[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Tests that take more than ten times build time.

Re-salut Cyril,

Le Wed, Feb 13, 2008 at 03:30:24AM +0100, Cyril Brulebois a écrit :
> What about *relative* numbers, e.g. what % of the time is spent in the
> build itself (gcc and friends), and what % of the time is spent in the
> testsuite? Anyway, (almost) less than 1 hour everywhere isn't what I
> call time-consuming from a buildd point of view.

Test time on arch (build time)
 1h05 on sparc (3 min),
 37 min on mipsel (2 min),
 39 min on mips (3 min),
 37 min on powerpc (2 min),
 19 min on hppa (2 min),
 6 min on amd64 (1 min)…

Of course, if this is an exception, there is no need to argue. But in
the Debian-Med packaging team, we have a few package whose tests are not
yet enabled (bioperl, emboss,...); I do not know if it would be wise to
do so systematically if they have a similar pattern of CPU usage...

Or maybe in the end it would be the role of the port responsibles to
micromanage this?

> Popcon is no absolute knowledge. Anyway, 25+ users is far from
> negligible AFAICT.

Sure, but we are talking og 0.1 % of 25+ users. So if for one popcon
report there are 10 instalations, it still makes only 0.25 user. And
this includes the persons who downoladed the package but are not using

Bonne journée,

Charles Plessy
Wakō, Saitama, Japan

Reply to: