Roland Mainz wrote:
> Stefan Teleman wrote:
>> Roland Mainz wrote:
>>> "Laszlo (Laca) Peter" wrote:
>>>> On Tue, 2008-04-15 at 05:20 +0200, Roland Mainz wrote:
> [snip]
>> There are more pressing problems with libtool which probably need to
>> be addressed, before we start worrying about libtool performance.
> 
> Performance is the item where the FOX people are screaming loudest right
> now...

Not really.

They are complaining about the exact same things everyone complains 
about with respect to libtool. And the idea of making unnecessary 
changes to libtool hasn't exactly been greeted with a warm welcome at 
Xorg.

Corollary: starting a libtool flamewar with Daniel Stone is, IMHO, 
*not* a Good Idea(TM).

> 1. Report these issues to upstream

I reported them to you. You indicated that you were interested in 
working on improving libtool.

> 2. Upgrade libtool
> 3. Convince our manager to allocate time that I can do the upgrade

Upgrading libtool to the latest flavor won't fix a single thing -- the 
annoying problems still exist in the latest version of libtool, and 
upgrading will achieve exactly one thing: we now have the very latest 
version of libtool exhibiting the exact same problems as the previous 
one, or the one from two years ago.

What exactly is the point of the upgrade ?

Reporting complaints upstream without a patch is the functional 
equivalent to reporting to /dev/null. It's a complaint, not a fix, and 
it will be greeted as such.

Is there any indication that the libtool maintainers would consider 
switching libtool from bash to ksh93, and that they would accept a 
shell change patch, which affects all consumers of libtool ?

> Grumble...
> ... do you really think I would fight for a 5% improvemnt ?

You mentioned an 18 minute total build time speedup.

What is an 18 minutes improvement for a 4 hours total build time, 
expressed as a percentage ?

--Stefan

-- 
Stefan Teleman
Sun Microsystems, Inc.
stefan.teleman at Sun.COM


Reply via email to