[/F] > > huh? are you seriously proposing to break every single C extension > > ever written -- on each and every platform -- just to trap an error > > message caused by extensions linked against 1.5.2 on your favourite > > platform? [Barry] > What makes you think that a crash will not happen under Unix > when you change the API? You just don't get the Windows crash. > > As this thread has pointed out you have no intention of checking > for binary compatibility on the API as you move up versions. I imtimated the following, but did not spell it out, so I will here to clarify. I was -1 on Barry's solution getting into 1.6, given the time frame. I hinted that the solution Guido recently checked in "if (!Py_IsInitialized()) ..." would not be too great an impact even if Barry's solution, or one like it, was eventually adopted. So I think that the adoption of our half-solution (ie, we are really only forcing a better error message - not even getting a traceback to indicate _which_ module fails) need not preclude a better solution when we have more time to implement it... Mark.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4