from recent forum threads elements i assume the following
'macro-situation':
(0) python libraries selected for a project should be thread-safe
(1) user uses python libaries at his own risk (there 's no mean to verify
it's thread-safe)
(2) in order to minimize risks user should put a lock around all code using
the libraries :
def dangerous() :
with lock_libraryA :
do something with libraryA
with lock_libraryB :
do something with libraryB
(3) in case of application crashes where libraries are involved means
libraries cannot be used and designers should look for alternatives
is this correct ?
As regard the dangerous code 'micro-situation':
I think it' ll be ok to lock between threads. One line of code instantiates
the main library object (constructor takes 2 files arguments). Is that a
potential problem at the process level ? (code is readonly nothing is ever
written to the files)
As regard processes concurrency :
Doing tests with several scheduler processes and lot's of tasks, i have had
deadlocks so i wondered if postgreSQL Lock command could help prevent that ?
https://www.tutorialspoint.com/postgresql/postgresql_locks.htm
thanks guys for your support…….
@Dave : quote by Nietzsche
« Partout on entend la voix de ceux qui prêchent la mort : et la terre est
pleine de ceux à qui il convient de prêcher la mort. »
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
You received this message because you are subscribed to the Google Groups
"web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to web2py+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
