Wendat 2024-01-29 to 2024-02-02
to : Martin Holmes
Minutes: 390
On Monday, worked a bit on example rendering, and fixed some issues with entries.
On Tuesday met with ML and planned a number of things, including backup signing auth so I can take over RA timesheets if necessary, and a couple of issues with the schema which we fixed. After committing schema fixes, ML made use of new options in encoding, and my XSLT broke because it wasn’t expecting the nesting now allowed.
On Wednesday, fixed the build break.
On Thursday, ML reported that rather than <hyph> elements in
<stem> elements having their wrapping <st>
child supplied automatically at preprocessing time, as we have been doing, there are
cases where the entirety of the <hyph> contents should not actually
be enclosed in an <st> element (for instance, when there’s a particular
kind of prefix which is not actually part of the <st>), so in fact child
<st> elements should be supplied at encoding time; that will need some
remediation from ML, so in the meantime I’ve modified the preprocessing stage so that
it only supplies a wrapper <st> in cases where there is no existing
<st> as a child of the <hyph>.
On Friday, I finally got some uninterrupted time to get to grips with the dictionary
build. I got the final remaining AJAX component, glossary entries, working, and then
started on the main HTML generation. This is going slowly because I’m trying to
refactor the original process, which used a confusing array of different modes, into
two modes only, template
and html
. The source page template is now
preprocessed for language and footer build information before it’s used to build
individual pages, which should save some time, and the individual pure-HTML site
pages are now rendering, with the exception of the processing-instructions for listings.