Re: [PATCH] Improve the strtok avoidance fix

[Benjamin Kosnik <bkoz at redhat dot com>]

>This version seems to me clearer, more efficient
>(half the strpbrk calls) and shorter.

Hard to argue with that, thanks to you both.

It looks like Jakub has put an early version of this on the
gcc-3_2-rhl8-branch. Since that branch is merged from gcc-3_2-branch,
I'm suggesting that you put the final version of this on the
gcc-3_2-branch as well. In general, I'm trying not to touch the
gcc-3_2-rhl8-branch (or any other vendor branches of gcc-3_2-branch) and
instead concentrating on getting all the bugfixes on gcc-3_2-branch. I
think everybody is treated fairly this way.

This is the only way I think I can manage these trees and remain sane.

I'm not quite sure how you should finesse this, but I suggest you start
by creating a GNATS bug report about the incident, say it's fixed on
mainline, provide the ChangeLog or patch for gcc-3_2-branch. Then it can
get checked in to the branch and closed.

Thoughts? We've got to be able to move more than strict regressions to
the branch. Some of the obvious bug fixes have to go as well. I think
this method is better than what we'd been doing, and at least documents
stuff in GNATS.

-benjamin