A new version of a patch should not be interpreted as reviewed

[ftomassetti]

@kodabb said

[26/09/2015 22:46:52] lu_zero: http://plaid.libav.org/patch/270/
[26/09/2015 23:39:26] Vittorio: che non c'è in
http://plaid.libav.org/project/libav/ o in
http://plaid.libav.org/project/libav/patches/new
[26/09/2015 23:40:16] lu_zero: perche` non e` new
[26/09/2015 23:40:32] lu_zero:
http://plaid.libav.org/project/libav/patches/reviewed
[26/09/2015 23:40:48] lu_zero: in compenso le patch con / vengono
interpretate maluccio
[26/09/2015 23:46:24] Vittorio: eh ma e' difficoltoso raggiungere le non new..
[26/09/2015 23:46:37] Vittorio: e poi io ho mandato una nuova versione
che dovrebbe far ritornare lo stato a new
[26/09/2015 23:46:37] lu_zero: come?
[26/09/2015 23:46:44] lu_zero: e` un link in prima pagina
[26/09/2015 23:47:00] lu_zero: sicuro che non l'abbia gia` commentata?
[26/09/2015 23:47:06] lu_zero: lu_zero pondera
[26/09/2015 23:47:31] Vittorio: di sicuro dovrebbe esserci un'altra
sezione per le "reviewed" in prima pagina
[26/09/2015 23:47:39] Vittorio: sotto new patches
[26/09/2015 23:47:46] Vittorio: e mi sembra che non l'avessi commentata
[26/09/2015 23:47:53] Vittorio: ma non so se fosse passata un'ora

[lu-zero]

53 is related.

[ftomassetti]

How do we want to implement this? If a patch is in the same thread of e-mails should we consider it a replacement? Should the patches look similar?

[lu-zero]

There is already an heuristic for single patches (based on the id provided by git send-email) and works well, the problem is that right now the code considers commented the patch instead of resetting it to new.

[ftomassetti]

Ok, I see. Can you give me any pointer about the point in the code where a patch is recognized as a substitute for an existing one?

[lu-zero]

https://github.com/lu-zero/plaid/blob/master/mailparse.py#L382