Built a new release and uploaded it into the system. Renamed the XAR to reflect the release number. The previous version was not preserved, but its release number was Version 6.2b, svn rev. 2028 2018-12-11 19:21:47 -0800, so if we need to build it again to roll back, we can do that. Had to build a couple of times because I forgot to increment the VERSION file.
CC asked me to add two features that were in the old site but hadn't yet been implemented in the new: automatic jumps to the first instance of a tagged term when following a link from a reference item to a text, and lists of texts containing reference items in the context of the text popup. I've now implemented both, but in the process I found a stack of bad xml:ids (accented chars etc.) and bad links, and had to fix all of those. I've also added a new Schematron rule to constrain ids properly. 180 minutes.
After seeing a problem in the rendering of a closer in Sonnet 1609, I discovered that we have no template handling closers, so I've added it to the generic block output template. Then CC and I worked to refine the rendering of one particular closer that was giving trouble.
We then addressed the issue of spaces before commas, which are possibly accidentals and possibly not; rather than remove them or regularize them in the TEI, we decided to remove the spaces in the normalized versions of the texts, so I've now added that to the hcmc:processText() function.
There were several instances in which document content such as figure and head elements was directly inside the front or body, without any intervening div. This was causing rendering issues -- head element styles were being lost, for example. I fixed all instances of this, and added a Schematron rule to prevent it in future.
We've also devised a bit of an encoding hack, involving nested date elements, to get around the TEI's decision not to allow @when alongside other attributes on dates. CC will implement it by examining all cases, and then I'll deal with any processing fallout as required.
Gradually figuring out the best way to include the list of other documents which link to a reference or a bibliography item. I'm coming around to the view that the way to do this is to put the list into the AJAX fragments (already done for references, not working for divs yet), and replace any existing link or reference in an HTML document with the content of the AJAX fragment (slightly processed to remove links to the current document itself, perhaps).
This document had some very weird structural issues caused during an earlier phase of encoding, which resulted in a large number of superfluous div elements and some chapter content finding itself outside of its chapter div. With CC, went through and fixed all these issues.
Make the introductions for each document type TOC collapse to an ellipsis. People need to see the TOC.
The reference entries used to be accompanied by a list of all the documents mentioning them, but the static site doesn't have that. Bring it back!
Mariage version 6.1 has been built and pushed to the live server. In the build process, found and reported a linkchecker bug.
I'm finally getting around to writing proper ODD-based documentation for the Mariage project, starting with a description of the use of Image Markup in the project, prompted by a request to talk about it. I've done an intro, and the section on the IMT and image markup in the project. More later...