Message236404
Posting RemoteDisconnected.v5.patch:
* Rebased and fixed minor merge conflict
* Change RemoteDisconnected base class from ConnectionError to ConnectionResetError
* Minor tweaks to tests
It seems that having a separate RemoteDisconnected exception (as in this patch) has at least two benefits:
1. It would allow the user to distinguish between a true ConnectionResetError (due to TCP reset or whatever) from a clean TCP shutdown
2. Backwards compatibility with user code that only handles BadStatusLine
The only disadvantage seems to be the bloat of adding a new exception type. But if some other comitter agrees that merging them is better and dropping backwards compatibility is okay I am happy to adjust the patch to go along with that. |
|
Date |
User |
Action |
Args |
2015-02-22 10:56:31 | martin.panter | set | recipients:
+ martin.panter, jhylton, mhammond, jcea, orsenthil, amak, rbcollins, cananian, r.david.murray, alanjds, agriffis, icordasc, demian.brecht, Yuri.Bochkarev |
2015-02-22 10:56:31 | martin.panter | set | messageid: <1424602591.68.0.0185997902504.issue3566@psf.upfronthosting.co.za> |
2015-02-22 10:56:31 | martin.panter | link | issue3566 messages |
2015-02-22 10:56:31 | martin.panter | create | |
|