2026/11/17
|
Tuesday, November 17, 2026 (#321)
|
|
References |
Main Work
- DBA
- ⇒ problem: dumping each schema using the previous schema's filename
- ⇒ improving logging output so I can figure out what's going on
- ⇒ new rich-text output system
- ⇒ trying to work out how to do styles that have structural dependencies
- ⇒ new rich-text output system
- ⇒ improving logging output so I can figure out what's going on
- ⇒ problem: dumping each schema using the previous schema's filename
From yesterday: «Maybe RText objects should optionally host some kind of generic "OStyleData" object»... -- I had started to implement this when it occurred to me that RText objects already host an optional Venue object ($OVenue) -- perhaps that's a good place to store this? It is, obviously, Venue-aware...
The part which bothers me is that we're then assigning a single Venue to an RT object which might want to output to multiple Venues (e.g. show on screen and log to file)... but I think as long as we're going those renderings separately/serially, there's no problem.
Also, the original justification for that object was to allow for a default in order to reduce the necessary fiddling around to locate the Venue object -- specifically, when we ask the Screen for an RT Table, why not have that object pre-provisioned with the Screen's Venue? I'm not sure how valid that really is, but if it works for this usage as well then maybe it's sort of an "elegant kluge".
13:05
I think I also need to add a field in the RText class to store intermediate string output, so attributes can iterate on that while also having access to the RText object without having to pass it up and down separately.
16:57
Ehh, no, that doesn't work -- we're crawling down a tree of multiple RText objects, so we'd have to pass the root object each time... or, wait, are we? Need to look at what's going on when Style does its own crawl...