2016-10-25 20:42:01 +0000 <mpickering> I looked quite a bit at inining/specialisation recently but couldn't come up with an example which failed to specialise so if you have one I would be interested to see. brb 30 mins
2016-10-25 20:42:04 +0000 <ryantrinkle> only A knows the actual instance to use
2016-10-25 20:42:54 +0000dju(dju@fsf/member/dju)
2016-10-25 20:43:03 +0000 <ryantrinkle> mpickering: perhaps we should verify that i'm actually seeing a failure to inline here
2016-10-25 20:43:11 +0000 <ryantrinkle> i'm not very experienced at reading core
2016-10-25 20:45:34 +0000 <ryantrinkle> is it fair to assume that if Main only exports 'main', and dump-simpl still has constraints of type Reflex (blah) getting passed around, it's not fully specialized?
2016-10-25 20:45:54 +0000 <ryantrinkle> (the export thing being relevant because that should mean that all the unused code is gone)
2016-10-25 20:47:20 +0000hamishmack(~hamishmac@121-73-30-206.cable.telstraclear.net) (Quit: hamishmack)
2016-10-25 20:48:05 +0000hamishmack(~hamishmac@121-73-30-206.cable.telstraclear.net)
2016-10-25 20:50:14 +0000Cale(~cale@2607:fea8:98e0:c3f:9d8f:73:82cc:fefa)
2016-10-25 20:51:55 +0000dedgrant(~dedgrant@192-40-237-82.dedicated.allstream.net) (Quit: Leaving)
2016-10-25 20:52:13 +0000newhoggy(~newhoggy@1.129.96.141)
2016-10-25 20:52:55 +0000hamishmack(~hamishmac@121-73-30-206.cable.telstraclear.net) (Ping timeout: 256 seconds)
2016-10-25 20:57:06 +0000dedgrant(~dedgrant@192-40-237-82.dedicated.allstream.net)
2016-10-25 21:01:37 +0000osa1(~omer@haskell/developer/osa1) (Ping timeout: 276 seconds)
2016-10-25 21:02:25 +0000 <mpickering> I would have to look at the core to say
2016-10-25 21:03:01 +0000 <mpickering> and if something is marked inlinable it should be specialised from what I understand
2016-10-25 21:03:25 +0000conal(~conal@104.255.240.10) (Quit: Computer has gone to sleep.)
2016-10-25 21:05:26 +0000conal(~conal@104.255.240.10)
2016-10-25 21:07:21 +0000 <ryantrinkle> mpickering: this code is under NDA, but it would be great if you could take a look at it; are you able to do contract work on this kind of stuff?
2016-10-25 21:29:55 +0000javjarfer(~javjarfer@78.250.221.87.dynamic.jazztel.es)
2016-10-25 21:31:55 +0000newhoggy(~newhoggy@1.129.96.141) (Remote host closed the connection)
2016-10-25 21:35:28 +0000hamishmack(~hamishmac@202-21-137-105.wlgcl1.acsdata.co.nz)
2016-10-25 21:37:53 +0000przembot1przembot
2016-10-25 21:43:43 +0000osa1(~omer@haskell/developer/osa1)
2016-10-25 21:47:25 +0000 <mpickering> niteria: Can you reliable reproduce the export list issue I posted about?
2016-10-25 21:48:52 +0000 <niteria> it reproduced twice
2016-10-25 21:49:03 +0000 <niteria> but only 3s difference
2016-10-25 21:49:50 +0000 <niteria> mpickering: here's my command: rm A.{hs,o,hi}; python gen12754.py > A.hs && time ghc A.hs
2016-10-25 21:51:08 +0000afarmer(~afarmer@199.201.64.133) (Quit: My Mac has gone to sleep. ZZZzzz…)
2016-10-25 21:52:55 +0000 <mpickering> I just got 26s vs 15s first run 24s vs 12s second run
2016-10-25 21:52:56 +0000 <mpickering> hmm
2016-10-25 21:53:13 +0000afarmer(~afarmer@199.201.64.133)
2016-10-25 21:53:42 +0000 <niteria> so it reproduces for you
2016-10-25 21:54:03 +0000 <niteria> I'm not running HEAD, it's ghc-8.0 with some extra stuff