2026/04/09
< 2026
Jump to navigation
Jump to search
|
Thursday, April 9, 2026 (#99)
|
|
References |
- 09:10 Starting task: try to sort out what's going on behind the scenes in order to cause the readout to look like it's still trying to send the 72 bytes of the schema-list from somewhere to somewhere else. Why is that even still in the buffer?
- 10:46 I decided this would be easier to diagnose if the Updater were attached to the Canal instead of individual Streams. This involved creating a new Canal clade -- and at that point it seemed best not to keep using the old events-not-really-a-system, so I've started an Event clade. So far, it is making the code simpler (mainly in places where one event gets passed to multiple sub-objects).
- 11:28 We have once again re-achieved The Loop of Trying to Send 72 Bytes for Some Reason; now to implement the switchover to a Canal-based Updater.
- 12:22 Oh, right, that's why we can't make it Canal-based: when the Engine Client is setting it up, the Canals haven't been created yet.
- Maybe it should be Session-based? Will have to ponder this.
- 17:20 (One shopping-trip, one midday snack, and one laptop internals-exchange later...) I did stuff and now the schema-fetcher is broken again. Quelle sue-preeze.
- 21:24 Made some relatively obvious changes and started seeing updates (without much info), so that was good.
- Made one small tweak which I thought might get me a little more info -- and it fixed the schema listing. So that's also good.
- The exports are only partly working: it looks like the data is actually being exported, but there are no screen updates while the data is coming in -- and because of that, it times out on a sufficiently large export (anything that takes over 10 seconds). Therefore: I just need to fix the screen updates.
- 22:00 ...and now we run into the problem of
CheckStream()needing to do different things depending on what operation we're doing. Setting a callback seems too complicated and difficult to debug; maybe a mode setting?