> On 15 Sep 2017, at 10:43, Frediano Ziglio <fzig...@redhat.com> wrote: > >> >>> On 14 Sep 2017, at 17:55, Frediano Ziglio <fzig...@redhat.com> wrote: >>> >>>> >>>> Wasn’t there a discussion about also renaming the function? I believe >>>> “pending_send” is a symptom, not the reason for disconnecting. >>>> >>>> Christophe >>>> >>> >>> The discussion was on red_channel_wait_all_sent, we decide a comment >>> was enough. Said that we can have always updates. >> >> OK >> >>> >>> Are you talking about red_channel_client_disconnect_if_pending_send ? >> >> Yes. What is the rationale behind disconnecting if “pending_send”? >> >> >> Christophe > > Do you mean the function name or the reason? The function name didn't change > so there was no much discussion on it. > The reason why we disconnect is that in these code path we want to flush > data to clients to make sure some resources are freed.
Hmm, now I’m even more confused. Do we disconnect to free resources? Or is disconnecting the only way to flush data to clients (I doubt it)? Or do we flush because we disconnect, and want to make sure the last drops are out? > Not that I 100% agree with disconnecting clients on these cases but this > is not a regression of this patch. No, but since this patch rewrites the function and its call sites anyway, we may as well give it a name that is slightly more explanatory. > > Frediano > _______________________________________________ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/spice-devel _______________________________________________ Spice-devel mailing list Spice-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/spice-devel
