<html><head><meta http-equiv="content-type" content="text/html; charset=UTF-8"><title>Conversation with #varnish-hacking at Mon 21 Jul 2014 11:55:00 CEST on slink@irc.linpro.no (irc)</title></head><body><h3>Conversation with #varnish-hacking at Mon 21 Jul 2014 11:55:00 CEST on slink@irc.linpro.no (irc)</h3>
<font size="2">(11:55:00)</font><b> The topic for #varnish-hacking is: Varnish 4.0.1 released! <a href="https://www.varnish-cache.org/lists/pipermail/varnish-announce/2014-June/000697.html">https://www.varnish-cache.org/lists/pipermail/varnish-announce/2014-June/000697.html</a></b><br/>
<font size="2">(11:55:00)</font><b> Topic for #varnish-hacking set by scn!lkarsten@185.7.63.45 at 13:56:45 on 24/06/14</b><br/>
<font size="2">(12:07:29)</font><b> maryam [<em>maryam@2001:840:4904:1:3c22:5dcd:261f:f922</em>] entered the room.</b><br/>
<font size="2">(12:07:31)</font><b> maryam_ [<em>maryam@2001:840:4904:1:3c22:5dcd:261f:f922</em>] entered the room.</b><br/>
<font color="#16569E"><font size="2">(12:13:48)</font> <b>slink:</b></font> Should we have a way to preserve state between bereq retries?<br/>
<font color="#A82F2F"><font size="2">(12:17:38)</font> <b>scoof:</b></font> yes. #1512<br/>
<font color="#16569E"><font size="2">(12:18:52)</font> <b>slink:</b></font> scoof: thanks<br/>
<font color="#A82F2F"><font size="2">(12:28:41)</font> <b>jw_irc.vs:</b></font> slink: aren't you supposed to enjoy your vacation?<br/>
<font color="#A82F2F"><font size="2">(12:28:50)</font> <b>jw_irc.vs:</b></font> someone should ban you for a while<br/>
<font color="#A82F2F"><font size="2">(12:31:42)</font> <b>Mithrandir / tfheen:</b></font> slink: how's the island?<br/>
<font color="#16569E"><font size="2">(12:31:56)</font> <b>slink:</b></font> haha<br/>
<font color="#16569E"><font size="2">(12:32:40)</font> <b>slink:</b></font> Kristina thinks it was too hot outside to take a walk, so I thought I could look after a problem our martin has<br/>
<font color="#16569E"><font size="2">(12:32:46)</font> <b>slink:</b></font> Hidra is GREAT<br/>
<font size="2">(12:32:59)</font><b> dberg [<em>dberg@pool-108-54-102-216.nycmny.fios.verizon.net</em>] entered the room.</b><br/>
<font color="#A82F2F"><font size="2">(12:33:37)</font> <b>scoof:</b></font> slink: scn recommended vmod-var as a workaround to somebody else. I'm not sure how far it scales<br/>
<font color="#A82F2F"><font size="2">(12:33:46)</font> <b>daghf:</b></font> vmod var doesn'<br/>
<font color="#A82F2F"><font size="2">(12:34:13)</font> <b>daghf:</b></font> doesn't quite work on the backend side of vcl in v4 yet.<br/>
<font color="#16569E"><font size="2">(12:34:47)</font> <b>slink:</b></font> On #1512: I think we should either have two rollbacks (one for client, one for bereq) and have retry not copy from bereq0 again, or we should make rollback sensitive to the context (rollback req in client, rollback bereq in backend). Opinions?<br/>
<font color="#A82F2F"><font size="2">(12:35:37)</font> <b>scoof:</b></font> retry should definitely not rollback. I like the idea of rollback being context sensitive<br/>
<font color="#A82F2F"><font size="2">(12:38:07)</font> <b>daghf:</b></font> I think context sensitive makes sense <br/>
<font color="#A82F2F"><font size="2">(12:38:21)</font> <b>daghf:</b></font> that said, it looks like bereq.http and req.http are both available in vcl_pipe<br/>
<font color="#A82F2F"><font size="2">(12:38:29)</font> <b>daghf:</b></font> ..<br/>
<font color="#16569E"><font size="2">(12:38:49)</font> <b>slink:</b></font> is phk working today?<br/>
<font color="#A82F2F"><font size="2">(12:38:53)</font> <b>scoof:</b></font> no<br/>
<font color="#A82F2F"><font size="2">(12:40:26)</font> <b>daghf:</b></font> I guess doing a rollback for bereq wouldn't make much sense in vcl_pipe anyway<br/>
<font size="2">(12:43:03)</font><b> dberg left the room (quit: Ping timeout: 301 seconds).</b><br/>
<font size="2">(12:43:32)</font><b> dberg [<em>dberg@pool-108-54-102-216.nycmny.fios.verizon.net</em>] entered the room.</b><br/>
<font color="#16569E"><font size="2">(12:44:07)</font> <b>slink:</b></font> daghf: seems to make sense to me as we can't restart from pipe<br/>
<font color="#A82F2F"><font size="2">(12:44:37)</font> <b>daghf:</b></font> Ah, right <br/>
<font size="2">(12:53:36)</font><b> dberg left the room (quit: Ping timeout: 301 seconds).</b><br/>
<font color="#16569E"><font size="2">(13:15:44)</font> <b>slink:</b></font> hm, the idea of context sensitive actions does not fit too well into the existing vcc code<br/>
<font color="#16569E"><font size="2">(13:16:23)</font> <b>slink:</b></font> having rollback_req and rollback_backend should be easy<br/>
<font color="#16569E"><font size="2">(13:16:37)</font> <b>slink:</b></font> or what about req.rollback and bereq.rollback ?<br/>
<font color="#A82F2F"><font size="2">(13:16:49)</font> <b>phk:</b></font> slink, you should not be able to do rollback_req on the backend side, req is read-only there.<br/>
<font color="#16569E"><font size="2">(13:17:31)</font> <b>slink:</b></font> I think on current master it would roll back<br/>
<font color="#16569E"><font size="2">(13:17:49)</font> <b>slink:</b></font> so this probably is a chance to improve things<br/>
<font color="#16569E"><font size="2">(13:19:23)</font> <b>slink:</b></font> req.rollback and bereq.rollback do not fit in either because this is the variables notation. So I'd want to go for req_rollback and bereq_rollback ...?<br/>
<font color="#16569E"><font size="2">(13:23:17)</font> <b>slink:</b></font> phk: got an opinion on https://www.varnish-cache.org/patchwork/patch/171/ ?<br/>
<font color="#A82F2F"><font size="2">(13:26:07)</font> <b>phk:</b></font> slink, not today, about to give a talk at thecamp.dk<br/>
<font color="#16569E"><font size="2">(13:27:25)</font> <b>slink:</b></font> cool! the kamp at the camp :D<br/>
<font size="2">(13:31:41)</font><b> dberg [<em>dberg@pool-108-54-102-216.nycmny.fios.verizon.net</em>] entered the room.</b><br/>
<font size="2">(13:37:55)</font><b> dberg left the room (quit: SSL Connection closed).</b><br/>
<font size="2">(13:58:32)</font><b> hpeter [<em>hpeter@80.149.237.118</em>] entered the room.</b><br/>
<font color="#16569E"><font size="2">(14:38:39)</font> <b>slink:</b></font> posted patches for #1512.<br/>
<font color="#16569E"><font size="2">(14:38:56)</font> <b>slink:</b></font> reviews welcome BY EMAIL, I am going back to HidraHolidayMode now :)<br/>
<font color="#A82F2F"><font size="2">(14:40:32)</font> <b>Mithrandir / tfheen:</b></font> slink: enjoy, and say hi to your family again<br/>
<font color="#A82F2F"><font size="2">(14:40:38)</font> <b>Mithrandir / tfheen:</b></font> it was great to see you all again<br/>
<font color="#16569E"><font size="2">(14:41:56)</font> <b>slink:</b></font> Mithrandir: Hello back from Kristina and yes, it was great meeting you in Oslo. And the food even passed my MSG test :D<br/>
<font color="#A82F2F"><font size="2">(14:43:02)</font> <b>Mithrandir / tfheen:</b></font> oh, you don't do msg?<br/>
<font color="#16569E"><font size="2">(14:43:30)</font> <b>slink:</b></font> Mono Sodium Glutamate that is<br/>
<font color="#A82F2F"><font size="2">(14:45:32)</font> <b>Mithrandir / tfheen:</b></font> yeah, I know what msg is<br/>
<font color="#16569E"><font size="2">(14:45:42)</font> <b>slink:</b></font> ah. yes, I get sick from it<br/>
<font color="#16569E"><font size="2">(14:45:55)</font> <b>slink:</b></font> allergenic reaction...<br/>
<font color="#16569E"><font size="2">(14:46:57)</font> <b>slink:</b></font> signing off for real now. Have a nice day everyone, we're leaving to enjoy the beauties of Norway \o/<br/>
<font size="2">(14:47:04)</font><b> The account has disconnected and you are no longer in this chat. You will automatically rejoin the chat when the account reconnects.</b><br/>
</body></html>