[PROPOSAL] To retire autoconf for 4.1

[Kai Blin]

On 2013-05-27 08:27, Volker Lendecke wrote:

Hi folks,

> On Mon, May 27, 2013 at 12:43:56AM +0200, Jelmer Vernooij wrote:
>>> The first two are certainly the most annoying ones, but
>>> probably the two that are hardest to fix. If Tridge failed
>>> to fix something (startup time), this is my very definition
>>> of "unfixable". So I can only ask everybody to keep the old
>>> build system around.
>> That's not a very constructive approach...
>
> Right, but if the main proponent, himself being Andrew
> Tridgell, one of the most intelligent people I know, fails
> to solve an issue, I know the mere mortals won't have a
> chance. If Tridge fails, I know this is way beyond my
> intellectual capabilities.

Trying to remember what Tridge and I were going for when we were working 
on the s3-waf build, I would say a lot of the decisions we made were 
mostly aimed at making the transition from make to waf easier for people 
used to the autoconf build.

Looking back, I think most of the "Oh, if we do it that way, it'll be 
easier on people" decisions came back to haunt us somehow. Our attempt 
to emulate the huge monolithic Makefile instead of splitting up 
wscript_build files per subsystem is the one that most annoyed me, but 
there's probably many others I've managed to forget about.

One thing I don't remember looking at is "how fast does it build", 
because we frankly didn't care. It's no use focusing on speed if you 
can't get it to build in the first place.

Not trying to fix a problem we could only start to care about after the 
build system was operational doesn't equal not being able to fix it. If 
I apply your "tridge didn't do it" logic in a more general pattern, why 
do we have Samba the way it is now? Obviously the first version tridge 
released didn't do a lot of the things we're doing now. But because 
tridge didn't write those featuresm, they should have been impossible. 
So how come "mere mortals" managed to write those features?

This reductio ad absurdum is just what Jelmer was criticising, and it's 
not helpful in my opinion either. Identifying the specific issues you 
were seeing was helpful, on the other hand. I think getting rid of the 
build groups and thus making the build faster has been a positive 
outcome of this discussion already.

Cheers,
Kai

-- 
Kai Blin
Worldforge developer http://www.worldforge.org/
Wine developer http://wiki.winehq.org/KaiBlin
Samba team member http://www.samba.org/samba/team/