Re: Bug squad: remember the regtest comparison

[Graham Percival]

On Wed, Aug 11, 2010 at 09:13:16AM +0300, Dmytro O. Redchuk wrote:
> On Wed 04 Aug 2010, 20:23 Neil Puttock wrote:
> > There's something fishy about the latest comparisons: a number of .log
> > changes suggest broken regtests,
> Sorry, could not catch that --- what exactly in test output[1] suggests broken
> regtests?

Look at the log for bookparts.log and page-spacing-markups.log,
for example:
-Fitting music on 1 page...
+Fitting music on 4 or 5 pages...
...
-Fitting music on 1 page...
+Fitting music on 3 or 4 pages...

This suggests that 2.13.29 has a problem with the page spacing
algorithm.  Something that fit onto 1 page (in .28) now requires 4
or 5 pages (in .29) !

I must admit that I never look at the logs, so it's great that
Neil found it.  But looking at the regtest comparisons isn't
either of our jobs; that's what you guys should be doing.  :)

> I've looked again in the manual[2] and couldn't find any relevant information.
> (Probably, something should be added to the docs, too?..)

As soon as you figure out what to add, I'll be happy to add it.

(no, this process isn't at all formalized -- but, as this bug
illustrates, it's very important!)

Cheers,
- Graham