Re: [Libreoffice] subsequenttest hang ...

On Mon, 2011-09-26 at 14:16 -0500, Norbert Thiebaud wrote:
> it is not new. the tendency of smoketest to hang was the reason it was
> not run in tinderboxes.

        Good - so we overcame one problem that can cause this.

> For instance. a build (the make itselft, excluding make check) with
> very good ccache hit typically take 13 minutes or so on my Linux
> Tinderbox... but with a bad ccache ratio (like a change a particularly
> pervasive header), that build time can climb to 55 minutes or so...

        Indeed, a crude top-level timeout would be rather hard to set right;
potentially ( assuming the build loop doesn't churn output on the
console ;-) we could read the output into a script that would allow a
timeout-without-printing, if we spent 30 mins not printing anything
during the build then we either are doing over-heavy LTO ;-) or it died.

> So I think the timeout safety should be in make test itself, and
> possibly with a  disable-switch to allow for debugging.

        Agreed; un-informed timeouts are pretty crude. Potentially we might
want to add some hooks to eg. vcl's Dialog::Show (or somesuch) to
trigger a fatal error in subsequent / unit / smoke-test mode.

        And of course de-coupling the 'svp' backend from the unx/generic code
would be nice too I guess :-)

        ATB,

                Michael.

-- 
michael.meeks@suse.com  <><, Pseudo Engineer, itinerant idiot