Re: The XZ patch?

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 5, 2009, 18:31 UTC
Message-ID: <7vhbwm5czg.fsf@alter.siamese.dyndns.org>
In-Reply-To: <m3zlaepl8v.fsf@localhost.localdomain>

Jakub Narebski <jnareb@gmail.com> writes:

Show 24 quoted lines

> Andreas Ericsson <ae@op5.se> writes:
>
>> Mark A Rada wrote:
>> >
>> > I was wondering what the fate of my XZ patch was (I didn't do
>> > something dumb, did I?). Never got a response after my last
>> > submit.
>
> I'm sorry, I didn't reply to the last version, did I? I liked both
> patches
>  
>> No comments usually meant noone cared enough about the implemented
>> feature to comment on it. Personally, I'd never use a compression
>> algorithm that hogs as much memory as XZ does. "Good enough" really
>> is just that, imo, and bzip2 and gzip are widely available pretty
>> much everywhere, whereas I've never heard of XZ before.
>
> Well, there were two patches in last series, and I'd rather liked the
> one that decoupled list of _known_ snapshot formats from the list of
> snapshot formats projects are _allowed to use_, when project specific
> override for 'snapshot' feature is turned on.  So for example one can
> allow project specific override (so projects can chose whether to have
> snapshot, and what formats to use) but for example disable chosing
> 'tbz' (bzip2 compression) or 'txz' (XZ / LZMA2 compression) formats.

I kind of liked the generic "disabling" support, even though I was uninterested in adding xz or any other formats. Perhaps it is more useful if the patches were in reverse order. First allow site administrators to selectively disable the current set of formats, with documentation updates for that feature, then finally add xz.