Camelia, the Perl 6 bug

IRC log for #darcs, 2012-12-13

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

All times shown according to UTC.

Time Nick Message
00:48 mizu_no_oto joined #darcs
01:20 _ilbot joined #darcs
01:20 Topic for #darcs is now can't talk? see -> | http://wiki.darcs.net/IRC | http://darcs.net/ | darcs 2.8 is out http://wiki.darcs.net/Releases/2.8
02:54 intripoon joined #darcs
03:12 mndrix joined #darcs
03:17 mizu_no_oto joined #darcs
04:14 mizu_no_oto joined #darcs
08:29 raichoo joined #darcs
09:48 schlaftier joined #darcs
10:14 xymox joined #darcs
10:37 owst joined #darcs
10:48 xymox joined #darcs
11:34 drdo joined #darcs
12:17 owst markstos: http://hub.darcs.net/markstos/darcs-​screened/patch/20121212224353-20bb4 has a typo: "patchese"
12:18 markstos owst: Thanks. Will obit and re-record.
12:22 markstos owst: An amended patch has now been pushed.
12:23 owst Cool :-)
12:27 owst sm: If I revisit markstos' patch link ^ hub times out - the patch doesn't exist. Is there no way to distinguish between "can't find the patch" and "timeout, since the patch is taking too long to render"?
12:28 xymox joined #darcs
12:36 Heffalump I just merged some of his patches to screened, btw
12:36 Heffalump I think the correctness of the others is unclear
12:41 markstos Good $localtime Heffalump. Are you located in ".li" ?
12:41 markstos Thanks for the merge
12:42 markstos Heffalump: should I open a ticket about "prompt deps" options being bi-directional, while the docs for them are uni-directional ?
12:43 Heffalump no, .uk
12:44 Heffalump markstos: I guess so. I wonder if we can find a formulation that makes it clear but is still precise
12:45 Heffalump (and is bi-directional)
12:45 Heffalump I guess it's also possible the language was intended in the sense of dependencies = you need this patch to do this particular command with this other patch
12:46 Heffalump but that'd be rather confusing
12:46 markstos Maybe there are some Unicode characters for handwaving?
12:46 Heffalump :-)
12:47 markstos Heffalump: I understand what you are saying. Perhaps the word "dependencies" itself is safe.
12:51 Heffalump I think darcs has a clear sense of "A depends on B if you must have B in a repo to have A"; overloading that to sometimes mean the opposite is asking for trouble
12:57 markstos Heffalump: what about these: --dont-prompt-for-dependencies ==  "select all dependencies for matched patches without prompting" and --prompt-for-dependencies = "prompt about dependencies for matched patches [DEFAULT]"
12:57 markstos To me, "dependencies for..." is still clear, but not directional.
12:57 markstos ( not as directional anyway )
12:59 markstos Different thread: What about renaming the "screened" and "reviewed" repos to be "alpha" and "beta" ? I think it would immediately clear to new users which one had the additional QA filter applied.
13:09 Heffalump sounds plausible
13:09 Heffalump suggest it on darcs-devel?
13:57 markstos I've updated my prompt/dependency patches with the language proposed here earlier: http://hub.darcs.net/darcs/darcs-screened/patches
14:18 mizu_no_oto joined #darcs
14:54 mizu_no_oto joined #darcs
15:06 Heffalump markstos: makes sense. I think your changes to --no-deps are going in the wrong direction (as it were) though
15:43 mizu_no_oto joined #darcs
15:54 markstos rebase question: Is the difference between "unsuspend" and "reify" that with unsuspend, "amendments" get merged into existing patches, while with "reify", the amendments become new patches?
16:00 mizu_no_oto joined #darcs
16:15 mizu_no_oto joined #darcs
16:59 mizu_no_oto joined #darcs
17:29 mizu_no_oto joined #darcs
17:30 amgarchIn9 joined #darcs
17:40 kowey joined #darcs
17:40 kowey hello!
17:40 * kowey thinks he just used darcs-bridge to accept a GitHub pull request on one of his hub.darcs.net repos
17:41 owst hey kowey
17:41 owst how did you do it?
17:41 kowey (man if only you could tell hub.darcs.net to make friends with your github account)
17:41 kowey and then have one click merge
17:41 owst maybe
17:41 owst (in the future!)
17:41 kowey uhm, not sure I remember
17:41 owst Careful of it being broken ;-)
17:41 kowey I think I merged in my GitHub repo
17:42 kowey and then forgot my GH repo was out of date
17:42 kowey and then used darcs-fastconvert sync to push my more recent darcs patches to the Git repo
17:42 kowey and then darcs-fastconvert sync the other way (darcs?) to grab the Git patch into my Darcs repo
17:42 kowey and then sync again to push my tweaks back onto Git
17:43 kowey do I now need to push all these weird tags onto hub.darcs.net?
17:43 kowey like   tagged darcs-fastconvert merge pre-target: f36ca5f8e427fb06cc5d0a3ce6323b26
17:44 kowey but I think I understand the basic structure: you need darcs repo, git repo, darcs bridge repo, git bridge repo
17:44 kowey and you have darcs -> d-bridge (via darcs)
17:44 kowey d-bridge <-> g-bridge (via darcs-fastconvert)
17:45 kowey and g-bridge <-> git (via git)
17:45 kowey if I'm reading the docs right
17:46 owst kowey: those tags indicate that you're using the bridge with a git repo that uses branches
17:46 owst "don't do that!"
17:46 kowey I am?
17:46 kowey I thought I had no branches
17:46 owst Yes, those tags are created when the darcs-import is given a branch
17:47 owst your merges will appear as branch/merges
17:47 owst your github merges*
17:47 kowey git branch just says master everywhere
17:47 kowey oh
17:47 kowey so don't use the GitHub merge gui thing?
17:48 owst No. If you'd rebased the git side it would be ok
17:48 kowey darn
17:48 owst right, gotta run
17:48 kowey see ya :-)
17:49 kowey I hope I can tell GitHub to go back to the premerge state
17:52 * kowey wanders off (sorry for continued absence, will likely continue till March, although apparently company policy is that we're supposed to take Xmas week off)
17:53 kowey am in the middle of a cross-country move
17:53 kowey that said, it kind of feels like this isn't actually subtracting all that much!
17:53 markstos kowey: cross-UK?
17:53 kowey oops
17:53 kowey English fail
17:53 kowey moving from UK to France
17:54 Heffalump did you get your UK citizenship first? :-)
17:54 markstos cross-channel.
17:54 kowey starting new postdoc (and hopefully getting to do my contract work on the side)
17:54 kowey Heffalump: turns out I would have needed Indefinite Leave to Remain (like everybody says)
17:54 Heffalump ah
17:54 kowey despite my many re-readings of the UKBA website, just failed to see that requirement
17:54 Heffalump what did you have?
17:54 kowey and so might be eligible in a year, and me moving away now is a bit silly
17:55 kowey I'm on an extension for my Tier 1 (General) visa
17:55 kowey which is good until 2014-01
17:55 kowey at which point I think I'm supposed to ask for ILR
17:55 kowey and then apply for citizenship 2015
17:55 kowey but now the best hope is to hang around France, maybe become French?
17:56 kowey or just go Home and be Uncle Eric in Florida
17:57 markstos kowey: you have US Citizenship?
17:57 kowey yeah, since about 2000 or so
17:57 kowey at the time, I even had to renounce my Malaysian citizenship to get it :-/
17:57 markstos And before that?
17:57 kowey (maybe you still have to do that)
17:58 markstos I hear the US is not as supportive of dual-citizenship as some other places.
17:59 * kowey has to dash
17:59 kowey see you all! :-)
17:59 markstos left #darcs
18:03 schlaftier joined #darcs
18:11 mizu_no_oto joined #darcs
19:15 stepcut joined #darcs
19:51 gh_ joined #darcs
20:43 nand` joined #darcs
21:22 mizu_no_oto joined #darcs
21:38 watermind joined #darcs
21:38 watermind hi
21:38 watermind this error is confusing me...
21:38 watermind I have a "darcs failed:  Refusing to apply patches leading to conflicts.", not the first time, seen this before
21:39 watermind usally that's because I changed a file in the remote repository
21:39 watermind eh wait
21:40 watermind nevermind...
21:40 watermind ok think I understand what happened
21:41 watermind the conflict was with a file that I had deleted in both
21:41 watermind the remote and local repository
21:41 watermind but it still needed me to pull the patch
21:41 watermind which didn't do anything
21:41 watermind I thought since I had deleted the file localy too it wouldn't need to pull the patch
21:44 carter_ joined #darcs
22:16 gpiero joined #darcs
22:18 mizu_no_oto joined #darcs
22:25 gpiero hi, can anybody confirm that there's no option that can be passed multiple times ?
22:26 gpiero better: that in no case an option passed once or multiple times legitimately produces a different effect ?
22:35 Heffalump you can certainly override defaults in config files by an opposite option on the command-line, so it's possible you can also do e.g. --foo --no-foo --foo on the command-line
22:43 gpiero mmh, which should be the effect ?
22:44 gpiero tried wh --summary --no-summary  and it solves to wh -s
22:45 gpiero so wh --summary --no-summary --summary doesn't change the effect either
22:46 gpiero maybe the --foo --no-foo pair is only valid when one of the two is from a config file ?
22:46 Heffalump yeah, quite possibly
22:46 Heffalump are you trying to script darcs somehow?
22:48 gpiero looking to issue2238 ( problematic wh -ll )
22:48 gpiero it can be solved removing duplicating -l in wh
22:49 gpiero anyway was thinking that a more general solution could be deduplicating every option passed
22:49 gpiero if it make sense
22:50 Heffalump hmm, I'd be more inclined to see what's happening to the arguments as they pass through the processing (Arguments and Flags modules)
22:54 gpiero going to look (well, maybe going to have a dinner before :) )
22:54 gpiero thx
22:57 schlaftier joined #darcs
23:09 Heffalump gpiero: I had a quick look at the code and I have one candidate for the problem
23:10 gpiero limited to wh ?
23:10 gpiero or to the --foo --no-foo ?
23:15 Heffalump limited to specifically how wh deals with --look-for-adds. I added a comment to the issue on the tracker.
23:17 gpiero yes, afaics it's right
23:19 gpiero substituting delete with filter (LookForAdds /=) seems working for me
23:19 Heffalump cool
23:19 owst joined #darcs
23:19 * owst should've come on IRC before responding to -users email :-)
23:20 owst Heffalump: the "read all the patches" for rollback seems strange
23:21 owst The selection context suggests it's doing LastReversed (i.e. end-segment, latest first)
23:21 gpiero anyway was thinking if using (nub opts) in RunCommands.hs (around line 100) could be a more general solutions
23:22 gpiero hi owst, thanks for caring
23:22 owst gpiero: no problem
23:23 owst nub would destroy the --foo --no-foo --foo case though?
23:24 Heffalump owst: apparently that case doesn't work cleanly anyway
23:24 owst Oh
23:24 owst I didn't read carefully enough
23:24 Heffalump but to me nub feels like a bit too heavy a hammer
23:25 owst Yeah
23:25 gpiero btw, just noted: diff -u -p issue2238 (on my repo matches 3 patches) just displays the content of the last patch matching... maybe I'm remembering wrong, but it always acted like this?
23:25 owst it should be up to the indvidual commands to handle repetitions
23:25 Heffalump even if we don't have any cases where duplicates are expected right now, I can see that it might happen
23:25 owst yeah
23:26 Heffalump gpiero: I suspect it's always acted like this. It has to explicitly realise the trees to diff, and unless the patches can be commuted to be next to each other that wouldn't be possible.
23:26 owst Yeah
23:26 owst diff should probably warn you that only a single patch is considered
23:27 Heffalump I think amend-record -p would behave similarly
23:27 Heffalump there is this implicit idea of "only works on a single patch" for some commands, but I don't think it's called out very clearly
23:28 gpiero no, amend let you select the patch... I was thinking also diff did
23:29 Heffalump oh, true
23:32 gpiero going to have dinner for true this time... if you'll both be online later I'l bother you about the download-all-the-patches-for-rollback issue
23:35 Heffalump I won't be online for long (bedtime), but one general point about repo laziness bugs is that we should find a way to write tests about them
23:36 Heffalump e.g. setup fake repos that fail when you try to download patches that are too old
23:44 owst fire off a concurrent shell "thread" that monitors the atime of the file in question and somehow kills the script if it notices it change?
23:44 owst Sounds dodgy and flakey :-)
23:46 owst Or something with inotify
23:47 dcoutts joined #darcs
23:51 gpiero owst, inotify would only work on linux... is it enough or the test should run on every platform supported by darcs?
23:54 owst yeah, I know
23:54 owst probably ok to test it only on linux
23:54 owst it seems that something similar can be coded for windows
23:55 donri joined #darcs

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