I just want to name one more restriction that is not well known (and  
perhaps not
followed anymore).  That is that patch file names were not reused.

That is if there are patches patch-aa, patch-ab, patch-ac and then at  
some point
patch-ab is removed and a new patch is to be added it is not added as  
patch-ab
it is added as patch-ad.

I was quite sternly told that this was the correct thing to do though  
I don't recall
if I was ever given an explanation why, I also don't know if all  
package maintainers
follow this rule.

As for the current naming scheme place a tick beside my name that  
says leave
things how they are, I don't see any value in changing the way things  
are done
and I agree that grep is a better solution to the can't find a patch  
for file x problem.

Tyler


On 19/04/2006, at 7:33 PM, Roland Illig wrote:

> Hi,
>
> I'd like to ease the restrictions on the names of the patch files  
> in the patches/ directory. Currently, only the characters [A-Za- 
> z0-9] are allowed after the "patch-". Further, the convention has  
> been established to name the first patch "patch-aa", followed by  
> "patch-ab", and so on.
>
> I haven't found documentation on why the names are so restricted.  
> When looking for a patch for a particular file, I have to look at  
> every patch file to find it. To avoid that, I suggest that the  
> patch file be named after the file that is patched, for example  
> "patch-configure", "patch-foo.c".
>
> For this, the set of allowed characters should be extended to [-A- 
> Za-z0-9_]. This would not allow the name "patch-foo.c", but that  
> one could be transformed to "patch-foo_c" (to avoid having patch  
> files with common filename extensions).
>
> Comments?
>
> Roland