4748-5698 subjects 5160-5423

[SOURCES] - Hierarchical/Modular Directory Structure
4919 [ilias@la ar ] The source-code structure should be simplified, lowering barriers for
+ 4951 [nobuyoshi.na] Sounds nice, but it would be debatable which is simpler, and if it is
| 4964 [ilias@la ar ] yes, that's the reason for this thread.
| 4975 [nobu.nokada@] My opinion is, current ruby implementation has less files
| + 4979 [chneukirchen] There is no real barrier.  Those who can't/don't want to code,
| | 4983 [ilias@la ar ] People which fail to pass those barriers will fail to contribute
| + 4982 [ilias@la ar ] this is not about adding files.
+ 4989 [ilias@la ar ] "Do it or not" should not be the topic of discussion.
  4990 [halostatue@g] Do it or not is of course the first topic of discussion. Like it or
  4991 [ilias@la ar ] I'm not in the condition to discuss essential & obvious things in this
  4996 [nobu.nokada@] I consider it just unnecessary.
  4997 [ilias@la ar ] It is your right to be an amateur.
  4998 [joaopedrosa@] What are you trying to do? A denial of service attack? :-) That has to
  4999 [ilias@la ar ] [...] (off topic babbling)
  5005 [ehuels@gm il] 1) You don't close a thread: I can and will respond to some of the
  5008 [ilias@la ar ] [movde into context]

PLEASE Stop Ilias Now
4925 [ryand-ruby@z] I'm spawning a new thread for the benefit of those with killfiles.
+ 4939 [batkins57@gm] ...
+ 4944 [ruby@ja es r] Yes, please.
+ 4961 [lists@be tr ] if you do, please let me know through this list because my

[ ruby-Bugs-1901 ] Makefile Error in OpenSSL Extension in 1.8.3 preview1
4931 [noreply@ru y] Bugs item #1901, was opened at 2005-05-12 16:17
4942 [nobuyoshi.na] Sorry, does this help?

ruby-1.8.3 preview1 - failure: cygwin.
4932 [hgs@dm .a .u] compiling bigdecimal
+ 4933 [hgs@dm .a .u] Excuse me replying to myself, but a build in /scratch/hgs works,
+ 4943 [nobuyoshi.na] Sorry, I hope this helps.
  4962 [hgs@dm .a .u] No need to apologise :-)  Thank you for the patch.
  4963 [hgs@dm .a .u] To be extra certain I re-tried the patch on a fresh distribution,

[ ruby-Bugs-1903 ] prefer '#include <errno.h>' over 'extern int errno'
4936 [noreply@ru y] Bugs item #1903, was opened at 2005-05-12 15:50

[ ruby-Bugs-1905 ] Delegate objects recurse infinitely on delegation to self
4940 [noreply@ru y] Bugs item #1905, was opened at 2005-05-13 11:39

[CODE] - Function Naming Conventions
4945 [ilias@la ar ] VALUE

[Fwd: 1.8.3 preview 1 failure - win32/openssl]
4953 [curt@hi bs c] Reply-To: ruby-talk@ruby-lang.org

[Fwd: 1.8.3 preview 1 failure - win32/openssl]
4955 [curt@hi bs c] Reply-To: curt@hibbs.com
4957 [ocean@m2 cc ] Please look at [ruby-core:4942].
4958 [curt@hi bs c] I apologize for the duplication. Thank you for letting me know.

Re: [SORES] - Hieroglyphical/Nodular Olfactory Fixture
4966 [matju@ar en ] ...
4969 [ilias@la ar ] [...] - (faulty bakery analogy)
4971 [ruby-lists@l] please mark all your messages [FLAME] so i can easily block you
4987 [ilias@la ar ] please get serious.

[FILE] - /todo, please update
4967 [ilias@la ar ] the "todo" file looks outdated (13 months).

[ ruby-Patches-1909 ] -Wall -W patch: bignum.c
4968 [noreply@ru y] Patches item #1909, was opened at 2005-05-13 11:19

Question about File.fnmatch
4970 [djberge@qw s] Maybe this is a silly question, but why is File.fnmatch defined in dir.c
4976 [nobu.nokada@] It is derived from Dir.glob and still shares the internal

[PATCH] RDoc consistency cleanup
4972 [tilman@co e-] ...

[PLAN] - Different Development Paths, where?
4974 [ilias@la ar ] this development-forum (http://ruby-lang.org/en/20020104.html)

[PATCH] Win32: Ruby & APR; build problems for Ruby Subversion SWIG bindings
4977 [ehuels@gm il] ...
4978 [nobu.nokada@] I think AC_TYPE_UID_T and AC_TYPE_PID_T also would be needed to
4980 [ehuels@gm il] ...
4981 [nobu.nokada@] Sorry, AC_TYPE_PID_T was inserted locally, but AC_TYPE_SIZE_T
4984 [ehuels@gm il] Not with APR, but probably with other software which defines and uses
4985 [nobu.nokada@] Thank you, could you?
4986 [ehuels@gm il] Consider it done.

[ ruby-Bugs-1917 ] Pathname#unlink fails on a symlink which points to a directory.
4992 [noreply@ru y] Bugs item #1917, was opened at 2005-05-14 21:33
+ 4993 [nobu.nokada@] The result of File.lstat can do it at a time.
+ 5000 [akr@m1 n. rg] I'll fix it as follows.
  5001 [ocean@m2 cc ] This code seems to work on Linux only.
  5002 [akr@m1 n. rg] Oops.
  5003 [ocean@m2 cc ] Yes, but Windows (bcc32 and mswin32) behaves differently....
  + 5004 [ocean@m2 cc ] OK, it's now working on Windows with following patch.
  + 5007 [akr@m1 n. rg] I see.

REXML::Encoding (Re: ruby 1.8.3 preview1)
4995 [speakillof@y] Is it possible to port changes of REXML::Encoding to 1.8.3?

Suggestion for avoiding incivilities
5006 [gsinclair@so] Recently there have been some questions and statements on this list
+ 5009 [ilias@la ar ] Even the most essential questions remain unanswered within a few threads.
+ 5025 [discordantus] I second this motion.
| 5030 [ruby-lists@l] /me nods sullenly and goes with it
+ 5029 [vjoel@pa h. ] On behalf of the silent majority, thanks, Gavin. (Oops, there goes my

[CODE] - Function Argument Conventions
5010 [ilias@la ar ] As a general rule, the first argument to a function should be the
+ 5011 [chneukirchen] We are eagerly awaiting your patches.
| 5012 [ilias@la ar ] I've no problem to provide patches, but before I provide patches, I like
+ 5015 [nobu.nokada@] More concretely.
| 5016 [ilias@la ar ] (object.c)
| 5018 [tilman@co e-] You are absolutely right!
+ 5017 [nobuyoshi.na] Could you show us the grounds your "general rule" is based on?  I

[FILE] - /class.c - seperate "module" specific code
5013 [ilias@la ar ] class specific code

Forwardable, reworked
5014 [djberg96@ya ] ...
5026 [florgro@gm i] Nice work! I wonder if you could also add a nicer interface while you
+ 5027 [Daniel.Berge] That *does* look nicer.  We should keep the older methods around for backwards compatability, though, unless folks don't care.  I don't get the impression that many people are using this. :)
+ 5028 [robert.mcgov] That is very sweet, if possible I'd like to see that style added.
  5032 [djberg96@ya ] def delegate(hash)
  5034 [drbrain@se m] delegate :@number, :to_i
  + 5035 [Daniel.Berge] It's six one way and half a dozen the other.  Both look good to me,
  | 5036 [drbrain@se m] Agreed, I only felt strongly enough to mention it here.
  + 5037 [gsinclair@so] The arrows are certainly valuable to me in understanding what's going

Problems with the "outer scope" operator
5020 [mailinglists] i don't know whats the name of the "::" operator in a "::module"
+ 5021 [ruby-ml@ma i] Conceivably such a class/declaration might be inside a module,
+ 5022 [hart.martin@] class Outer
  5023 [mailinglists] Okay thats a good example that i understand.
  + 5024 [evanwebb@gm ] People (myself included) put :: at the beginning to be absolutely sure
  + 5031 [discordantus] module One; module Two; module Three; end end end

[ ruby-Patches-1923 ] Forwardable, reworked
5033 [noreply@ru y] Patches item #1923, was opened at 2005-05-17 07:50

: [BUG] unknown node type 0
5038 [juergen.mang] ...

[ ruby-Bugs-1930 ] CGI::escapeHTML escapes the ampersand in an existing escape command
5039 [noreply@ru y] Bugs item #1930, was opened at 2005-05-19 11:45
5041 [nobuyoshi.na] It is exactly expected behavior.
5044 [rubyzbibd@ub] OK.  It would be nice if that was in the documentation somewhere, so
+ 5045 [nobuyoshi.na] $ ri CGI.escapeHTML | cat
| 5046 [rubyzbibd@ub] Konbanwa,
| + 5048 [rubyzbibd@ub] I might have spoken too soon.
| + 5061 [nobuyoshi.na] Rather I'm afread about &lt; and &gt; especialy.  They may lead
+ 5054 [tom@in oe he] Not that I know of, although that would be nice.

[ ruby-Bugs-1931 ] Suggestion to add alias for underscore style method names to camel case named class methods in CGI
5040 [noreply@ru y] Bugs item #1931, was opened at 2005-05-19 11:46
5047 [dblack@wo bl] The place for suggestions for non-bug changes to Ruby is RCRchive
+ 5049 [halostatue@g] irb(main):001:0> require 'cgi'
+ 5050 [rubyzbibd@ub] I always thought the RCR were more language design proposals as
  + 5051 [djberg96@ya ] Well, really it's a feature request, but since they
  + 5052 [dblack@wo bl] Sorry -- I think I read your suggestion backwards!  If you're
    5059 [rubyzbibd@ub] Yes, I want underscore names.  I proposed making an alias so that

[ ruby-Bugs-1932 ] Regexp#+@ can't be overridden
5042 [noreply@ru y] Bugs item #1932, was opened at 2005-05-18 21:19
5043 [pit@ca it in] Seems to be a parser problem. You have to put parentheses around a regex

[PATCH] Silly bug in 1.8's YAML
5055 [tilman@co e-] ...
5166 [matz@ru y- a] Thank you.  I will merge this.  By the way, I think you need to
5171 [tilman@co e-] I didn't notice this before. Yes, it looks like the same change should

SSL patch for preview 3
5057 [evanwebb@gm ] Could the patch attached to [ruby-core:4663] be merged into 1.8 and
5062 [gotoyuzo@no ] Sorry, I overlooked that mail.
5063 [evanwebb@gm ] This is fine, just as long as the use of random session_id is the

[ ruby-Patches-1934 ] Add underscore method name aliases to the CGI class methods that currently have a camel case style method name.
5058 [noreply@ru y] Patches item #1934, was opened at 2005-05-20 11:46

[ ruby-Patches-1939 ] Pathname, totally revamped
5065 [noreply@ru y] Patches item #1939, was opened at 2005-05-21 12:34
5067 [nobu.nokada@] Where is it?
5068 [djberg96@ya ] module Facade
+ 5069 [dooby@d1 .k ] Related - [ruby-core:3919], short enough to quote in full ...
| 5070 [djberg96@ya ] #---------------------------------------------------------------------------
| + 5072 [discordantus] I think the point that is being made here is that it modifies the
| | 5073 [djberg96@ya ] How would one end up with "/path/to/libdir" on Win32?
| + 5074 [akr@m1 n. rg] How different?
| | 5075 [Daniel.Berge] assert_equal('../../d',
| | + 5077 [nobu.nokada@] It differs from the former two.  I can't see your reason.
| | + 5079 [sroberts@un ] But I don't even know, and can't...
| | | 5080 [matju@ar en ] ...
| | | 5120 [sroberts@un ] Uhm, well, IMHO, thats so abstract its hard to know if its accurate,
| | | 5121 [matju@ar en ] ...
| | + 5094 [akr@m1 n. rg] The current pathname.rb returns '../../d' too, because the dots left
| + 5083 [gsinclair@so] I don't like Pathame inheriting String.  It just doesn't make sense to
|   5096 [akr@m1 n. rg] Currently I think to_str should be used only for a subtype of String.
+ 5071 [nobu.nokada@] Other DOSISH platforms, djgpp, OS/2 and human68k.
  5076 [halostatue@g] On 5/22/05, nobu.nokada@softhome.net <nobu.nokada@softhome.net>
  5078 [nobu.nokada@] Do you mean that open('//server') should try to access
  5082 [halostatue@g] No. That's the bizarre part. Simply doing //server will result in an
  5084 [nobuyoshi.na] Thank you, understood.  How should we deal with that invalid path?  In
  5085 [halostatue@g] I donno. EINVAL is thrown if I open a file in my root directory with
  + 5087 [dooby@d1 .k ] (... unless someone else bumps it ;)
  + 5089 [nobuyoshi.na] Do you mean Dir.glob("//server/*")?  opendir() is using FindFirstFile

[ ruby-Bugs-1940 ] Fixnum#** slower than Bignum#**
5066 [noreply@ru y] Bugs item #1940, was opened at 2005-05-22 00:15
5086 [matz@ru y- a] We have been too lazy to calculate fixnum ** float case directly,

Liskov Substitution Principle (was: Pathname)
5081 [matju@ar en ] ...

select() on non-sockets in eval.c
5088 [BArmstrong@d] In eval.c, select is used to synchronize IO in ruby threads, but on some
5090 [nobuyoshi.na] It is replaced within win32/win32.[ch].
5091 [nobuyoshi.na] Like the following patch, though I haven't even done compilation.
5093 [BArmstrong@d] Ah!  Thanks.  I'll have a look at that code.
5095 [BArmstrong@d] OK.  Incidentally, VMS still lacks GNU configure, so we have to simulate
5113 [BArmstrong@d] So we now have socket fds in rp, wp & ep.
5114 [nobu.nokada@] Correct.
5124 [BArmstrong@d] Nobu,

FTP, mostly questions, but a patch, too.
5092 [hgs@dm .a .u] I was looking at RFC959 -- the FTP spec, because I wondereed if it
5097 [warrenbrown@] Hugh,
5100 [hgs@dm .a .u] OK, I'll have a look, but what I saw about those (on rfc-editor.org)

shim for select() for VMS & general porting questions
5098 [BArmstrong@d] The DECC man page for select() says that I need to #include <time.h>.
5101 [nobu.nokada@] missing.h included by ruby.h should define it.
5102 [nobuyoshi.na] Moved all to vmsruby_private.[ch].  Does this work?
5103 [BArmstrong@d] Nobu,
+ 5105 [BArmstrong@d] Nobu,
+ 5109 [nobuyoshi.na] Indeed.
  5125 [BArmstrong@d] I have yet to try all of these out.  I'm still working from Masamichi's

Re: Trouble compiling ruby 1.8.2 on Intel ia64
5099 [sam.overdorf] I found the following info on a mail thread.

will callable objects be more general in Ruby 1.9?
5104 [eric_mahurin] In Ruby 1.9 CVS, you can "call" local variables. i.e.
5106 [matz@ru y- a] In fact, I am thinking reverting this "call" change.  It causes code
5107 [eric_mahurin] I was kind of looking forward to the further unification of
+ 5108 [matz@ru y- a] I'm afraid it's just like making Common Lisp into Scheme.
| 5115 [eric_mahurin] I would assume it would be workable because you already have to
| 5116 [matz@ru y- a] But you can distinguish local variables and method invocations from
| 5117 [eric_mahurin] I like the Scheme/Python way, but this would cause too many
+ 5111 [pit@ca it in] Please don't. Otherwise, adding a method to a class could break
  5112 [halostatue@g] Concur. I had a related situation while developing and debugging

Shift-JIS encoded strings and YAML.load bug
5110 [rubyzbibd@ub] Some of my colleagues discovered a problem loading YAML with

[ ruby-Patches-1967 ] win32ole.c
5122 [noreply@ru y] Patches item #1967, was opened at 2005-05-30 00:34

[ ruby-Bugs-1972 ] CGI::Cookie::parse does not properly parse set-cookie headers
5126 [noreply@ru y] Bugs item #1972, was opened at 2005-05-30 11:24

XMLRPC charset
5127 [zn@mb .n ft ] XMLRPC library's default charset is us-ascii,

ext/dl/dl.c compile time
5136 [ryand-ruby@z] 501 % gcc --version
5137 [ryand-ruby@z] Oops. Left out an important detail, this is only on HEAD, not ruby_1_8.
5139 [chneukirchen] Note that dl.c includes the huge file (>32000 loc) callback.h... so I

[ ruby-Bugs-1988 ] objects never get garbage collected when requested via ObjectSpace._id2ref
5138 [noreply@ru y] Bugs item #1988, was opened at 2005-06-03 13:02
5140 [matz@ru y- a] _id2ref does no caching. There's no guarantee that any object to be

Where to put non-GNU build system components
5141 [BArmstrong@d] VMS doesn't have a complete GNU environment yet.  (See gnv.sf.net for
5142 [nobu.nokada@] I vote this plan.  ext/extmk.rb and lib/mkmf.rb already do it
5146 [BArmstrong@d] They look good, so long as the generated makefiles are more GNUish than

Win32API segfault in 1.8.3p1
5143 [ntalbott@gm ] I'm on Windows XP, using VC7 to compile. I've previously gotten a good
5144 [nobu.nokada@] Is the line [148,0,0,0,0,swCSDVersion].pack("LLLLLa128") ?
5145 [ntalbott@gm ] GetVersionEx.call(OSVERSIONINFO)
5149 [nobuyoshi.na] I could reproduced it using the version compiled with VC++ 7.1.
5150 [ntalbott@gm ] Thanks, Nobu, the patch worked great.

fcntls only partially supported on non-Unix platform
5147 [BArmstrong@d] if defined?(Fcntl::FD_CLOEXEC)

[PATCH] thread.rb documentation
5148 [drbrain@se m] ...
5162 [matz@ru y- a] I will merge the patch.  Thank you.

COPY and INSTALL on Windows
5151 [ntalbott@gm ] 1.8.3p1 has changed the defaults for the COPY and INSTALL Makefile
5156 [dooby@d1 .k ] I thought it was just bccwin that had this problem because Borland
5157 [nobu.nokada@] INSTALL_{PROG,DATA} have been fixed already in CVS, and "/"
5158 [dooby@d1 .k ] Bizarre, but true.
5159 [dooby@d1 .k ] ... but the first would be of *more* use than the second ...

[PATCH] 1.8.3 p1 segfault in array.c- bccwin32 - bcc5.5 (free) compiler bug
5152 [dooby@d1 .k ] There's a line in array.c which confuses the Borland 5.5 compiler
+ 5153 [Daniel.Berge] I don't know assembler very well, so forgive my ignorance, but does it
| 5155 [dooby@d1 .k ] ... does it help to add parens?
+ 5154 [matju@ar en ] ...
+ 5286 [dooby@d1 .k ] I discussed this on the Borland forum[*] and it was confirmed as an
  5370 [ocean@m2 cc ] Umm, this is difficult to answer...
  5385 [dooby@d1 .k ] Hmm - and that's for C++Builder 6.0
  5394 [dooby@d1 .k ] * array.c (sort_2): get rid of yet another bcc's bug.