[pacman-dev] namcap 2.8.1 and namcap 2.99

[Dan McGee]

On Sun, Feb 6, 2011 at 5:56 PM, Rémy Oudompheng
<remyoudompheng at gmail.com> wrote:
> Hello,
>
> I am willing to release a 2.8.1 version of namcap corresponding to the
> namcap-2.x branch of the repository. It includes the following
> changes:
> * revert a dictatorial choice of valid filename characters to include
> all reasonable ASCII characters
> * fix the emptydir rule that did not work
> * the extravars rule was never run (and didn't know options was a
> standard variable)
> * an even bigger test suite.
>
> The rules with no test cases are : depends, elffiles, kdeprograms,
> licensepkg, lotsofdocs, missingbackups, perllocal, rpath,
> scrollkeeper, symlink.
Looks like you didn't push the 2.8 tag to the official repo yet. But
2.8.1 sounds fine.

I've also noticed you are breaking git commit message convention a
bit- you should have a one line summary, then a blank line, then your
message. See http://projects.archlinux.org/namcap.git/commit/?id=eb13cb04c7a2bc3264c63e6fa882633a0a54e773
for why it looks odd doing it the way you have a few times.

> In the other branches (python3 in my repo), I have a version of namcap
> that runs with Python 3 (the bytes vs. string battle was ended
> prematurely). The plans:
> * finish the test suite to have at least a test case for each rule
> * add basic split package support
> * read tarballs in one pass
Where is this branch? I'm not seeing it in either your repo or the
official one. I'd also point out that pushing this to the official
repo would get more people to follow it.

> I'd like to have beta-level intermediate releases (something like
> version 2.99) because the changes made to the code are becoming quite
> large, and it needs many testers to spot regressions.
Just please don't let it turn into 2.99-b-324324-asdfasdf like xz did;
that is just silly. If you have confidence in your test suite, there
is no reason you can't just release it as 3.0 and then have a 3.1
release to fix the problems in that.

-Dan