Camelia, the Perl 6 bug

IRC log for #darcs, 2013-08-06

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:37 kofno joined #darcs
01:01 kofno joined #darcs
01:18 kofno joined #darcs
02:30 edwardk_ joined #darcs
03:13 edwardk_ joined #darcs
03:21 preflex_ joined #darcs
03:25 mizu_no_oto joined #darcs
03:45 dolio joined #darcs
03:46 enthropy sm: the version with the unsafeCoerce seems to work (after all it's passing the same values as code which used to work)
03:46 enthropy sm: nh2's branch also works just fine without modifications
03:47 enthropy that "check" code is either missing from nh2's branch, or it got rearranged (so I didn't find it)
03:47 mizu_no_oto joined #darcs
05:11 dolio joined #darcs
05:21 burp joined #darcs
05:30 kofno joined #darcs
05:52 konundra joined #darcs
06:34 burp joined #darcs
07:09 raichoo joined #darcs
07:19 alexei_ joined #darcs
08:53 owst joined #darcs
08:58 whaletechno joined #darcs
10:28 donri joined #darcs
10:41 gh_ joined #darcs
11:10 konundra joined #darcs
11:51 kofno joined #darcs
12:05 konundra joined #darcs
12:12 favonia joined #darcs
12:22 pointfree-w530 joined #darcs
12:49 mtp joined #darcs
12:49 mtp joined #darcs
12:57 mtp joined #darcs
12:57 mtp joined #darcs
14:02 bfrank_ joined #darcs
14:10 bfrank_ is there already compiled versions of darcs-fastconvert?
14:21 owst bfrank_: no
14:34 mizu_no_oto joined #darcs
14:41 bsrk joined #darcs
14:57 whaletechno joined #darcs
15:14 raichoo joined #darcs
16:03 whaletechno joined #darcs
16:08 gh_ joined #darcs
16:34 adnap joined #darcs
16:49 * Heffalump appears
16:49 konundra joined #darcs
16:57 mizu_no_oto joined #darcs
16:58 bsrk joined #darcs
17:16 * Heffalump reappears
17:17 Heffalump bsrk: how's it going?
17:19 bsrk Hi
17:19 bsrk I am working on bundles
17:19 bsrk specifically displaying a patch in detail
17:20 bsrk I could not figure out making a post request using send command
17:21 bsrk The server is down at the moment
17:21 bsrk so you will have to try out bundles at your local instance for now
17:23 Heffalump re the post request, what have you tried?
17:23 bsrk I have setup a dummy handler
17:23 bsrk which I then tried to make a post request to
17:24 bsrk I usied the logs to verify the requests being made
17:24 bsrk but there was never a post request
17:24 Heffalump how does displaying a patch in a bundle differ from displaying a patch in a repo?
17:25 Heffalump did you investigate why darcs isn't making the post requests?
17:25 bsrk The source is different, obviously; we will want to show it in a similar manner.
17:26 bsrk but we will have to recode the details.
17:26 Heffalump recode what details?
17:26 Heffalump I'd expect that it'd share most of the code, just be addressed differently
17:26 edwardk joined #darcs
17:27 bsrk We will reuse the pattern, but we can't reuse the code
17:27 bsrk because it assumes that the patch is a part of a repository
17:27 Heffalump how come?
17:28 Heffalump (sorry, didn't see your reply before I said that0
17:28 Heffalump can you abstract the code so that they share the core bits?
17:28 bsrk Yes
17:29 bsrk that is the preferred way
17:31 Heffalump ok, cool
17:31 Heffalump so, getting back to the post request, have you investigated what's happening on the darcs side? Did you try my suggestion to make sure darcs is actually configured to send patches?
17:32 bsrk yes
17:32 bsrk you mean the http flag?
17:32 Heffalump no, I mean the one that means that by default darcs doesn't send emails anymore
17:33 bsrk I tried putting a --to=url too
17:33 bsrk did not work
17:35 Heffalump I'm talking about the recent(ish) change in maniline where darcs doesn't send emails at all by default
17:36 Heffalump I think you just need to pass --mail
17:36 bsrk but we are not trying to send emails.. still I will try to add --mail and see what happens.
17:37 Heffalump I know we're not trying to send emails, but it's probably the same logic that handles both given that it's the darcs send command.
17:37 Heffalump You could look at the source of darcs send to see how this all fits togheter.
17:37 bsrk Okay
17:38 Heffalump also, if you don't understand my suggestions, ask :-)
17:38 bsrk :-)
17:38 bsrk I misunderstood, not did not understand.
17:40 Heffalump ok, but if you'd answered my email asking if you'd tried it, we could have cleared up the misunderstanding more quickly
17:40 bsrk yes
17:41 Heffalump anyway, if the --mail option doesn't help, try adding trace statements to the source of darcs send to see how far it's getting in trying to post the patch
17:41 Heffalump the code is there, so we need to understand what makes it be called or not
17:41 bsrk sure, I'll keep trying
17:45 bsrk is there anything else; if nothing else I'll go to sleep. :-)
17:45 Heffalump where did we get to with the relationship between repository compare and the existing fork display?
17:46 bsrk I revamped the original code, so that it would do both.
17:47 Heffalump OK - did you ask sm for feedback on how it looks now?
17:47 favonia joined #darcs
17:47 bsrk Oh, yes. He did give feedback
17:47 bsrk and I made changes.
17:48 bsrk It had to do with limiting the number of patches being displayed.
17:48 Heffalump and has he seen those chanegs?
17:48 bsrk ah, I think so? They have been there for a while.
17:48 Heffalump The last I remember was the conversation last Thursday/Friday: http://irclog.perlgeek.de/​darcs/2013-08-01#i_7400593
17:49 Heffalump which isn't that long
17:50 bsrk yes, he is making changes in parallel too, I have imported some of them.
17:51 Heffalump ok, I'll take a look
17:52 Heffalump I think that's it for tonight, i'll mail you about the next meeting when I know my availability
17:52 Heffalump good night :-)
17:52 bsrk bye. :-)
18:31 preflex joined #darcs
18:33 sm hmm
18:34 edwardk joined #darcs
18:42 Heffalump hmm?
18:43 alexei_ joined #darcs
18:48 sm oh, just disappointed that nothing has changed with http://hub.darcs.net:8901/simon/darcsden/patches or http://hub.darcs.net:8901/simon/darcsden/compare
18:50 sm top alignment bsrk ? I don't feel you use these screens
18:53 edwardk joined #darcs
18:53 sm merge checkboxes gone..
18:54 sm no response re patch cleanup & code review.. I'm holding off changing stuff so as not to create a bunch of work..
18:56 sm yes I'm whiny today.
19:00 Heffalump /patches looks very busy/confused
19:23 gh_ joined #darcs
19:28 dolio joined #darcs
19:40 schlaftier joined #darcs
19:54 mizu_no_oto joined #darcs
20:06 gh_ joined #darcs
20:31 edwardk joined #darcs
20:47 Heffalump what's running on port 8901, Aditya's latest code or just the head of darcsden mainline?
20:58 sm Heffalump: Aditya's latest darcsden-gsoc
20:59 sm or rather... the latest trunk + his latest darcsden-gsoc patches
20:59 Heffalump ok
20:59 Heffalump should we organise another meeting between you and him?
21:00 sm I dunno
21:00 sm have you had a chance to look at his unmerged patches ?
21:01 * sm does that now
21:02 sm wait, I'm working. another time
21:02 Heffalump :-)
21:02 Heffalump I'll take a look, I've been keeping an eye on them but not reading each one in detail.
21:03 sm will do it tomorrow morning, maybe he'll be around
21:04 sm I wanted them cleaned up but it's probably too much work at this point
21:05 Heffalump cleaned up in terms of removing conflicts, or other stuff?
21:06 lelit joined #darcs
21:07 sm combining into more complete logical patches, and removing a conflict resolution patch
21:07 Heffalump I'm fine with asking him to rebase if you want a nicer history
21:09 sm that would be my preference, these patches are pretty fragmentary and there's no real need to be committing conflict resolutions
21:10 Heffalump well, if he does rewrite history it screws up other people tracking his branch like me, so there are arguments against
21:10 Heffalump but I'm not that fussed
21:10 sm yes you'd have to re-get. Are you doing work on top of his ?
21:11 Heffalump no, just using my local repo state to help me keep track of what I've seen
21:14 sm I figure not, so it shouldn't be a big deal. But I'm not super attached either. The easy way is to pull everything and trust things will keep improving in trunk. I'm inclined to do this except I don't see him actively fixing issues reported. Also the chance to code review is quite valuable
21:14 Heffalump I'd rather your criterion for acceptance is "this would be acceptable even if no further work is done"
21:15 Heffalump since getting patches accepted is an important goal from the point of view of GSoC, that provides (more) certainty in terms of things being done acceptably
21:16 sm yes, and it's currently not in that state. I'd help with polish but haven't had the tuits just lately. So I should not be in a hurry
21:16 Heffalump well, the first thing I think about repository comparison is it should have a nicer UI so we can see what's in the queue :-)
21:17 Heffalump at the moment it just displays "Large number of patches, showing only a few", doesn't say which few, and doesn't give a way to see more.
21:17 sm yes.. it's a ui regression right now
21:22 * sm wonders if alex cole is here
21:23 sm how bad is it that hub uses http not https ?
21:26 Heffalump good in that it means darcs has a hope of uploading patches to it :-) (the post code in darcs send uses the haskell HTTP package which doesn't support https)
21:28 sm I'm amazed as
21:29 sm many technically minded people as there clearly are there have signed up
21:29 sm given this how security problem transmitting passwords in plaintext!"
21:29 sm (email excerpt)
21:38 Heffalump but of course noone would reuse their password on another site
21:38 Heffalump so they're only risking their access to this being compromised in future
21:39 Heffalump in practice, MITM attacks are at least a couple of orders of magnitude less frequent than sites themselves being compromised
21:39 konundra joined #darcs
21:39 sm I think any smart user should assume this site is likely compromised already
21:40 sm but still, if I'm doing something boneheaded I should stop
21:41 Heffalump I don't think it's boneheaded, but https wouldn't really hurt, apart from forcing us as darcs devs to fix darcs to support https post somehow
21:42 sm that should be doable with http-conduit I assume
21:42 Heffalump yeah
21:54 alexei_ joined #darcs
21:55 mizu_no_oto joined #darcs
22:08 edwardk joined #darcs
22:09 pointfree-w530 joined #darcs
22:43 dolio joined #darcs
22:48 alexei_ joined #darcs

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary