Message158540
| Author | pitrou |
|---|---|
| Recipients | Jim.Jewett, amaury.forgeotdarc, asvetlov, dstanek, kristjan.jonsson, loewis, pitrou, rhettinger, stutzbach, tim.peters |
| Date | 2012-04-17.10:48:30 |
| SpamBayes Score | -1.0 |
| Marked as misclassified | Yes |
| Message-id | <1334659650.3338.9.camel@localhost.localdomain> |
| In-reply-to | <1334659080.54.0.88058540041.issue9141@psf.upfronthosting.co.za> |
| Content | |
|---|---|
> I _think_ the only python related things you can do from tp_clear() is > Py_DECREF(), this is what I mean by trivial. Well, Py_DECREF is not trivial at all, since it can invoke arbitrary Python code (through e.g. weakref callbacks, or by releasing the GIL). Therefore, I would say any code is allowed from tp_clear :-) > If file.close() can be an arbitrary python method, then it can no more > be called from gc, than an object's __del__ method. This would not be > a regression, this would be a fact of life. I don't believe it. I don't see what's magical about being called by the gc. Again, a Py_DECREF in tp_dealloc can invoke arbitrary Python code. |
|
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2012-04-17 10:48:31 | pitrou | set | recipients: + pitrou, tim.peters, loewis, rhettinger, amaury.forgeotdarc, kristjan.jonsson, dstanek, stutzbach, asvetlov, Jim.Jewett |
| 2012-04-17 10:48:30 | pitrou | link | issue9141 messages |
| 2012-04-17 10:48:30 | pitrou | create | |