Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2015-07-31

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

All times shown according to UTC.

Time Nick Message
00:17 mizu_no_oto joined #darcs
00:39 mizu_no_oto joined #darcs
01:06 mizu_no_oto joined #darcs
02:04 mizu_no_oto joined #darcs
06:29 Heffalump fr33domlover: slow loading in what sense, of individual pages?
06:29 Heffalump notdan: how's the rebasing going?
06:54 alexei___ joined #darcs
07:31 fr33domlover Heffalump, http://dev.rel4tion.org
07:32 fr33domlover It just gets stuck in the middle of page loading. Perhaps it's time for me to learn how to debug Haskell code and see what the code does :P
07:32 fr33domlover While loading is slow, there is no rise in CPU usage, so it's not like darcsden is busy generating some page
07:33 fr33domlover (could even be a bug due to bumping some dependency version or something)
07:43 Heffalump perhaps try strace
09:44 amgarchIn9 joined #darcs
11:58 fr33domlover joined #darcs
11:58 fr33domlover hello again
11:59 fr33domlover looks like the slowness was due to some network issue,either way after network reconnect it seems to work fast now
12:00 fr33domlover URLs contain port 8900 which is invalid, but I'll fix that ...
13:46 mizu_no_oto joined #darcs
15:04 mizu_no__ joined #darcs
15:17 neto joined #darcs
15:25 sea-gull joined #darcs
15:26 gh_ joined #darcs
15:27 gh_ hi
15:27 gh_ Haskell Platform 7.10.2 was released: http://mail.haskell.org/pipermail/haskell-cafe/2015-July/120605.html
15:40 xauth joined #darcs
16:46 notdan Heffalump: eh could be better, but I am working on it. Currently trying to squash together various cleanup patches
16:47 notdan Heffalump: also the day before yesterday I told you I almost rm'd my homedir: turns out I *did* manage to rm my ~/.emacs.d/.. Which I found out today by trying to start emacs
16:52 notdan lucky i seem to have a backup
16:58 mizu_no__ joined #darcs
16:59 Heffalump notdan: would following sm's suggestion of re-recording from scratch make sense?
17:52 mizu_no_oto joined #darcs
18:16 mizu_no_oto joined #darcs
18:16 Riastradh joined #darcs
18:24 mizu_no_oto joined #darcs
18:37 notdan Heffalump: well, I don't want to loose a lot of history
18:51 Heffalump you don't have to record just one patch
18:56 notdan Oh yeah, I mean I would loose intermediate patches -- sometimes you can read off them (with the help of darcs annotate) some explanations for this or that choice
18:56 notdan sm: pingue
18:58 notdan sm: what do you think about those patches?
19:07 Heffalump history is valuable but not worth days and days of work to preserve
19:17 mizu_no_oto joined #darcs
19:20 notdan OK, you have a good point. I will try to do that
19:22 notdan I mean the history will still be saved *somewhere*
19:26 sm hi.. this is an old debate in the git world (keep all gory details vs record a clean logical history)
19:28 sm if there's so much gory detail we can't easily review and merge, then cleanup is needed
19:28 sm imho
19:31 Heffalump I think the point here is that a clean logical history exists in theory, but is painful to construct
19:32 Heffalump so we need a compromise that maybe skates over some of the logical history but is mergeable
19:35 sm I'm not asking for anything extreme. Just some independent mergeable consecutive smaller patch groups
19:38 sm I mean: some smaller groupings of related patches
19:39 sm that I can understand
19:39 sm why, why is this taking so long :)
19:42 notdan sorry guys :( i just.. happen to suck at this, as it turns out
19:43 notdan sm: I realized that I haven't included a link to my previous message: http://hub.darcs.net/simon/darcsden-dev/compare/co-dan/darcsden-to-merge
20:02 sm notdan: nice, thank you
20:02 sm what does "Read repository website from _darcs/darcsden/website + test" affect ?
20:04 sm or do ?
20:04 mizu_no__ joined #darcs
20:06 amgarchIn9 joined #darcs
20:09 notdan It adds a feature (only for the local backend), it reads the website information (that is normally stored in couchdb in the regular backend) from the file _darcs/darcsden/website
20:09 notdan and the tests for that: checking that it indeed is displayed correctly
20:10 sm notdan: ok thanks. Don't be shy about dumping quick notes like that in the long description, it makes patches much more digestible (I am trying to learn that myself)
20:10 sm what's the security impact of "Implement HTTP Basic auth" for someone running a public darcsden instance like darcs hub ?
20:12 sm also I see you added the errors package, are we handling errors differently in different parts of the app do you think ?
20:15 sm maybe you just use it for the hush function
20:16 notdan Ok, so the security implication is that now people can login in by using HTTP auth and adding ?login=true to the URL. It should be as secure as the default login procedure: http://hub.darcs.net/co-dan/darcsden-to-merge/browse/src/DarcsDen/WebUtils.hs#162
20:16 notdan And yeah, I only use the errors package for their utility function
20:16 notdan functions*
20:16 notdan I use hush in that patch, and I think I am using more of the functions from that package later on
20:17 notdan yeah I also make use of `readErr'
20:20 sm with basic auth always on, I could post links to something private that log you in automatically, right
20:21 sm (notdan)
20:23 gh_ joined #darcs
20:35 sm if someone publishes links like that, would it let search bots index private stuff ?
20:37 notdan hm
20:37 notdan I guess it's possible. If search bots can accept cookies
20:41 sm notdan: adding always-on basic auth to a production web app feels a bit weird, a move in the wrong direction
20:41 sm how do others feel about it
20:43 sm in future when I hope we start requiring HTTPS always, will basic auth still work ?
20:44 notdan Yep
20:44 notdan but perhaps you are right, we should make some sort of trigger
20:45 notdan maybe we can skip this for now, it's fairly autonomous (I hope!)
20:45 sm it it's not a huge effort, making it a config option or enabled only for testing/local use would be nice
20:46 sm or if #darcs says it's fine as is, I'll merge it now. I've merged the preceding three
20:47 notdan ++ for the config option
20:47 notdan let me see if those patches commute with the rest
20:47 notdan i am currently doing a clean re-record, as you've suggested
21:11 Heffalump it's probably safest to make it local-only
21:11 Heffalump (basic auth)
21:14 mizu_no__ joined #darcs
21:59 notdan OK, I mananged to reformat all the changes
22:00 notdan http://hub.darcs.net/simon/darcsden/compare/co-dan/darcsden-to-merge
22:01 mizu_no_oto joined #darcs
22:07 notdan hm let me put more comments in the commit messages
22:11 fr33domlover hello again, another newbie question: suppose I move all my repos from plain darcsweb to darcsden. Is there a known safe process to migrate when 1) I upgrade darcsden by darcs pull; 2) I move from darcsden-1.1 to the upstream darcsden
22:11 fr33domlover just to be sure it's safe to move the repos etc. and have their tickets and other metadata safely persistent
22:11 fr33domlover I assume it
22:12 fr33domlover 's possible since sm does it somehow :P
22:19 notdan Well, usually upgrading darcsden should not cause any troubles, unless an upgrade changes the database schema or something.. But in order to be safe I would backup the repos and the couchdb documents
22:24 notdan sm: can you check out <http://hub.darcs.net/simon/darcsden/compare/co-dan/darcsden-to-merge> and some additional explnations at <https://privatepaste.com/4844ae4b93>?
22:24 notdan sm: i must appologize for torturing you (and ganesh) with this merge
23:56 mizu_no_oto joined #darcs

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