Packaging .py files

[Tomasz Pala]

On Thu, Jul 17, 2008 at 17:17:21 +0200, Mariusz Mazur wrote:

> I know, my point is, that there are specific cases, where an 'error' is too 
> widespread to try to fix everything and it might make more sense to just stop 
> enforcing our way and do what everybody else does. On the other hand, I'm 
> quite attached to PLD being for example FHS-strict.

Our policy seems to be the winning one - not only other distros try
harder to keep standards, app developers too. Have you ever faced
rejecting some bashizm patch? World wants to be standarized, it's
popular.

> second upgrade. I really do think that such integration nightmares as OO or, 
> dunno, big java apps (especially considering java has it's own standards for 
> allmost everything and you don't gain anything by recompiling bytecode) 
> aren't worth trying to force our ways onto and it makes more sense to make 
> more of an effort to accommodate the stuff that's released by upstream. It's 
> a separate discussion though.

I agree. Because in this case we are 'dumb monkeys' trying to recompile
everything. However it's not /bin/sh case.

> At a certain complexity level it might not just be possible/worth it, to do it 
> The Right Way.

Fixing bashizm is not complex. After all one can change just bang line.
I'm far from making Oracle FHS-compliant.

>> Doesn't our patches go upstream? If they are rejected it usualy means,
>> that authors are really dumb or don't give a shit. Either way we do The
>> Right Thing.
> 
> A) Authors often have different goals then distributions, especially 

Shell scripts are usually beyond any goals, they exist just because they
are handy.

> non-mainstream ones, like PLD. So I'd guess more often then not, they'd be 
> saying we're the idiots.

Some examples of rejecting bashizm patch?

> B) We can't save the world. Having more and more 
> pld-specific patches makes it harder to maintain PLD so in specific cases it 
> might make more sense to just give up and do what everybody else does.

FHS is much more complex than bash/pdksh issues, as well as handling
compressed %doc in internal help browsers.

> I'm in favor of PLD being a compromise between being a geek's dream and 
> something that's actually usable without having to patch your way trough 
> every app.

There's only ca. 30 bash related patches in SOURCES. It's not every app.

-- 
Tomasz Pala <gotar at pld-linux.org>