Re: [gentoo-dev] Re: [gentoo-commits] repo/gentoo:master commit in: eclass/

[Mike Frysinger]

On 31 Oct 2015 09:08, Michał Górny wrote:
> On Sat, 31 Oct 2015 03:06:21 -0400 Mike Frysinger wrote:
> > On 30 Oct 2015 18:20, Michał Górny wrote:
> > > On Fri, 30 Oct 2015 12:03:59 +0000 (UTC) "Justin Lecher" wrote:
> > > > --- a/eclass/distutils-r1.eclass
> > > > +++ b/eclass/distutils-r1.eclass
> > > > @@ -322,8 +322,7 @@ distutils-r1_python_prepare_all() {
> > > >
> > > >  	_distutils-r1_disable_ez_setup
> > > >
> > > > -	if [[ ${DISTUTILS_IN_SOURCE_BUILD} && ! ${DISTUTILS_SINGLE_IMPL} ]]
> > > > -	then
> > > > +	if [[ ${DISTUTILS_IN_SOURCE_BUILD} && ! ${DISTUTILS_SINGLE_IMPL} ]]; then
> > >
> > > This was intentionally wrapped to stay within 72-column line width. Not
> > > saying the eclass is perfect in keeping text width, especially with
> > > others committing random changes to it, but that's no reason to
> > > introduce further offenders.
> >
> > Gentoo has never had a hard 80-col rule let alone 72-cols.  forcing a wrap
> > here makes no sense and the new version is an improvement.
>
> For years, Gentoo was unable to make *any* *sane* *global* rules. Which
> doesn't mean everything needs to be as crappy as the overall Gentoo
> 'quality'.

you forgot "imo".  we've had commonly accepted standards that people generally
kept to, some of which were more fuzzy than others.  but we've never had a hard
80 col rule and claiming that 72 col is an improvement is pretty hard to
swallow.  there's no reasonable argument for such minimal restrictions in
todays's world, and "it's always been that way" doesn't fly.

> So what's this improvement exactly? 'I like this style better, so it's
> an improvement'? As I see it, it's a pointless, changing-nothing-really
> commit that causes huge cache regen for no good reason except someone's
> fancy.

one thing that has been consistent is cuddling of the initial command.  we do
not write:
	if foo
	then
	for x in 1 2 3
	do
	while true
	do
we have always written:
	if foo ; then
	for x in 1 2 3 ; do
	while true ; do

as for "why consistency", i think your e-mail is a bit confused here.  one on
hand you beoman lack of official hard style rules while on the other claiming
that none exist and being inconsistent is fine.

as for cache regen due to changed eclasses, meh, that happens every day, and for
eclasses more widely used than this.  it's a non-issue considering the caches
are generated on servers and distributed to users.
-mike