public inbox for insight@sourceware.org
 help / color / mirror / Atom feed
From: Syd Polk <spolk@redhat.com>
To: "Eray 'exa' Ozkural" <erayo@cs.bilkent.edu.tr>,
	insight@sources.redhat.com
Subject: Re: Packaging insight for debian
Date: Mon, 09 Oct 2000 10:49:00 -0000	[thread overview]
Message-ID: <4.2.0.58.20001009104557.00a306c0@pop.cygnus.com> (raw)
In-Reply-To: <20001009203218.A9794@borg.cs.bilkent.edu.tr>

At 08:32 PM 10/9/00 +0300, Eray 'exa' Ozkural wrote:
>Syd Polk wrote:
> >
> > Factoring out the internal versions of tcl/tk is not a good idea. We have
>local
> > changes to these which could cause random failures, particularly when
>running
> > the test suite.
> >
>
>Do you think making major changes to a forked tcl/tk is a good idea?

No, it is not a good idea. Unfortunately, it has been this way for quite 
some time, and requires manpower to fix. We are working on fixing it, but 
it is slow going.

>Also, this didn't seem to play very nicely with the debian way of
>doing things. Which of course goes something like "if you use a
>general-purpose lib or lang. make it available in a generally usable
>and uniform way" This is how we cope with perl or tcl/tk. What would
>debian be like if every perl application distributed its own perl with it?

The problem is when somebody upgrades the tcl/tk or perl version and breaks 
your application.

>A suggestion thus put forth was to put the cygnus versions of those
>libs in some place like /usr/lib/insight, but it didn't prove to
>be easier than factoring out the libs which seemed to be the right-er
>way of doing things.

While being "right" in the long term, we cannot support it because it has 
not been tested.

>Anyhow, cygnus's changes to tcl/tk are not major extensions or anything
>really new. Otherwise, I wouldn't be able to port it to tcl/tk 8.3
>It now works, and is probably bug-compatible with the pristine
>cygnus (red hat) version. And if you'd tell me that you might be
>introducing something major in the future, you'd better introduce it
>into the upstream version of tcl/tk. ;)

The main thing is that tcl/tk 8.3 does not build for cygwin in a cross 
environment. While this is not really a concern for debian, it is a concern 
for us.

Plus, regular expressions changed between tcl 8.1 and 8.2, and that could 
break some of the code in interesting ways.

Have you run "make check" from the gdb build directory? You should compare 
the results with our version of tcl/tk and with your version; they should 
be the same.

>I'd be glad if you could give the package I made a try. Most of the
>cygnus changes are build-related, and not very relevant to tcl/tk itself.
>(or any of its extensions that insight uses)
>
>I didn't just strip insight:
><<
>orion:debian$ cat strip-insight
>#! /bin/bash
>
>cd insight-5.0
>
>rm -rf opcodes libiberty opcodes etc texinfo tcl tk itcl itk tix
><<
>
>I left readline, intl, and bfd in there. Was a bit hesitant about
>bfd actually but then you can tell me whether this is the same thing
>as in binutils-2.10.0.26 :) Is it? I've managed to strip bfd too in
>an experimental package, so if they're the same version I can safely
>let it be that way.

The bfd that ships with insight is more current than binutils-2.10.0.26. 
However, this is more cygwin stuff. You may not share.


> > Shipping the two versions. The next version of SN will have the same
>version as
> > Insight, and will have been tested correctly.
>
>That will be a bit difficult, but I'll try to do that. Are they
>separate branches or just different versions of the same thing, and
>any major differences between the two?

Actually, I lied. Insight is based on tcl 8.0.5, and sn is based on tcl 
8.1. We will have the entire system upgraded to 8.3.2 at some point in the 
future. It is also our goal to use whatever version of Tcl/Tk is installed; 
however, this is some months away.

>Insight is a very nice debugger [interface], and I'd like to thank all
>developers from cygnus and red hat; we just want to have a proper
>debian version. :)
>
>Looking forward to your remarks,
>
>Thanks,
>
>
>PS: By the way, I've understood how difficult it might be to debug a debugger.
>Kudos for all the hard work!
>
>PS2: I'm having to mail this from a remote machine because my ISP's
>mail server is on the ORBS. Is the only way to talk to the mail admin
>at my ISP?
>
>--
>Eray (exa) Ozkural
>Comp. Sci. Dept., Bilkent University, Ankara
>e-mail: erayo@cs.bilkent.edu.tr
>www: http://www.cs.bilkent.edu.tr/~erayo

Syd Polk		spolk@redhat.com
Engineering Manager	+1 415 777 9810 x 241
Red Hat, Inc.



  parent reply	other threads:[~2000-10-09 10:49 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2000-10-09 10:26 Eray 'exa' Ozkural
2000-10-09 10:47 ` Chris Faylor
2000-10-09 10:49 ` Syd Polk [this message]
  -- strict thread matches above, loose matches on Subject: below --
2000-10-08  9:05 Eray 'exa' Ozkural
2000-10-09  8:35 ` Syd Polk

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4.2.0.58.20001009104557.00a306c0@pop.cygnus.com \
    --to=spolk@redhat.com \
    --cc=erayo@cs.bilkent.edu.tr \
    --cc=insight@sources.redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).