tm: reset T_ASYNC_CONTINUE flag in t_suspend

[grumvalski]

Related to http://lists.sip-router.org/pipermail/sr-users/2017-January/095772.html
Resetting the flag in t_suspend allows adding new branches in a failure route while suspending the transaction.

[miconda]

Probably it's ok -- couldn't see a side effect right now. Is it tested/used anywhere else than src/modules/tm/t_reply.c:1123?

[grumvalski]

Currently it's only used in that t_reply check.

…

On Mon, Jan 30, 2017 at 1:01 PM, Daniel-Constantin Mierla < ***@***.***> wrote: Probably it's ok -- couldn't see a side effect right now. Is it tested/used anywhere else than src/modules/tm/t_reply.c:1123? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#955 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADip4w48vD_4NkGmo1-ISG2b8JWDK1Uzks5rXdEJgaJpZM4LwvBV> .

[miconda]

Unless you think it should be better to be reset when t_continue() is finished instead of inside t_suspend(), you can merge it.

[grumvalski]

You're right, I'll try this 2nd approach.

…

On Mon, Jan 30, 2017 at 1:19 PM, Daniel-Constantin Mierla < ***@***.***> wrote: Unless you think it should be better to be reset when t_continue() is finished instead of inside t_suspend(), you can merge it. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#955 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADip4-uUCfyyOtIztbsfOXue3BnqzQh_ks5rXdVHgaJpZM4LwvBV> .

[kamailio-sync]

Thinking it better, wouldn't moving the flag reset at the end of t_continue make the check in t_reply useless? I have to admit that I don't fully understand that check anyway :) On Mon, Jan 30, 2017 at 1:21 PM, Federico Cabiddu <notifications@github.com> wrote:

…

You're right, I'll try this 2nd approach. On Mon, Jan 30, 2017 at 1:19 PM, Daniel-Constantin Mierla < ***@***.***> wrote: > Unless you think it should be better to be reset when t_continue() is > finished instead of inside t_suspend(), you can merge it. > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#955 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ADip4- uUCfyyOtIztbsfOXue3BnqzQh_ks5rXdVHgaJpZM4LwvBV> > . > — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#955 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AF36ZYAsvBRJ6KyaWuVm-Jg6pmA9CZoIks5rXdXKgaJpZM4LwvBV> . _______________________________________________ sr-dev mailing list ***@***.*** http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev

[miconda]

It's a bit foggy the purpose right now, or better said the duration when that flag is needed. I didn't have the time to dig into it too much and don't remember the full purpose for it right now. I expected to be related to the fact of choosing the final response while a suspended branch is still processed.

But I think your patch is better in this regard, because it preserves the old behaviour (which didn't have a complaint on this flag) longer, until a new suspend is done. So can be merged as it is.

[grumvalski]

Thanks, should I backport it too?

…

On Mon, Jan 30, 2017 at 4:06 PM, Daniel-Constantin Mierla < ***@***.***> wrote: It's a bit foggy the purpose right now, or better said the duration when that flag is needed. I didn't have the time to dig into it too much and don't remember the full purpose for it right now. I expected to be related to the fact of choosing the final response while a suspended branch is still processed. But I think your patch is better in this regard, because it preserves the old behaviour (which didn't have a complaint on this flag) longer, until a new suspend is done. So can be merged as it is. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#955 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADip44rHstoc23NanawiPEhAdBUkWDJwks5rXfx4gaJpZM4LwvBV> .