32370-33220 subjects 32803-33708

rdtool and RDoc
32594 [ptkwt@sh ll ] I haven't tried RDoc yet, but I plan to real soon.

open source projects, sponsoring
32602 [tobiasreif@p] would it make sense to setup a Wiki page for exchange of sponsoring offers?
+ 32604 [ptkwt@sh ll ] I can see that it would be easy to get Open Source developers to
| 32625 [tobiasreif@p] Ideas are welcome :)
| 32659 [ptkwt@sh ll ] Is SourceXchange still in operation?
| 32662 [rich@in oe h] As a former CTO, I can say that one way to convince your company's business
| + 32682 [nhodgson@bi ] Another one to use is "keeping this closed does not contribute to
| + 32683 [nhodgson@bi ] Another one to use is "keeping this closed does not contribute to
+ 32613 [chadfowler@c] offers?
  32678 [ianm@ac iv s] Yep you're thinking of SourceExchange ( sponsored by HP ) and CoSource - run

Anyone in Bangalore, India?
32615 [chadfowler@c] Now that Dave has started this trend....

Array#splice method
32617 [x@ic im nk .] I'm rewriting the examples in "Mastering Algorithms with Perl" (the
+ 32620 [matju@sy pa ] The closest Ruby equivalent to Perl's splice(\@a,$i,$n,$v) function is
| 32693 [x@ic im nk .] Thanks (to everyone) for the great answers. Once again Ruby excels at this
| 33243 [matju@sy pa ] Okay, so my 2nd proposal was buggy. The former, iirc, was better.
+ 32621 [matz@ru y- a] Most of them can be accomplished by other methods; that's why I didn't
| 32630 [tobiasreif@p] 1. This   ^ caused a parse error in irb, it worked without the dot.
| 32632 [matz@ru y- a] Just a typo.  Sorry.
+ 32622 [pcs3@ma lh s] Just from checking perldoc it seems splice is uneccessary in ruby, since

Problem with Ruby Tk
32627 [peter@se an ] I'm having some problems with the following code. When I click on either
32631 [matz@ru y- a] Since Ruby/Tk's "self swap" hack, you can't access instance variables
32635 [peter@se an ] Brilliant, it works now.
32706 [nagai@ai ky ] And, another sample is here. :-)
32713 [peter@se an ] Ooh, very nice, I like it.

segfault without --enable-shared
32645 [paul@at es .] #include <ruby.h>
32647 [decoux@mo lo] I can't reproduce the coredump with 1.6.5
32651 [paul@at es .] That seems to solve the problem.  In my real-life program, I was using

popen3 and buffering
32646 [paul@at es .] loop do
+ 32652 [ysantoso@je ] I'd use gets and puts instead of the more primitive read and write. Using gets and puts, I get the lines as they are being outputted. Ruby 1.6.5.
| 32655 [paul@at es .] This doesn't seem to help.  The problem is not in read/write, but in
+ 32657 [nobu.nokada@] $>.sync = true	# or $defout.sync = true
| 32665 [paul@at es .] So a call to setvbuf ought to solve the problem...
| 32669 [harryo@zi wo] You should be able to use
| 32692 [nobu.nokada@] Sorry, it didn't look obvious.
| 32727 [paul@at es .] [pin, pout, perr, $stdin, $stdout, $stderr].each { |io| io.sync = false }
+ 32738 [paul@at es .] Here's one possible solution.  Does anyone know how to get rid of this
  32785 [decoux@mo lo] What I've not understood is why you want to use a constant (PIPE_BUFFERS)
  + 32786 [nobu.nokada@] If you mean IO instance object's instance variable, I guess he
  | 32789 [decoux@mo lo] Sorry but I don't understand : ruby mark instance variables, where is the
  | 32792 [nobu.nokada@] Sorry, I was stupid.  You're right.
  + 32801 [paul@at es .] I'm not sure I understand why either.  An instance variable makes
    32802 [decoux@mo lo] Something like this, in this case (not tested)
    32806 [paul@at es .] I'm doing something similar to that.
    + 32811 [decoux@mo lo] It depend of your system
    + 32900 [j.travnik@sh] I had similar problem, see

C API/method arguments
32648 [cbgb@u. as i] i hope i'm not being a complete list newbie, but i was wondering if
32649 [decoux@mo lo] Well, you have in README.EXT
32667 [cduck@u. as ] so chalk that up as a newbie not reading the documentation well enough

problems with string interpolation
32653 [ahoward@fs .] i'd like do a gsub such that some characters are escapes for later use
32654 [decoux@mo lo] Well, it's normal you are replacing the string by the original string
32740 [ahoward@fs .] that you
32743 [decoux@mo lo] Well you have 2 steps with this construct
32753 [josb@cn ds .] Slightly off-topic: what's confusing for me personally here is that in a
32784 [decoux@mo lo] Can you give me the name of this P language ?
32867 [josb@cn ds .] Oh I am such an idiot, I take that back. This P language behaves identically.

'make' errors on AIX with Ruby 1.6.6
32656 [tgphelps@ig ] I just downloaded Ruby 1.6.6, and did "configure" and "make" on my AIX
+ 32701 [drosih@rp .e] This is probably the same problem I ran into on one of our (RPI)
+ 32702 [matz@ru y- a] /* -*- C -*-
+ 32846 [kloster@su f] As of 1/30/02, I am having the exact same issue installing on AIX
  32859 [tgphelps@ig ] The fix, from Matz, is to replace ext/digest/defs.h with the following

howto copy files from NT box to NT servers
32660 [Desai.Dinaka] I have NT box and five NT servers on LAN. I would like to copy few files from
32661 [tgphelps@ig ] "copy" is a builtin command of the CMD shell. You probably have to run a
+ 32663 [chris.morris] You might also try these Windows-ized versions of system and `
+ 32690 [nobu.nokada@] Win32API version.
  32721 [chris.morris] Also, ftools.rb (in the distribution) has a copy routine. I haven't tried it

Documenting Fortran 90 source via Rdoc's engine?
32666 [W.L.Kleb@La ] Any chance this could be made to work?  Easily?
32668 [Dave@Pr gm t] How compatible would the intermediate object model be? Right now I
32679 [W.L.Kleb@La ] "Fairly straightforward" I write, expecting something to
32680 [Dave@Pr gm t] Not at all. You see, if someone writes "this is fairly

What is special about TCPServer?
32670 [john.carter@] I have a wee problem inheriting from TCPServer
32687 [r2d2@ac .u u] Because TCPServer implements TCPServer.new.
+ 32695 [nobu.nokada@] It has changed in recent CVS version.  It works in 1.7.
+ 32700 [john.carter@] Aha! Thanks. All I needed to do was override TCPServer.new and invoke

ANN: Automation Server DLL can by used by Ruby on Win32
32672 [mkelly2002NO] I created an InProcess Automation Server .dll wrapping

Test connection InterBase with Apache
32676 [marnaudo@in ] i am testing the right way to connect Interbase via Apache and html
+ 32703 [matz@ru y- a] I don't think so.  Interbase module itself do nothing with Apache.
+ 33099 [marnaudo@in ] I don't think so.  Interbase module itself do nothing with Apache.

32677 [harryo@zi wo] Sorry, people.  There were a bunch of messages from me a couple of

Cardinal - project looking for a leader
32681 [ptkwt@sh ll ] A few weeks back I proposed the Cardinal project - basically a project to
32856 [Erik.Bagfors] From talking to the parrot-people (well, dan anyway), trying parrot and
+ 32864 [alan@di ik t] I'm interested in working on this project, but have a time problem too
+ 32868 [ptkwt@sh ll ] Good point: working with the parrot folks now lets us give them input.

require problem - uninitialized constant Set (NameError)
32684 [dcorbin@im e] I took the ruby code here (http://www.hypermetrics.com/rubyset.html) and
32696 [rich@in oe h] require "set" # note the quotes

(the diff) Re: TypeError verus NoMethodError
32685 [chr_news@gm ] ...
+ 32708 [matz@ru y- a] The point is: when we agree Fixnums and Symbols are NOT singletons,
+ 32836 [chr_news@gm ] this did not make it to my news-server  so I  grabbed it from blade
  32879 [matz@ru y- a] I think I will go for this.

(the diff) Re: TypeError verus NoMethodError
32686 [chr_news@gm ] ...

Weekly RCR Summary
32688 [RubyGarden@t] This is an automatically generated list of Ruby Change Requests.

ruby extensions in C and booleans
32689 [mmaciaszek@g] I was trying to create some simple windows with curses. This
+ 32691 [nobu.nokada@] RTEST(bf)
+ 32710 [rokosm@kl ka] From ruby.h

Open3 (more on  thread)
32694 [lists.ruby-t] I was searching through the ruby-talk.org archive (very helpful!) on
32699 [nobu.nokada@] Like this?
32709 [lists.ruby-t] another script that ended with 'exit 1'.  I then ran this mess like
32715 [nobu.nokada@] Otherwise the script blocks till the child process exits.  If

32705 [Jean-Francoi] does someone knows why I can't get the size of the items displayed on a
32707 [nagai@ai ky ] Am I misunderstanding your question?

Hal Fulton's Set class
32716 [dcorbin@im e] I'm attempting to use this class
+ 32733 [matt@li ke .] See the union! method.
| 32734 [dcorbin@im e] def union!(other)
| 32737 [matt@li ke .] Yep, looks like you have to create a new Set that has the new elements
+ 32778 [hal9000@hy e] Ahh... I really should remove that thing or rewrite it.
  32796 [dcorbin@im e] I might do so, but I'm sure I have less experience with Ruby that you
  32797 [jason@jv eg ] bin/viewcvs.cgi/*checkout*/rubycollections/rubycollections/rbc/set.rb?

FOX - undefined method `repaint'
32717 [vladare@ya o] I receive this message when I start the program under Windows 2000, but
32726 [ljohnson@re ] Could you maybe send me your program (or another that demonstrates the

68k Mac version
32720 [west@io .c m] I seem to recall reading that there was a 68k Macintosh version of Ruby
32723 [hisa@im sy o] First, I praise that you are using the old Mac carefully. But

32722 [Eugene.Scrip] Yes, it is ok. It is my misunderstanding of mod_ruby. I add paths in
32728 [nobu.nokada@] Hmmm, I'm not sure about mod_ruby, but it seems to initialize

mailing list problems
32724 [avdi@ya oo c] I hate to post this to the list at large, but I haven't
32725 [nobu.nokada@] In the header,

[ANN] Log4r 1.0.1 - remote logging and assorted updates
32729 [leon@ug s. a] Added remote logging (with ROMP), log reports via email, and various

FOX newbie help, hooking up button click to method
32731 [chris.morris] Whipped up my first FOX dealy ... and it don't work. It creates the form,
32748 [ljohnson@re ] Yes, I can see what you're doing wrong ;)

[ANN] RbProf 0.2.1
32732 [feldt@ce ch ] RbProf is a profiler included in the AspectR distribution. Readme is

newbie help String#method!
32736 [L4@st ts n. ] ...
+ 32741 [matt@li ke .] You can replace one string value with another with the String#replace
| 32749 [L4@st ts n. ] Yeah, I figured something like gsub!(/(\A|\s)\w/) { $&.upcase }
| 32756 [matt@li ke .] self.[](0..-1, @rs)
+ 32742 [decoux@mo lo] Well, there are many ways. A stupid one is
  32745 [dempsejn@ge ] a quick question: in perl, using $& hits your whole program with a
  32781 [decoux@mo lo] Yes, you don't have this problem in ruby.

do not delete ! get rich - hurry up and act now !
32750 [whynot@ge ri] ...

Learning the APIs
32751 [dcorbin@im e] 9 months ago, when I last tried to find time to mess with Ruby, I seem
+ 32752 [tobiasreif@p] the incredibly awesome RDoc
+ 32754 [armin@ap ro ] I guess you meant ri, written by Dave and Andy.
+ 32755 [alwagner@ua ] the music is nothing if the audience is deaf."

strange behavior
32757 [mengx@ni ls ] Below is a simple program
32760 [matz@ru y- a] * constants defined in the class/module.
+ 32767 [paul@at es .] If a constant is visible in an outer class, is there a good reason for
| 32774 [matz@ru y- a] I don't think so.  It makes constants globally scoped (well, mostly).
| 32820 [mengx@ni ls ] I think the problem here may be how "include" works.
| + 32823 [mengx@ni ls ] An indirectly related topic, A class method can not be called
| | + 32826 [alan@di ik t] How would you differentiate between the following?
| | + 32827 [paul@at es .] class A
| |   + 32829 [mengx@ni ls ] Exactly confusing. I myself never use the same name for both class
| |   | + 32831 [mengx@ni ls ] Introspection can always be played to get things done. What I meant
| |   | | 32842 [matz@ru y- a] Ruby is following old Smalltalk convention.  For me, C++/Java/whatever
| |   | + 32866 [r2d2@ac .u u] Well, there are plenty of examples in the standard library
| |   + 32840 [chr_news@gm ] Why not do this?
| |     32841 [paul@at es .] IMO, a method should not have to know the name of the class in which it
| |     32855 [chr_news@gm ] One problem is that at the ``time of method body declaration''
| |     32857 [paul@at es .] The context is, however, known at the time that the method is called.  I
| |     + 32863 [chr_news@gm ] The problem might be that  ``class_name" a.k.a.  "thismethod'' would have
| |     + 32896 [decoux@mo lo] This not this that you want ?
| |       32901 [mengx@ni ls ] That was actually what I meant.
| |       + 32903 [mengx@ni ls ] Looks like I have to yield since matz is doing this way.
| |       | + 32905 [decoux@mo lo] You have the response in [ruby-talk:32842]
| |       | + 32911 [alan@di ik t] I think that requiring the qualifier is desireable because it
| |       |   32914 [mengx@ni ls ] I agree. Basically, we end up with trading one convenience for another.
| |       |   32915 [mengx@ni ls ] So Ruby does not "follow symlink" when it tries to look up in calling module's
| |       |   32917 [Dave@Pr gm t] Think of 'include' as inserting a reference to (a proxy for) the
| |       + 32906 [matz@ru y- a] That's the point.  Altough you might expect that, it shouldn't.
| |       + 32907 [Dave@Pr gm t] That's a cut-and-paste include. Ruby's is by reference, and is closer
| + 32843 [matz@ru y- a] Have you tried B.constants?  B.constants => []
+ 32919 [paul@at es .] Here's a problem related to the original poster's problem.  When I do
  + 32921 [alan@di ik t] hmm.. include is different from extend.
  + 32931 [nobu.nokada@] You need to use singleton class to access a costant in an
  + 32940 [matz@ru y- a] Just in case you don't get the Alan's point, "extend" extends the
  + 32945 [decoux@mo lo] Write it like this
    32959 [mengx@ni ls ] As Dave mentioned, "include" use a proxy for the included module.
    + 32963 [decoux@mo lo] C don't ask for the constants of B
    + 32964 [pit@ca it in] I didn't thoroughly follow the many posts in this thread, but I think
      32965 [mengx@ni ls ] I meant the implementation of Ruby could be changed to search
      + 32966 [decoux@mo lo] How many scripts do you break ?
      + 32971 [matz@ru y- a] No limitation.  It's intended behavior.  I advise you to try to

[ANN] PerfCtr 0.1.0
32758 [feldt@ce ch ] Small wrapper extension for high-resolution timing on Windows. When PAPI

Ruby and Morphic?
32759 [ljohnson@re ] So I'd like to look into an extension for FXRuby that allows you to (easily)
+ 32763 [pcs3@ma lh s] It's slow, but if you can live with that it's an excellent OO API.
| 32765 [vruz@ru y- s] The original Morphic for Self  runs only on some Mac OSes, right ?
| 32770 [pcs3@ma lh s] The Self 4.1.4 distribution also runs on Solaris (Sparcs only), and OSX
+ 32764 [matju@sy pa ] I've always wished for a _simplified_ Morphic written in Ruby.
| 32798 [ljohnson@re ] Based on your response and some of the others, it's becoming clearer that
| 32805 [ jimm@io co ] True. It's an object scripting system as well. Every Morphic object is
+ 32766 [vruz@ru y- s] Squeak/Smalltalk
+ 32768 [vruz@ru y- s] [massive snip]
+ 32775 [vruz@ru y- s] [massive snip]

RDoc, imagemaps, help
32776 [Dave@Pr gm t] Well, I've managed to get client-side image maps working with RDoc
+ 32777 [pit@ca it in] Thanks, Dave, that's great. Since I've been the first who mentioned
| 32791 [Dave@Pr gm t] Yup - that was it. It seems to work in all the browsers I have now.
| 32845 [W.L.Kleb@La ] Can you commit it?  I'm giving a short (non-networked) Rdoc pitch
| + 32852 [Dave@Pr gm t] It should have been committed (as they say) earlier this morning.
| + 32853 [W.L.Kleb@La ] My error, it was already there...
|   32862 [Dave@Pr gm t] Done.
+ 32782 [djkea2@mu ca] Well I don't have NS to test it but the SHAPE attribute of AREA should

Ruby on AIX
32787 [dazh_ribrogg] Does anybody know where a SMIT-installable (for AIX) image of Ruby
+ 32790 [schneiker@ju] Over a year ago.
| 32819 [moumar@ne co] I successfully compiled Ruby on AIX. If you want i can email you the archive.
+ 32812 [tgphelps@ig ] Ruby 1.6.4 compiles and installs just fine on my AIX 4.3.3 machine, with

32793 [Eugene.Scrip] I can't. We use .htaccess and some VirtualHosts, but RubyAddPath don't work there.

trouble changing mtime on symlinks
32794 [0bz63fz3m1qt] I need to change the time on some symlinks.
32795 [nobu.nokada@] Remove the symlink and link it again.
32800 [0bz63fz3m1qt] This was NOT what i asked for.
32804 [nobu.nokada@] If your system has lutime(2), you can use Kernel#syscall.
32815 [nobu.nokada@] Shell script to make system call module, needs GCC.
32821 [0bz63fz3m1qt] Thanks you are my hero ;-)

$: in mod_ruby
32799 [nobu.nokada@] Then, can you set $: in handler itself?  If not, I have no
32957 [Eugene.Scrip] I had the same idea but there was something wrong with it, I don't