[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Status of install-info
From: |
Ian Zimmerman |
Subject: |
Re: Status of install-info |
Date: |
18 May 2007 20:04:25 -0400 |
User-agent: |
Gnus/5.09 (Gnus v5.9.0) Emacs/21.4 |
Ian> Well, I don't know if anyone is still interested in this since
Ian> there was zero response to my previous mails, but here's a python
Ian> version. It actually runs a little bit faster than the perl one.
Nicolas> There are still interests in your script.
Nicolas> I sent a proposal in April (see the thread
Nicolas> http://lists.debian.org/debian-dpkg/2007/04/msg00031.html). You
Nicolas> were not in copy since I think you are subscribed to the Dpkg's
Nicolas> mailing list.
Wrong.
Nicolas> I planed to rely on your script for update-info-dir.
Nicolas> I just had a quick look at you script. You are no more using
Nicolas> ginstall-info (nor install-info). Was it really needed? (for
Nicolas> performance issues?)
Ah, I should have explained better. What I personally do is dpkg-divert
the official install-info script with this:
#! /bin/sh
exec /usr/local/bin/generate-info-dir \
--path=/usr/share/info --path=/usr/local/share/info \
--ignore='GNU ' --ignore='Gnu ' --ignore='Software ' \
--substitute='Math|Mathematics' \
--keyword=Programming --keyword=programming --keyword=Text
In other words, I ignore the original arguments and forward to the script
you have seen.
Nicolas> I'm not sure it would be fine to add another install-info
Nicolas> implementation.
Nicolas> IIRC, you had a implementation based on the GNU install-info
Nicolas> code. Were there some problems with this implementation?
Yes, I have sent a patch for the GNU install-info (C implemented) to Karl.
There was no reply from him or other GNU people either.
The Perl and Python versions are just a demonstration of how much cleaner
and less bug prone this could be if this strategy (regenerate the index
after each install/deinstall) were adopted exclusively.
Well, for me there are more than a demonstration, since I have used them in
"production", and I'll likely stick with the Python version even if the
patch is accepted.
Best, Ian.
--
This line is completely ham.