Re: on branching...

[Michael Käppler]

Am 05.10.2020 um 15:17 schrieb Karlin High:
> On 10/5/2020 7:33 AM, Michael Käppler wrote:
> > Sure, it may be that there are
> > still lots of problems there, waiting to be uncovered. But debating
> > whether
> > we need more testing at first does not help, when we do not have
> > enough people to do it.
>
> What would a reasonably-adequate testing effort look like? Maybe that
> could be crowd-sourced if step-by-step instructions are given.
We could try to create a test suite consisting of a number of scores,
from small ones up to huge ones.
The testers would download the current unstable release for their OS and
run a batch script (python or even a
shell script) that runs LilyPond / lilypond-book and maybe other scripts
on the test suite. The output would
then be compared against a reference (pdfs, pngs etc. graphically, other
output as diffs) and the result
together with the corresponding log files sent back to us. Ideally there
would be no need for
manual interaction apart from downloading the release and the test
suite, running the tests and
sending the results back.

Note that 'we' and 'us' are highly under-determined in this set of
equations. :) :)

Cheers,
Michael