[Teammetrics-discuss] Broken config file

Sukhbir Singh sukhbir.in at gmail.com
Sun Nov 15 02:11:24 UTC 2015


Hi Andreas,

> Uhhhhmmm, that's a shitty reason.  I'm wondering why ConfigParser is
> dealing that badly with this instead of issing a helpful error message.

I agree; there was nothing helpful about the error message.

> On the other hand I have no idea how this broken config file came here.
> There is no calteam and debconf13-localteam was mentioned twice.
>
> I checked the Git repository:
>
> $ git diff HEAD^^^^^^ listinfo.conf
> diff --git a/etc/teammetrics/listinfo.conf b/etc/teammetrics/listinfo.conf
> index 6c09ed4..a88de85 100644
> --- a/etc/teammetrics/listinfo.conf
> +++ b/etc/teammetrics/listinfo.conf
> @@ -15,6 +15,7 @@ lists = debhelper-devel
>          pkg-haskell-maintainers
>          pkg-horde-hackers
>          pkg-java-maintainers
> +        pkg-javascript-devel
>          pkg-kolab-devel
>          pkg-libvirt-maintainers
>          pkg-libvirt-discuss
>
> In other words the file was somehow disrupted since new bytes were added
> but the pkg-javascript-devel list was missing.
>
> This makes me wonder even stronger to skip the concept to move these
> config files to /etc at all.  It was choosen to finally create a package
> from the teammetrics code.  I admit I do not see this happen.  In the
> current situation its always a pain to git pull and to remember that
> you need to manually copy some files to /etc.  What do you think about
> withdrawing this and just use the config in <gitclone>/etc/teammetrics.
> This would be way more consistent.

That's true -- the reason we did this was so that we can have
standardized installations of the config files, but remembering to
manually update is probably not a good idea. So the two ways of
resolving this area: by changing the path to etc/ instead of /etc, or,
by having a git hook that copies the file /etc whenever something is
committed in teammetrics/. Let me know which one you prefer and I can
implement that.



More information about the Teammetrics-discuss mailing list