Re: stencils: let some stencils carry a box-extent; issue 3255 (issue 9295044)

[pkx166h]

I saw that there were some translation merges and David put in his elisp
fix as well, but I think I already ran my tests against that yesterday.

Anyway, to save time for pathcy, when ever I see a merge has taken place
I test a random patch (regardless if it fails) so that Patchy has a new
baseline to make check against when real patches come around - saves
about 20 minutes for the first new patch since a merge.

Anyway, just picked this patch again and it failed make doc, so I did it
again and it still failed make doc. I've not yet dug through the logs,
but could this be anything to do with the last two checkins?

I'm having a hard time thinking it could but would like a clarification
as I am at a loss to this apparent randomness of failure. I know how
troublesome it is for others to pinpoint make doc issues with
incremental changes because it is so time consuming.

....

I looked at the logs in more detail again and still it seems to start to
process

chords-funky-ignatzek.ly

then I get the child returned error at this point - same as a few days
ago.

Is there something special about this reg test that might break
something intermittently?

http://lilypond.org/doc/v2.17/input/regression/38/lily-1d33dc26.ly

It looks like something specifically TextScript based

https://codereview.appspot.com/9295044/