2026/07/10
|
Friday, July 10, 2026 (#191)
|
|
References |
Main Work
- Ferreteria FileSys stuff in order to support sftp in FTM.
09:11
The last couple of hours of yesterday's work involved splitting off the array-oriented pieces of Nodes so I could be clearer about where it needed to be kept open and where it could be shut.
This morning, I figured out where to shut each folder (after loading all the entries), and now it looks like the open-file-count is staying where it should be. Testing in progress.
10:25
We have a "How do we get here?" at Pair line 149 and a "MESSAGES: Actual-open of file (tot: 54300)". Hmm.
11:30
...except now the *files* aren't being closed (again), and even the force-shut call I put in doesn't do it. I guess I need to do some logging...
12:59
I really hate it when my brain is finally ready to dive into fixing a longstanding issue but I have to go out for an hour instead.
I really, really hate it.
22:03
There is clearly a quantum-code effect at work here: I spent most of today's focus-hours writing a logging system so I could figure out where the increase in open files was happening.
As soon as the logfile contains the information I need, the problem goes away. (Caveat: I haven't yet tried removing the key information to see if the problem comes back.)
owait -- may have been a bug in the accounting in my information-calculations... (It's bugs all the way down, isn't it.)
Aha!
I'm shutting the files properly, but for some reason they aren't actually being closed...
23:10
...because I was decrementing the number-of-opens counter before calling the actual-shut, and then the actual-shut was also checking the counter and declining to close the file if it had reached zero (probably a remnant of an earlier way of doing stuff). Tentatively, looks good -- but as before, more testing needed (tomorrow).