted as taintedness, frozen and
+ 6591 [kevin.jackso] ...
  6594 [ruby-ml@ma i] Mm, not quite sure that is clear enough.. if you want to draw a direct
  6595 [halostatue@g] The text should be tighter than that.
  6597 [ruby-ml@ma i] I think clarity should be overriding :) How about the original then?
  6598 [halostatue@g] The problem is that your rewrite added lots of words with little
  6599 [ruby-ml@ma i] My apologies, I thought the section at the bottom was just a quote

Encoding pragma and $KCODE
6592 [nohmad@gm il] pragma each source would have its own. But I could not find the patch
6593 [nobuyoshi.na] Already committed.  But now I noticed pragma for encoding, "coding",

Unpack Endian Bug
6601 [andre@bl et ] Please smack me if this is the wrong list on which to discuss this
6602 [Daniel.Berge] Presumably this is the same issue that was discussed in
6603 [blair@or aw ] Or Fink, which is now at 1.8.3.
6613 [halostatue@g] ...which is not recommended at this point if you want backwards
6614 [blair@or aw ] Does 1.8.4 prelease 1 have a fix for this bug?
+ 6615 [halostatue@g] I don't know. I would suggest that you look at the changelog. I am
+ 6617 [langstefan@g] No, the problem is still there.

Sandboxing without $SAFE
6604 [ruby-core@wh] I've been playing with Ruby sandboxing alot over the past several
+ 6605 [jeremy@bi sw] I wonder whether I could craft foo such that Marshal.load(foo)
| 6607 [ruby-core@wh] from (irb):6:in `load'
+ 6608 [ruby-ml@ma i] In general, this may be unnecessary work. If you run Ruby in a chrooted
| 6609 [ruby-core@wh] Well, I'm trying to boil it down: require 'sandbox'.  No other setup.
+ 6751 [erlercw@gm i] ...
| 6766 [ruby-core@wh] Yes, this is a danger.  I do replace $stderr, $stdout, $stdin.  Your
+ 6752 [erlercw@gm i] ...
+ 6753 [erlercw@gm i] ...

Wildness: Purpose of NOEX_PUBLIC Flag in rb_add_method?
6619 [ruby-core@ha] Several Different references to 'noex'
6620 [mailing-list] NOEX_PUBLIC seems to be the flag for saying that a method is public,
6621 [ruby-core@ha] That's what I thought --- But I wanted to make sure -- I am trying to
6622 [ruby-core@ha] if (SCOPE_TEST(SCOPE_PRIVATE) || node->nd_mid == init) {
6623 [mailing-list] It seems to mean that super can't be called from this method.
6624 [ruby-core@ha] Thanks again --- Seems Simple when you describe it --- I am beginning to

[ ruby-Bugs-2824 ] Array::fill causes segfaults after many calls
6625 [noreply@ru y] Bugs item #2824, was opened at 2005-11-14 23:11
6626 [matz@ru y- a] Fixed.  Thank you.
6627 [andrew@wa ro] Could you expand on the cause a little? Also when the problem was introduced?
6628 [matz@ru y- a] Array#fill sometimes expands the memory region remaining some part of

Strange error messages using DRb/TupleSpace
6629 [drbrain@se m] Using
6630 [james@gr yp ] $ ruby -v

YAML breaks RDOC against head
6632 [ryand-ruby@z] rm -rf rdoc; cvs up -dPA; make ruby; ./ruby -Ilib ./bin/rdoc -Y -o rdoc
+ 6634 [nobuyoshi.na] Install before rdoc or do `make RIDATADIR=rdoc install-rdoc'.
+ 6635 [ocean@m2 cc ] Try this.

alarming changes
6636 [ara.t.howard] is this by design?
+ 6646 [assaph@gm il] ruby 1.8.2 (2004-12-25) [i386-mswin32]
+ 6647 [dblack@wo bl] irb(main):006:0> /(42)/.match(C.new("42")).captures[0].class
  6648 [ara.t.howard] i don't understand how it's supposed to work unless copy constructor semantics

[ ruby-Bugs-2832 ] %x ignores Process.euid with shell metacharacters
6638 [noreply@ru y] Bugs item #2832, was opened at 2005-11-16 14:56
6644 [matz@ru y- a] Is it possible to show us the results for each platform you've tried?

Tuple Class
6639 [transfire@gm] If I put together a good Tuple class for Ruby could it go into core? I
+ 6640 [drbrain@se m] I thought this is what rcrchive.net was for.
| 6642 [ara.t.howard] several of my packages have them internally.  they all use arrayfields and
+ 6641 [dooby@d1 .k ] You can talk /about/ -core on ruby-talk.
  6643 [transfire@gm] Sigh,
  6645 [matju@ar en ] Agreed

Fixnum Limits
6649 [transfire@gm] Will these constants for fixnum become standard?
6829 [transfire@gm] A "NO", "YES", or "IN EFFECT" wiil do.
6830 [mailing-list] At least give us MIN and MAX (for sentinels).

REXML Update Please
6650 [zdennis@mk e] I submitted this as an RCR, but I didn't know that RCR's aren't for the stdlib. Matz commented on
6651 [matz@ru y- a] Usually, ruby-core should be fine.  Unfortunately we haven't seen
6652 [ruby@ja es r] Isn't Sean looking after REXML?
6653 [ser@ge ma e-] charset="iso-8859-1"
+ 6654 [zdennis@mk e] No, there is no RCR. It got closed. The request is my original posting which started this thread. I
| 6658 [ser@ge ma e-] charset="iso-8859-1"
+ 6656 [matz@ru y- a] Oops, sorry, Sean.
  6657 [ser@ge ma e-] charset="iso-8859-1"

Array#insert bogs down under heavy iteration
6655 [djberg96@ya ] It seems that Array#insert bogs down quite heavily
6659 [jim@fr ez .o] Search for Eric Mahurins patch to Ruby. This is probably

Ruby on Neko ?
6660 [ncannasse@mo] I've been working recently on a project called Neko (http://nekovm.org).
+ 6662 [matz@ru y- a] interesting.
| 6663 [transfire@gm] How fast is Neko? Also, can it be used an interpretive mode?
| 6664 [ncannasse@mo] NekoVM is very fast, and it will continue to improve. I'm planing to
+ 6665 [florgro@gm i] Hm, but would it not be easier to convert YARV byte code to Neko byte code?
  6666 [ncannasse@mo] I don't think it would map correctly.
  6667 [chneukirchen] Show how will one map singleton methods, runtime method definitions
  6668 [matju@ar en ] "Objects are some kind of optimized hashtables. All fields names are
  + 6669 [chneukirchen] It is O(0), of course!!1!
  | 6670 [ncannasse@mo] It's actually O(-1), so everytime you're accessing a field you're making
  | + 6671 [ser@ge ma e-] charset="iso-8859-1"
  | + 6676 [matju@ar en ] According to the theory of uncomputing, that requires an computer made of
  + 6674 [shyouhei@ic ] I think that can hardly happen. Have you ever seen a object includes
    6675 [matju@ar en ] I think that you don't understand what I mean. I mean: as long as the

[ ruby-Bugs-2855 ] FileUtils.ln documentation wrong
6661 [noreply@ru y] Bugs item #2855, was opened at 2005-11-19 09:51

[ ruby-Bugs-2858 ] testing for hardlink with "test(?-, ...)" flawed on Windows
6672 [noreply@ru y] Bugs item #2858, was opened at 2005-11-20 16:35
6673 [nobuyoshi.na] I'm not sure if this works on Win9X or not.
6677 [dooby@d1 .k ] It fixes the bug; - I'm not sure if it should do more?
+ 6678 [nobu.nokada@] Shortcuts are neither hard links or symbolic links.
+ 6679 [nobu.nokada@] BTW, may I add a new method File.identical? to 1.8 now?
  + 6680 [ara.t.howard] please do!
  | 6681 [Daniel.Berge] What would it do exactly?
  | 6682 [nobu@ru y- a] Same as test(?-, file1, file2).  Returns true if file1 and
  + 6683 [dooby@d1 .k ] Sorry, I didn't know how to test.
    6690 [nobuyoshi.na] AFAIK, Win9X family don't support NTFS and others, which is
    6691 [djberg96@ya ] raise NotImplementedError unless NTFS
    + 6692 [matju@ar en ] Should I upgrade my PC-DOS 2.00 ?
    | 6706 [hgs@dm .a .u] Apparently I'll get a new PC for work in January, so I can leave
    + 6694 [nobuyoshi.na] Why unless NTFS?  Shouldn't test(?-, f, f) return true on all
    + 6711 [langstefan@g] The FAT file system is still very common. Just think about USB

semenatics of if/unless/while statement modifiers
6684 [skaes@gm .n ] I found that
6685 [ mfp@ac .o g] See [ruby-talk:166828].
6686 [skaes@gm .n ] And how do I get to [ruby-talk:166828]?
+ 6687 [drbrain@se m] Eric Hodel - drbrain@segment7.net - http://segment7.net
| 6689 [skaes@gm .n ] Thanks a lot.
| 6693 [matz@ru y- a] I believe it's a strict application of the "local variables should be
| + 6695 [skaes@gm .n ] It is definitely a compiler issue. Just like the recent discussion of
| | + 6698 [drbrain@se m] f options[:x] if options.include? :x
| | | 6700 [skaes@gm .n ] This doesn't solve the problem of accessing the hash twice and computing
| | | 6701 [shyouhei@ic ] f(options.fetch(:x){raise}) rescue nil
| | | 6702 [skaes@gm .n ] Neither elegant nor efficient. And don't ask me whether it's intuitive
| | | 6709 [ara.t.howard] i use this
| | | 6710 [skaes@gm .n ] I'm afraid I don't understand what you're driving at, although I have
| | | 6713 [ara.t.howard] well that's a start ;-)
| | | 6715 [skaes@gm .n ] ah. this true. nice solution. which I have used in the past too. but
| | | 6727 [matju@ar en ] In Ruby, there are nine different conditionals. All of them are
| | | 6732 [skaes@gm .n ] Only that Perl allows you to assign to $something in
| | + 6699 [matz@ru y- a] Hm, suppose it is a compiler issue, then almost every syntax issues
| | | 6703 [skaes@gm .n ] You don't have to agree.
| | | 6704 [matz@ru y- a] It's complex, but it's not harder to read than it really is, just
| | | 6705 [skaes@gm .n ] I perfectly understand your argument. Anyone writing code like this
| | | 6712 [matz@ru y- a] Because you shoot all who write in that way. ;-)
| | | 6716 [skaes@gm .n ] Just keep it.
| | | 6774 [matz@ru y- a] Hmm, I think you mean we don't have to change the rule except for
| | | 6775 [skaes@gm .n ] Exactly. Although I'm a bit confused by the term RHS identifier. In
| | | 6776 [dooby@d1 .k ] 3.  YACC
| | | 6778 [skaes@gm .n ] Already revoked my statement. For different reasons though. YACC has
| | + 6707 [dblack@wo bl] What would you want to happen here?
| | | 6708 [skaes@gm .n ] That's easy to answer: the same thing that would happen for
| | | 6714 [dblack@wo bl] Ugh.  Well, I normally would avoid the name clash anyway, but ugh :-)
| | | 6717 [skaes@gm .n ] All appreciated. Still you can't decide syntactically whether the = sign
| | | 6736 [dblack@wo bl] if a = 1
| | | 6737 [skaes@gm .n ] But the two forms are not equivalent :-(
| | | 6740 [dblack@wo bl] point ever testing it for truth.  (I purposely chose an example where
| | | + 6748 [skaes@gm .n ] Sorry, I was in a hurry, so missed that a = 1 would, of course, always
| | | + 6754 [ser@ge ma e-] charset="iso-8859-1"
| | |   6756 [dblack@wo bl] irb(main):011:0> def foo(x); if n = x; end; end
| | |   + 6758 [skaes@gm .n ] Exactly.
| | |   | 6759 [dblack@wo bl] We're going around in circles, but the point still doesn't seem to be
| | |   | 6763 [skaes@gm .n ] I missed that one, sorry.
| | |   | 6767 [dblack@wo bl] But how annoying it would be to get warned every time, even when you
| | |   + 6771 [ser@ge ma e-] charset="iso-8859-1"
| | + 6718 [meta@po ox c] I'm really not a fan of
| | | + 6719 [decoux@mo lo] What is a post-test loop ? This
| | | | + 6720 [meta@po ox c] <URL:http://www.rubycentral.com/faq/rubyfaq-6.html>
| | | | | + 6729 [matju@ar en ] Well, pretend you have a real one... and it'll behave like a real one.
| | | | | | 6786 [meta@po ox c] We're discussing language semantics here, 'picky' comes with the
| | | | | + 6735 [dblack@wo bl] There's no "do" in the example; it's begin...end.  The "begin" isn't
| | | | |   6785 [meta@po ox c] It's the 'end' that's extra, not the 'do'. I guess I elided too
| | | | + 6741 [matz@ru y- a] Don't use it please.  I'm regretting this feature, and I'd like to
| | | |   + 6742 [dblack@wo bl] I'm surprised.  What do you regret about it?
| | | |   | 6745 [matz@ru y- a] Because it's hard for users to tell
| | | |   | 6768 [mailing-list] If anything comes out of this discussion (I'm still doubtful), I hope
| | | |   + 6743 [transfire@gm] <amazed> an actual discussion on core!? </amazed>
| | | + 6722 [skaes@gm .n ] No one forces you to use it. But it is already there and probably won't
| | |   + 6723 [ara.t.howard] until you are a native english speaker you won't be able to understand how we
| | |   | 6724 [skaes@gm .n ] you probably meant "unless", because native speaker isn't somthing you
| | |   + 6726 [steven.jenki] It's also valid English, but not what people would typcially say. If the
| | |   | 6755 [ser@ge ma e-] charset="iso-8859-1"
| | |   | + 6757 [dblack@wo bl] Until ready, vamp.  Until golden brown, bake.  'Til you drop, shop.
| | |   | | 6772 [ser@ge ma e-] charset="iso-8859-1"
| | |   | | 6773 [dblack@wo bl] That's my point -- there are a lot of expressions in English that
| | |   | | 6779 [matju@ar en ] Compatibility problems in software are relatively easy to overcome
| | |   | + 6760 [jim@we ri hh] My favorite example of this is Bruce Eckel's comment on Ruby that the
| | |   | | + 6761 [warrenbrown@] Stefan,
| | |   | | | 6764 [skaes@gm .n ] I missed it. mea culpa.
| | |   | | + 6769 [mailing-list] Haha, yes, I saw that too when I read the article :-).
| | |   | |   6770 [dblack@wo bl] If you grew up Pennsylvania Dutch, it also makes sense; end.
| | |   | + 6762 [steven.jenki] Of course they are. I didn't claim otherwise. Either order is
| | |   + 6731 [matju@ar en ] It's valid structure in French too. It's certainly not a germanic-only
| | |   + 6783 [meta@po ox c] The funny thing about programming is that I find languages with
| | |     6790 [matju@ar en ] Aren't rescue-expressions pretty close to "come from"? It wants to "come
| | |     6791 [vjoel@pa h. ] Along with event handlers, callbacks, observers, queue waiters, ... The
| | + 6725 [matju@ar en ] You're not "forced" to do it like that.
| |   6728 [skaes@gm .n ] We had that one covered already. In
| |   6733 [matju@ar en ] What can I say. Sorry for not replying quickly enough.
| |   6734 [skaes@gm .n ] Rightly so.
| |   6738 [matju@ar en ] Why? it would've gotten rid of superfluous keywords.
| |   6749 [skaes@gm .n ] ...
| |   6750 [brent@mb ri ] I've found this aspect of Ruby's otherwise very intuitive
| + 6696 [nohmad@gm il] irb(main):001:0> a
|   6697 [matz@ru y- a] It's to determine which identifiers to be local variables.  You see an
+ 6688 [dblack@wo bl] David
threads.html
top