LEMDO 2026-07-06 to 2026-07-10
to : Martin Holmes
Minutes: 390
On Monday, transitioned two diagnostics to Schematron rules, which is something we should do more often; it can only really be done when the number of instances is low (in this case, one case of each problem), and when we can assume the remaining issues re trivial to fix. Did some other diagnostics triage too.
On Tuesday, raised an issue on how we should handle localCit pointers
to <pb> elements; got a response from JJ, and implemented
the processing both for <pb> and <lb>.
On Wednesday, did some proofing and checking for the upcoming QME release,
and handled a ticket calling for the processing of WSB <idno>s
to be changed to match their new URL pattern.
On Thursday, spent much of the afternoon troubleshooting the QME release; once it was done, we discovered that the archiving code which creates a package for Zenodo was not aware of videos, and that it had a hard-coded user id; I fixed these issues, but then we found that the release was faulty and so it was rolled back. Bugs were fixed, and a new release will be done. I also updated the cron code that creates a listing of filenames for checking links so that it includes both editionImages and videos, which it was missing, and tested that, so we can now add a diagnostic to catch bad image links.
On Friday, spent an inordinate amount of time trying to debug the new diagnostic,
with IN. I rewrote the cron task to produce a complete set of full URLs to the target
images and videos, and rewrote the diagnostic to compare expanded pointers to that
list of URLs. However, the approach I always take to this simply didn’t work;
neither of us were able to determine why a straight comparison of
a string to a sequence of strings using the = operator was failing, when a call to
contains() using the untokenized text file was working. I ended up
leaving the latter comparison in place. Given the structure of the strings in
question, it’s unlikely that a mismatch or a missed match could occur in this
case, but it is annoying not to be able to see where the problem is.