today i noticed that my scel autocompletion was not working right, because sclang-symbol-table was always ending up as nil after sclang-start. this behavior persisted even after restarting Emacs. whenever this happened, sclang-command-process-previous would contain a very long string with a partial symbol table plus some other command results at the end, so it seemed like something was going wrong in sclang-command-process-filter. despite my best debugging efforts, i could not figure out how exactly this happens.
i did however notice that scel always informs SC about all open sclang buffers, which at this point i had a lot of (500+ buffers overall, possibly 100+ sclang). on a hunch i decided to trim my buffer list down to just a handful, and sure enough, now sclang-start seems to work normally again.
sorry for the rather vague report. i need to set up some automation to open lots of buffers and see if i can find a breaking point...
also, this may well be the root cause of #57
Emacs 28.2, SC 3.13.0, macOS 10.14
today i noticed that my scel autocompletion was not working right, because
sclang-symbol-tablewas always ending up asnilaftersclang-start. this behavior persisted even after restarting Emacs. whenever this happened,sclang-command-process-previouswould contain a very long string with a partial symbol table plus some other command results at the end, so it seemed like something was going wrong insclang-command-process-filter. despite my best debugging efforts, i could not figure out how exactly this happens.i did however notice that scel always informs SC about all open sclang buffers, which at this point i had a lot of (500+ buffers overall, possibly 100+ sclang). on a hunch i decided to trim my buffer list down to just a handful, and sure enough, now
sclang-startseems to work normally again.sorry for the rather vague report. i need to set up some automation to open lots of buffers and see if i can find a breaking point...
also, this may well be the root cause of #57
Emacs 28.2, SC 3.13.0, macOS 10.14