Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#18215 - [vi] package improvements

Attached to Project: Arch Linux
Opened by Davorin Učakar (phantom) - Sunday, 07 February 2010, 17:10 GMT
Last edited by Paul Mattal (paul) - Sunday, 07 March 2010, 21:03 GMT
Task Type Feature Request
Category Packages: Core
Status Closed
Assigned To Tobias Kieslich (tobias)
Paul Mattal (paul)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

As the maintainer of the ex package that current core/vi is based on, I have a few suggestions:

1. part of du.patch is obsolete; it handled offsetof bug for GCC 4.4.0 on x86_64 but it has been fixed in GCC 4.4.1,
2. du.patch should be renamed to something like navkeys.patch to make more descriptive name (I named du.patch by my initials due to lack of imagination :) ) or may also be removed as it's only intention is to make vi a bit more user-friendly,
3. PKGBUILD may be simplified; most of the sed lines can be replaced with parameters to make.

Look updated ex package for details:
http://aur.archlinux.org/packages.php?ID=26805
This task depends upon

Closed by  Paul Mattal (paul)
Sunday, 07 March 2010, 21:03 GMT
Reason for closing:  Implemented
Additional comments about closing:  These updates are in vi-050325-3.
Comment by Paul Mattal (paul) - Saturday, 06 March 2010, 14:33 GMT
This looks much cleaner to me! Thanks for pointing us to it, Davorin.

Anyone object to my cleaning things up along the lines suggested in the AUR PKGBUILD? If not, I'll do that later today.
Comment by Paul Mattal (paul) - Sunday, 07 March 2010, 18:24 GMT
It looks like your navkeys.patch removes one mod that was in du.patch, to a regexp function. This still seems like a valid patch-- is there some reason we should remove it?

diff -ru ex-050325/libuxre/regnfa.c ex-050325-du/libuxre/regnfa.c
--- ex-050325/libuxre/regnfa.c 2005-02-06 15:16:14.000000000 +0100
+++ ex-050325-du/libuxre/regnfa.c 2009-06-25 21:33:30.000000000 +0200
@@ -535,7 +535,7 @@
* the alignments (given the varying length of rm[])
* are potentially nontrivial.
*/
- n = offsetof(Context, rm) + np->rmlen * sizeof(regmatch_t);
+ n = ((size_t)&cp->rm - (size_t)cp) + np->rmlen * sizeof(regmatch_t);
i = 4;
cpp = &np->avail;
while ((ncp = malloc(n)) != 0)
Comment by Paul Mattal (paul) - Sunday, 07 March 2010, 18:30 GMT
I'm slowly piecing this together. Was this to address a gcc bug? The one referenced in your ChangeLog as:

ex-050325-5
- removed x86_64.patch, since offsetof bug for GCC on x86_64 has been fixed
in GCC 4.4.1.

If so, it sounds like I can remove this now, as it's no longer needed, correct?
Comment by Davorin Učakar (phantom) - Sunday, 07 March 2010, 18:31 GMT
Remove it, it was only a workaround for broken offsetof operator in GCC 4.4.0 on x86_64. It was fixed in GCC 4.4.1.
Comment by Paul Mattal (paul) - Sunday, 07 March 2010, 19:35 GMT
This is now in svn. Will release to testing shortly in vi-050325-3.

Loading...