Moses 2025-08-25 to 2025-08-29
to : Martin Holmes
Minutes: 525
On Monday I started working on the last tiny final issues in the PDF output, first addressing the need to post-process the XSL:FO produced initially so that the turned ash character can be replaced by a regular ash which is superscripted using style; that was relatively straightforward, but the following issue, the problem display of the capital V with acute accent, proved to be really awkward, and every transformation-based approach I took failed. Eventually, I decided to hack the actual fonts to add a new composed character in the Private Use Area at U+E000, and I did a first draft of that just to make sure it was workable. I now need to re-edit those characters to move the accents a little higher, then I think it will work OK. Very time-consuming job, though. I have hacked fonts in temp/abFonts on my drive right now, but when they’re finished and tested, they need to end up in the repo, but also in the xep/fonts folders on both my computer and Saga.
On Tuesday, went back to tweaking the position of the acute accents, and I believe I have them looking pretty good. Rebuilt the fonts and generated a test build of the dictionary to see how it looked. However, the regenerated fonts were no use at all; the process of creating them seems to have borked all sorts of parts of the fonts, making the result unusable. Then I tried working with the glyphr studio online tool, which generates OTF files; I was able to convert those to TTF using the Python otf2ttf command from fonttools. This time I got decent character output, but the characters were positioned wrongly, so I guess the kerning had not been inherited when I pasted the vectors in, and other characters were also borked in the process, so we have now given up on this.
On Wednesday, worked with SK on a number of rendering issues in the PDF, and got most of them fixed.
On Thursday, got to work on the latest set of changes in the latest TODO list, just completed. Got through three changes, and discussed a couple of other potentially problematic ones.
On Friday, took a look at the one remaining issue in my TODO list, discussed it with SK, and left it for a decision by the team. Also discovered that we had been bitten by the TEI 4.10.0 release bug, and so rebuilt the schema to fix build breaks.