Camelia, the Perl 6 bug

IRC log for #mojo, 2013-05-19

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

All times shown according to UTC.

Time Nick Message
00:41 stephan48 vervain: you mean Mojolicious::Plugin::Subdispatch and Mojolicious::Plugin::RelativeUrlFor?
00:43 stephan48 memowe: you got two pull requests for fixing the modules problems
00:47 stephan48 contenticious itself seems to pass the tests
00:48 stephan48 and a quick function test makes it look likes it work
00:49 stephan48 works*
01:05 Hammerfist joined #mojo
01:30 whitebook joined #mojo
01:53 whitebook joined #mojo
01:53 KindTwo joined #mojo
01:57 Meiermann joined #mojo
02:10 human39 joined #mojo
03:02 preflex_ joined #mojo
03:32 mikegrb joined #mojo
03:42 fildon_ joined #mojo
04:00 preflex_ joined #mojo
04:06 suy joined #mojo
04:08 d4rkie joined #mojo
04:43 basiliscos joined #mojo
05:17 KindOne joined #mojo
05:39 KindTwo joined #mojo
05:49 Britzel_ joined #mojo
05:59 Adura joined #mojo
06:35 Mike-PerlRecruiter_ joined #mojo
06:37 Miked joined #mojo
06:54 denisboyun joined #mojo
06:55 jamesw joined #mojo
07:02 denisboyun_ joined #mojo
07:17 Vandal joined #mojo
07:18 mire_ joined #mojo
07:41 jzawodn joined #mojo
07:41 rihegher joined #mojo
07:41 rihegher left #mojo
07:50 mattastrophe joined #mojo
08:21 dod joined #mojo
08:25 abra joined #mojo
08:48 dod joined #mojo
08:52 denisboyun joined #mojo
09:03 denisboyun_ joined #mojo
10:21 abra joined #mojo
10:38 delias joined #mojo
10:39 rem_lex|pivo joined #mojo
10:39 denisboyun_ joined #mojo
10:40 amirite joined #mojo
10:40 amirite what happened to the memorize helper?
10:41 amirite i assume it got taken out and won't get put in again so i'll just remove the code
11:03 sh4 joined #mojo
11:24 denisboyun_ joined #mojo
11:27 whitebook joined #mojo
11:57 * jberger is watching the fowler talk
12:21 dpetrov_ joined #mojo
12:41 moltar joined #mojo
12:56 denisboyun_ joined #mojo
13:02 mattastrophe joined #mojo
13:06 whitebook joined #mojo
13:23 Miked joined #mojo
13:25 gryphon joined #mojo
13:29 bowtie joined #mojo
13:42 MichaelRpdx joined #mojo
14:04 whitebook joined #mojo
14:27 suy joined #mojo
14:27 dotan Looking at https://metacpan.org/release/Mojolicious I noticed a file called  examples/.DS_Store
14:28 dotan (I went there looking for the changes file, to see all the removed APIs the mailing list is grumbling about)
14:28 d4rkie joined #mojo
14:28 sri we know
14:28 sri it's harmless
14:44 sri the whole deprecation policy discussion is annoying
14:44 sri it's like they don't actually understand the consequences
14:46 d4rkie joined #mojo
14:52 sri vti must be a sociopath or so... "FUCK YOU AND EVERYTHING YOU DO! oh want to do an interview for my perl magazine?"
14:55 sri just a word of warning in case someone here is tempted to do an interview with him
14:55 Vandal sri, hi is Russian right?
14:56 Vandal or I'm confused him with some one else?
14:56 sri ukrainian i believe
14:57 Vandal good
15:25 autarch joined #mojo
15:26 autarch hi, I'm using Mojo::DOM and getting a bunch of warnings like this - Subroutine style redefined at (eval 491) line 2.
15:27 autarch aha, I'm caling $element->attr( style ... )
15:27 autarch and it should be attrs
15:28 autarch but shouldn't I just get a real error?
15:30 autarch also, after I style the style attr, this doesn't seem to be reflected in the parent DOM object
15:32 jberger autarch, can you make a quick example?
15:32 autarch sure, one sec
15:33 jberger also, you may be interested in dumping the tree via tree to inspect thing
15:33 jberger via ->tree
15:33 autarch https://gist.github.com/autarch/5608009
15:33 * jberger looks
15:34 autarch so basically, if I dump the element object after setting the styles, I see them set, but the parent $dom object isn't altered
15:35 jberger IIRC it should be
15:35 autarch it's hard to see a how a DOM that didn't do this could be useful
15:35 autarch so I kind of assumed it would work too ;)
15:36 autarch doh, severe PEBKAC
15:36 autarch yep, it's all me
15:36 autarch I was returning the unaltered HTML
15:36 autarch sorry about that
15:37 jberger haha
15:37 jberger ok
15:37 jberger no worries
15:38 autarch the upshot is that it works perfectly for what I'm doing
15:38 autarch I'm inlining CSS styles for an HTML email
15:38 jberger its a cute example actually
15:39 jberger I was going to be very sad if it didn't work as it looked like it should
15:40 jberger I starred the gist, if you don't mind
15:40 jberger I might refer to it as a fun use of a DOM
15:41 jberger autarch, with your approval of course
15:41 autarch sure
15:41 jberger win-win
15:41 jberger yaya
15:41 jberger yay
15:44 whitebook joined #mojo
16:00 rem_lex| joined #mojo
16:01 denisboyun joined #mojo
16:01 KindTwo joined #mojo
16:04 good_news_everyone joined #mojo
16:04 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/5DpWfA
16:04 good_news_everyone mojo/master d1e9066 Sebastian Riedel: removed a few generic Changes entries
16:04 good_news_everyone left #mojo
16:05 sri it's true, those are pretty pointless most of the time
16:05 jberger :-)
16:05 sri my ocd strikes again
16:06 sri and chinaxing actually delivered a test case \o/
16:07 jberger see! I had hope!
16:07 jberger 🎩
16:12 KindOne joined #mojo
16:13 autarch left #mojo
16:15 bpmedley joined #mojo
16:20 KindTwo joined #mojo
16:27 sri hehe, koorchik has talked himself into a corner on the list
16:29 KindTwo joined #mojo
16:31 Topic for #mojo is now 🎩, indubitably | http://mojolicio.us | http://irclog.perlgeek.de/mojo/today
16:32 abra joined #mojo
16:33 KindTwo joined #mojo
16:34 russum joined #mojo
16:34 russum left #mojo
16:38 KindOne joined #mojo
16:46 good_news_everyone joined #mojo
16:46 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/0X0iVQ
16:46 good_news_everyone mojo/master 565d075 Sebastian Riedel: fixed small drain event bug in Mojo::IOLoop::Stream (closes #491)
16:46 good_news_everyone left #mojo
16:50 sri was actually a pretty interesting bug
16:53 dbr joined #mojo
16:55 good_news_everyone joined #mojo
16:55 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/5Py3_Q
16:55 good_news_everyone mojo/master bd4c6d5 Sebastian Riedel: test tweaks
16:55 good_news_everyone left #mojo
16:56 KindOne joined #mojo
17:03 KindOne joined #mojo
17:04 good_news_everyone joined #mojo
17:04 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/bubKXA
17:04 good_news_everyone mojo/master 57a882f Sebastian Riedel: made Mojo::IOLoop::Stream code a little more defensive
17:04 good_news_everyone left #mojo
17:05 alnewkirk joined #mojo
17:16 sri NOOOOOOOOOO
17:16 sri yahoo bought tumblr
17:17 Vandal mwahaha
17:23 good_news_everyone joined #mojo
17:23 good_news_everyone [mojo] kraih tagged v4.01 at 78a27d6: http://git.io/HQJNcA
17:23 good_news_everyone left #mojo
17:27 sri hmm, Mojo::DOM::attrs would have been a candidate for removal too
17:27 sri or renaming at least
17:27 sri the similarity to Mojo::Base::attr keeps causing trouble every now and then
17:28 sri it's not so bad since we added $dom->{foo} though
17:28 sri well, maybe in 5.0
17:29 sri (not that i can think of a better name)
17:34 denisboyun joined #mojo
17:40 KindTwo joined #mojo
18:17 denisboyun_ joined #mojo
18:27 * sri turns the deprecation policy discussion around
18:33 xaka joined #mojo
18:35 Mike-PerlRecruiter_ joined #mojo
18:43 jberger sri: I'm not sure there are any apples to compare the mojo release cycle to
18:44 jberger the speed of mojo development is unmatched as far as I can tell
18:44 jberger to compare to any other project I think would implicitly as mojo development to slow dow
18:44 jberger n
18:48 whitebook joined #mojo
18:59 yko jberger: and probably that is what people trying to actually achieve
18:59 jberger yko, again, use Carton or elsewise to have a steady development
19:00 jberger s/development/environment/
19:00 yko hey, hey, it's not about me. it's about people
19:00 jberger understood, :-)
19:01 yko 6-8 months ago I had something internal runnning 1.6 or 2.22, I'm partially fine with that
19:01 yko I mean, did team rejected yearly release cycle/inbranch development? I was not following the discussion
19:02 crab what is carton?
19:03 yko Carton is https://metacpan.org/release/carton
19:03 yko Perl module dependency manager (aka Bundler for Perl)
19:04 yko A way to lock yourself from upgrading :) and so on
19:04 crab i see.
19:11 sri jberger: that's what i'm trying to find out
19:12 sri so far it sounds like "we want the exact same development speed but a little slower"
19:15 jberger well I would like a pony, but I would like it shaped like a mule
19:15 jberger :-)
19:16 yko one of the points there - people want to have libraries on cpan. it's hard to chase you
19:17 jberger I would say that CPAN modules are the hardest point
19:17 sri i'm not sure i would survive a yearly release schedule
19:17 yko jberger++ # i would like my bicycle back. i would like druggy who stole it to die in fire
19:18 jberger sri: what about a short official freeze before a major release, with widely publicized changes list, aimed at CPAN module authors
19:18 jberger say maybe a week
19:18 yko sri: it's totally up to you, I'm just asking, not trying to push you to do something
19:19 sri jberger: the only proposal so far has been 3 months
19:19 sri jberger: you're welcome to post it to the list, i'm ok with a one week freeze before a major release
19:20 jberger let me start again, the only real problem that I see is that its actually impossible for modules not to break in the face of a mojolicious major release
19:20 jberger until you release it, module authors cannot release ...
19:20 yko in the end that would not help. people will argue at you anyway as only you break their libraries on cpan
19:20 sri jberger: that is correct
19:20 jberger after you release it, modules are broken until they do
19:20 jberger no matter how much lead time, there must be a gap
19:21 sri jberger: that's my point with truly breaking changes, in one version of mojolicious your plugin is broken anyway
19:21 jberger perhaps, rather than a defined freeze, have a defined release day, maybe known in advance
19:21 jberger doesn't have to be firm
19:21 sri that doesn't work for me
19:21 jberger "we intend to release mojolicious 4.0 on day X, module authors should look at breaking changes"
19:22 jberger this doesn't have to be policy even, just an announcement
19:22 sri we need to be able to make last minute changes and re-freeze
19:22 jberger sure, we can push it back
19:22 jberger as we did
19:23 jberger I'm just saying, the best case scenario is to have CPAN module authors ready to release updates right after a major mojo release
19:23 sri and the freeze would be limited to breaking changes, so docs/tests can be fixed
19:23 jberger an announcement a week ahead of time might help them
19:23 jberger approximately a week-ish
19:23 jberger and where you reserve the right to push it back, of course
19:24 yko after all, somebody might take care to test mojolicious-dependant libs on cpan and inform authors
19:25 yko optional thing, though
19:27 russum joined #mojo
19:28 * jberger didn't notice the topic
19:28 jberger hehe
19:28 jberger when was that changed?
19:29 yko couple of hours ago. tempire did
19:30 sri "A major release is signaled by a new major version number and a unique code name based on a Unicode character, before every major release there needs to be a stabilization period of at least one week in which no breaking changes can be made."
19:34 xaka joined #mojo
19:34 jberger I like the concept, as I said, I almost think the announcement is more important than the freeze
19:35 jberger though of course major changes after the announcement probably are asking for trouble
19:35 sri there are no rules for announcements
19:35 jberger right
19:36 jberger I'm only asking if the freeze should actually be policy, or just a guideline
19:36 * jberger is arguing back-and-forth
19:36 sri post it to the list and see if it gets accepted
19:37 jberger remember, part of the reason we skipped the last few numbers of 3.XX was that something (I can't remember what now) was broken and needed an immediate breaking fix
19:37 jberger I kinda would hate to lose that flexibility
19:39 marty joined #mojo
19:40 jberger seems that stefan is going down my avenue as well
19:46 sri Mojo::URL
19:46 sri the data thing
19:46 sri yea
19:47 sri you bring up a good point though
19:48 sri what about breaking changes for major security flaws
19:48 jberger that's why I think this should be "we hope" not "we will"
19:48 sri it's bound to happen again at some point
19:48 jberger of course
19:48 * jberger posted to the list
19:48 sri (yes, we've been there already a few years back)
19:49 sri oh, i just posted the same ;p
19:50 sri well... not really...
19:50 jberger doesn't hurt to show that we are discussing
19:51 sri i actually kinda announced an imminent 4.0 release on twitter
19:51 sri we just happened to make a few more breaking changes afterwards
19:51 jberger and I had my "are you excited" tweet and blog post
19:51 sri oh right
19:51 sri might be worth bringing up as well
19:51 jberger yeah, the render_* were pretty late in the game
19:53 jberger on a different topic: there are times when I just want to shout: LEARN SOME PERL BEFORE ASKING: http://stackoverflow.com/questions/16637333/c​onvert-string-in-to-hash-in-perl-using-split
19:53 jberger I know we were all noobs once, I was
19:53 jberger recently enough
19:54 jberger but saying "I'm new to Perl" doesn't excuse "I choose not to read about split or regexen"
19:54 jberger grrr
19:54 * jberger just needed to vent, sorry
19:58 sri look at that, jberger is not a list admin yet
19:59 denisboyun joined #mojo
19:59 * jberger barely uses the list
19:59 * sri chains jberger to the list
20:09 sri meh, koorchik is not listening at all
20:10 sri render_* method semantics would have changed as well, it was no clean removal :S
20:11 sri think i'll just leave the discussion, repeated myself way too much already
20:14 * sri runs away screaming
20:16 sri tempire++ # classy topic
20:22 KindOne joined #mojo
20:25 gtodd left #mojo
20:25 gtodd joined #mojo
20:28 jberger I think that actually making a policy decision now is a bad idea
20:28 jberger we can revisit in a few months and assess coolly
20:28 jberger right now we are likely to swing too far one way or the other, depending on who is loudest
20:43 * sri agrees
20:44 * sri never considered that people might be ok with announcement only deprecations
20:44 sri "Mojolicious::Controller::render is going away in 3 months"... BOOM!
20:45 sri "In 3 months we are going to apply branch jberger-rocks, which breaks a few dozen things"... BOOM!
20:47 * sri will make sure to try it some time now that he knows >:->
20:47 KindOne joined #mojo
20:47 jberger honestly I think that sounds much more evil
20:48 jberger deprecation path with warnings is good
20:48 jberger once a year, people can wait to upgrade
20:48 jberger and, you know, test things first
20:48 jberger :-)
20:49 sri we are kind of low on things left on the todo list
20:50 tempire deprecation policy conversations always seem to boil down to, "I want cpanm to manage my application dependency issues"
20:50 jberger aka, why Carton exists :-P
20:51 * jberger forces sri to work on Mango!
20:51 * sri works on Mango
20:51 jberger sri: I liked that Fowler talk you posted
20:52 sri i'm actually a little stuck with mango, still the connection pool
20:53 jberger nothing groundbreaking in the talk, but good to hear that I wasn't crazy that the "schema" question is just kicked down the road a little
20:53 sri not even sure anymore which direction to go, fixed pool size with queue or dynamic pool with maximum connections kept alive
20:53 jberger and I liked seeing the idea of aggregates as things, which is the "default" for nosql
20:53 sri Fowler is great
20:54 jberger sri: is it very hard to support both types of pool? is there some way to abstract that detail?
20:54 sri yea, pretty hard
20:54 * jberger guesses there probably isn't or it wouldn't be a problem
20:54 sri the queue changes everything
20:55 jberger * probably IS
20:55 sri worst is that there are two levels of queues
20:56 * jberger flips a coin
20:56 sri one for normal operations and one for priority operations like getnonce and authenticate
20:56 jberger in the end, can't you emulate a fixed pool queue with dynamic pools?
20:57 sri the other way around would be easier
20:57 jberger probably costly to use them like that
20:58 jberger my knowledge of the theory of this stuff is limited
20:58 jberger I too often fall back to the convar idea, its the one I understand the best
20:58 sri the big question really is if you say $mango->max_connections(5) what happens to the 6th parallel operation?
20:58 * jberger should probably study this stuff more
20:59 jberger meaning, should it wait or ... what?
20:59 sri does it get a 6th temporary connection, or does it wait in a queue for its turn
20:59 jberger I would think it would wait in a queue
21:00 sri Mojo::UserAgent gives it a 6th connection
21:00 sri connection limit there is only a keep alive queue
21:00 sri of course it is used very differently most of the time
21:01 sri who connects to more than one server with Mango
21:01 jberger that almost make it sound like min_connections IMO
21:01 sri the definition there is "max connections to keep alive"
21:01 jberger yeah, I think connecting to one database is far and away the most common
21:03 jberger my admittedly uneducated thought process would be "max number of worker available, though they don't all have to be acting if they aren't needed"
21:04 sri not all mongodb operations are fast, if you use it as a pub/sub system you could end up waiting a long time if the limit is too low
21:05 jberger like I said, I don't know any of the details
21:05 jberger :-)
21:05 sri node.js driver uses a hard connection limit though
21:05 jberger you could make the default nice and high
21:07 * jberger trusts sri
21:09 sri "Docker is a rewrite of a 2 year old python codebase which has seen a lot of action in production. We wanted the operational learnings but not the spaghetthi code." https://news.ycombinator.com/item?id=5732413
21:09 sri haha!
21:10 sri Go sure gets a lot of love
21:12 KindOne joined #mojo
21:24 KindTwo joined #mojo
21:32 lukep joined #mojo
21:53 whitebook joined #mojo
21:55 KindTwo joined #mojo
22:02 mattastrophe joined #mojo
22:19 KindTwo joined #mojo
22:28 trone_ joined #mojo
22:30 whitebook joined #mojo
22:31 sri oh... mr tempire https://metacpan.org/source/COOLMEN/Test-​Mojo-More-0.02/lib/Test/Mojo/More.pm#L234
22:31 sri oops
22:31 sri wrong link
22:31 sri https://metacpan.org/source/TEMPIRE/Moj​olicious-Plugin-ConsoleLogger-0.05/lib/​Mojolicious/Plugin/ConsoleLogger.pm#L39
22:33 sri Test::Mojo::More is the worst offender when it comes to private api use though
22:34 jberger is there an approved way to check for a static response?
22:34 * jberger looks
22:34 sri not really
22:34 sri but you can emulate it yourself
22:34 jberger should there be?
22:34 jberger ->is_static?
22:34 sri hook after_static => sub { shift->stash->{my.static}++ };
22:35 jberger yeah, thats true
22:35 sri we can't just add methods to Mojolicious::Controller
22:35 jberger oh right
22:35 Akron joined #mojo
22:35 jberger :-)
22:36 jberger 5.0!!!!
22:36 sri some using mojo.routed or mojo.finished are pretty bad
22:36 sri mojo.secret is an interesting case
22:36 sri we use that stash key so embedded apps can inherit the secret from the main app
22:37 Akron sri: I just read your response to the mailing list and I was wondering: Is it offending to use private API? I just thought, you are on your own if you do it?
22:37 jberger Akron, true in your own app. Naughty in CPAN modules
22:37 sri you are totally on your own if you do it
22:38 sri in fact, if you do it on cpan i hate you! :o
22:38 jberger sri: mojo.secret suggests a possible solution
22:38 jberger mojo.public.*
22:38 sri -.-
22:38 Akron 8-O
22:40 Akron In fact I don't know how private Mojolicious::Routes::Route is ... I might violate that with Mojolicious-Plugin-Util-Endpoint somehow - but as I need to fix it, I will take care of it.
22:40 sri oh,  just noticed that Test::Mojo::More is using mojo.secret wrong anyway
22:41 sri Akron: looks very public to me
22:41 sri class with a documented api is public
22:42 sri when i say private i mean underscore methods and mojo. keys and the like
22:42 Akron Okay - thanks. However - I guess I have to remove given/when now ... so releasing 4.0 now was a good match for fixing broken plugins anyway. ;)
22:42 sri haha ->pattern->tree is walking the line though ;p
22:43 sri the data strucutre itself is documented as being dynamic
22:43 sri so it may change with optimizations
22:44 Akron That's fine. cpantesters will inform me about that ... I will have a look how to fix that tomorrow.
22:44 sri not very likely, but something to be prepared for
22:45 sri it's the same with the Mojo::DOM and Mojo::Template trees
22:45 sri html healing just needs to get a little better and suddenly you have a totally different tree
22:45 KindOne joined #mojo
22:46 Akron sri: I am misusing the DOM-tree in a cpan release. I am aware of that. In fact, as far as I remember I used it in a way you told me back then.
22:47 Akron But the release is marked with bold words of an experimenatl status, so I guess I am okay. ;)
22:50 BeDa joined #mojo
22:50 jberger Akron, which module needs DOM tree?
22:51 Akron jberger: XML::Loy
22:51 jberger I haven't needed to use it since the next/previous methods were added
22:51 jberger the tree that is
22:52 jberger oh, but yours is definitely more involved than I usually get
22:52 Akron You are right - I should have a look at it, if my direct accesses are now obsolete.
22:53 sri next/previous are very recent additions
22:55 sri also text_before/text_after
22:56 sri http://mojolicio.us/perldoc/Mojoli​cious/Guides/Cookbook#Web_scraping # last example here is pretty nice
22:56 sri extracting text + alt attributes
22:56 sri (if that's the kind of thing you used the tree for)
22:57 Akron No - in fact it was more of "creating methods that are namespace aware" etc.
22:58 Akron text_before and text_after is more for dealing with mixed content, which is rather rare in the scope of XML::Loy.
22:59 sri ah

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