| 1 |
Hi! |
| 2 |
|
| 3 |
On Tue, 23 Apr 2013, Ulrich Mueller wrote: |
| 4 |
> > I appericiate the work done by Tobias utmost too but I have to agree |
| 5 |
> > this is not something I want to see running automatically, or even |
| 6 |
> > from within ebuilds. |
| 7 |
> |
| 8 |
> +1 |
| 9 |
> |
| 10 |
> In Tobias's list, I count some 80 packages that need fixing. That's |
| 11 |
> way too few to merit a general solution. Also this number will |
| 12 |
> decrease when files are fixed upstream. |
| 13 |
|
| 14 |
I see two problems with this approach: |
| 15 |
|
| 16 |
- What do we do with unresponsive-yet-not-dead upstreams? |
| 17 |
- What about future packages that ship broken files? I mean not |
| 18 |
just existing packages that keep issuing broken PNGs but also |
| 19 |
future packages that we just didn't cover now? |
| 20 |
|
| 21 |
The former has two and a half solutions: |
| 22 |
|
| 23 |
- Wait until itmagically fixes itself or upstream comes around. |
| 24 |
This is only 1/2 a solution, IMO. |
| 25 |
|
| 26 |
- Add a separate tarball or the like that the Gentoo maintainer |
| 27 |
generates from the broken PNGs. Use this tarball to overwrite |
| 28 |
the broken results of equivalent_of("make install"). |
| 29 |
|
| 30 |
- Have a convenience function that can be used for known-bad |
| 31 |
packages to fix broken IDATs. Basically calling my script or |
| 32 |
the binary Samuli mentioned. |
| 33 |
|
| 34 |
The second problem, however, is trickier. We can rely on people |
| 35 |
noticing the error messages/broken packages and hope they file |
| 36 |
bugs. The other option is to have a QA-like check for it, again |
| 37 |
using the simplest possible binary to do so. |
| 38 |
|
| 39 |
Regards, |
| 40 |
Tobias |
Archives