| Apache/2.4.7 (Ubuntu) Linux sman1baleendah 3.13.0-24-generic #46-Ubuntu SMP Thu Apr 10 19:11:08 UTC 2014 x86_64 uid=33(www-data) gid=33(www-data) groups=33(www-data) safemode : OFF MySQL: ON | Perl: ON | cURL: OFF | WGet: ON > / usr / lib / python2.7 / dist-packages / twisted / internet / iocpreactor / | server ip : 104.21.89.46 your ip : 172.69.214.16 H O M E |
| Filename | /usr/lib/python2.7/dist-packages/twisted/internet/iocpreactor/notes.txt |
| Size | 890 |
| Permission | rw-r--r-- |
| Owner | root : root |
| Create time | 27-Apr-2025 09:56 |
| Last modified | 19-Mar-2008 00:35 |
| Last accessed | 07-Jul-2025 07:25 |
| Actions | edit | rename | delete | download (gzip) |
| View | text | code | image |
test specifically:
failed accept error message -- similar to test_tcp_internals
immediate success on accept/connect/recv, including Event.ignore
parametrize iocpsupport somehow -- via reactor?
do:
break handling -- WaitForSingleObject on the IOCP handle?
iovecs for write buffer
do not wait for a mainloop iteration if resumeProducing (in _handleWrite) does startWriting
don't addActiveHandle in every call to startWriting/startReading
iocpified process support
win32er-in-a-thread (or run GQCS in a thread -- it can't receive SIGBREAK)
blocking in sendto() -- I think Windows can do that, especially with local UDP
buildbot:
run in vmware
start from a persistent snapshot
use a stub inside the vm to svnup/run tests/collect stdio
lift logs through SMB? or ship them via tcp beams to the VM host
have a timeout on the test run
if we time out, take a screenshot, save it, kill the VM
failed accept error message -- similar to test_tcp_internals
immediate success on accept/connect/recv, including Event.ignore
parametrize iocpsupport somehow -- via reactor?
do:
break handling -- WaitForSingleObject on the IOCP handle?
iovecs for write buffer
do not wait for a mainloop iteration if resumeProducing (in _handleWrite) does startWriting
don't addActiveHandle in every call to startWriting/startReading
iocpified process support
win32er-in-a-thread (or run GQCS in a thread -- it can't receive SIGBREAK)
blocking in sendto() -- I think Windows can do that, especially with local UDP
buildbot:
run in vmware
start from a persistent snapshot
use a stub inside the vm to svnup/run tests/collect stdio
lift logs through SMB? or ship them via tcp beams to the VM host
have a timeout on the test run
if we time out, take a screenshot, save it, kill the VM