Message6364
But there's probably actually 2 issues here:
1) Explicitly closing a file in this situation doesn't work. The EDT file writing in this example is using the with statement and even that doesn't flush the file. This is probably due to a little logic bug with the recent closer changes
2) However even with a quick fix for #1 I think we're still going to have the same problem when you *don't* explicitly call close on the file. 2.5.1 would flush that file on exit too
I think we can probably fix #2 for command line jython by joining all non daemon threads before having the main thread call cleanup().
Command line CPython does this. It calls threading._shutdown (which joins all threads registered with the threading module) before exiting.
Jython doesn't call threading._shutdown, but it wouldn't help this bug if it did: because the EDT thread is never registered in the threading module (threads created from pure Java are only registered with it lazily when the threading module is first used) |
|
Date |
User |
Action |
Args |
2011-01-30 21:32:56 | pjenvey | set | messageid: <1296423176.55.0.445240875958.issue1701@psf.upfronthosting.co.za> |
2011-01-30 21:32:56 | pjenvey | set | recipients:
+ pjenvey, fwierzbicki, zyasoft, alex.gronholm |
2011-01-30 21:32:56 | pjenvey | link | issue1701 messages |
2011-01-30 21:32:56 | pjenvey | create | |
|