Re: wip/mk/git-package.mk cleanups

[Aleksej Saushev <asau%inbox.ru@localhost>]

Richard Hansen <rhansen%bbn.com@localhost> writes:

> On 2014-05-06 08:30, Thomas Klausner wrote:
>> I have one question and a stylistic comment.
>> 
>> On Mon, May 05, 2014 at 07:54:17PM -0400, Richard Hansen wrote:
>>>   * create NO_CVS_CACHE, NO_GIT_CACHE, and NO_HG_CACHE for consistency
>>>     with NO_SVN_CACHE (so users can easily disable caching)
>>>
>>>   * create NO_{CVS,GIT,HG,SVN}_CACHE.${id} so that packages can
>>>     disable caching
>> 
>> We usually avoid negative variable names, if possible, since e.g.
>> NO_CVS_CACHE=no is a bit hard to understand. Can you make these
>> positive? Perhaps CREATE_HG_CACHE or CACHE_HG or something similar?
>
> I actually added positively named variables to cvs-, git-, and
> hg-package.mk before I noticed that svn-package.mk already had
> NO_SVN_CACHE.  Are you OK with breaking compatibility by renaming
> NO_SVN_CACHE?  If so, I'd be happy to rename everything to
> CACHE_{CVS,GIT,HG,SVN} (I prefer positive variable names too).

I'm fine with breaking compatibility as long as there's clear way to adapt.
This is WIP anyway.

> Alternatively I can support both CACHE_SVN and NO_SVN_CACHE and add a
> deprecation warning if defined(NO_SVN_CACHE).

I'd say, no, don't do that.

My opinion is that we can discuss inverting the logic a bit later.

>>> Subject: [PATCH 05/13] use :Q modifier to escape special shell characters
>> 
>> Did this actually fix anything or is it just for safety?
>
> Just for safety.  Sorry for not mentioning that in the commit message;
> I'll update it in the next re-roll.

I'm sceptical of such changes too. Could you keep them for another time?


-- 
HE CE3OH...