On Mon, Sep 19, 2011 at 04:10:05PM +0200, [email protected] wrote:
> On Sep 19, 2011, at 4:04 PM, Neil Puttock wrote:
> 
> > On 19 September 2011 13:05, [email protected] <[email protected]> 
> > wrote:
> > 
> >> Perhaps create two new make options:
> >> 
> >> make unsafe-test-baseline
> >> 
> >> and
> >> 
> >> make unsafe-check
> >> 
> >> These would not have the fail-if-unoptimized check.

Why not do it the other way around?  Leave the default alone, but
add a separate target for -requires-unoptimized.

> Graham - how difficult would it be to change LilyPond's build in
> GUB such that it used different commands to run its regtests?

1-5 hours.  If you haven't run "make bootstrap" yet, then add
another 1-5 hours to that total, to accommodate fixing bugs in GUB
to allow it to complete that step on your computer.

These estimates do not include the 5-20 hours that your computer
will spend compiling software in the background.  Do not even
think about attempting this inside virtualbox.

> Also, it may be worth it to consider scrapping optimized binary
> altogether - it'd be good to test how much overhead the
> unoptimized version introduces with respect to the optimized
> version.

I don't think we want to be shipping unoptimized binaries.

Cheers,
- Graham

_______________________________________________
bug-lilypond mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to