Re: Minified files and source code requirement
Raphael Hertzog, 2011-10-27 09:08:37 +0200 :
>> I think this is exactly the same as xserver-xorg-video-nv, which
>> contained obfuscated C code instead of the actual source code. I
>> personally considered that a DFSG violation but I guess you would not?
> I would consider it a DFSG violation. It's all a matter of intent.
There's a saying about the road to hell.
> Obfuscated != minified.
Intent is all very well, but if the effects of the operation make the
resulting "code" unusable, even for the best of reasons, then said code
can't be said to be the source.
> In one case we have unreadable C code where you don't have access to
> the unobfuscated version and this was done on purpose by the upstream
> authors to make it difficul to understand what the code does.
> easily find the corresponding source code on numerous places
> (sometimes in the corresponding Debian package, sometimes in an older
> version of it on snapshot.debian.org and generally on its upstream
> website). The minification was not even done by the upstream that uses
> that file but by the original project who delivers both files. The
> minification is not done to make it more difficult to understand the
> code or make changes, but to save time when downloading the file.
Convenience is good, but actually providing source is a requirement
for Debian. Unusable source isn't enough.
> Requiring the non-minified file to be provided in the same source
> package is not a very productive use of our time.
Right. In the same way that providing the source for our binaries
isn't very productive, I guess, because who's going to use it when they
have the pre-built binaries? I know this is an exaggeration, but
there's no substantial difference between the two cases.
> In particular if you don't go to the step beyond which is to modify
> the upstream build process to regenerate the minified file from the
> original one. Otherwise modifying that file in the source and
> rebuilding it does not have the expected result.
> I think it's great to encourage this sane behaviour, but it's not a
> bug that's worth a serious severity.
With no particular hat on (except the one of a maintainer of a webapp
that uses a few JS libraries), I respectfully disagree.
>> What is the preferred form for modification for a work (aka source)
>> is highly context-dependent.
> I share entirely the opinion of Russ who replied to this specific point.
> We should not mix the minimal requirements that we have and our own
> personal ideals in terms of what's needed to modify in sane conditions
> the stuff what we are releasing.
I thought that the minimal requirements were precisely designed to
reflect what's needed to do modifications in a sane way, but maybe
that's just me.
The price of liberty is eternal housekeeping.
-- in Odds and Gods (Tom Holt)