Camelia, the Perl 6 bug

IRC log for #mojo, 2012-04-16

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

All times shown according to UTC.

Time Nick Message
00:51 mattastrophe joined #mojo
01:02 scott joined #mojo
01:31 asarch joined #mojo
01:48 bird joined #mojo
02:48 lbr joined #mojo
03:28 asarch joined #mojo
03:34 trexy joined #mojo
04:06 mire joined #mojo
04:18 Leandr joined #mojo
04:38 Foxcool__ joined #mojo
04:54 l3l1p joined #mojo
05:01 * sri yawns
05:23 Leandr joined #mojo
05:26 plu sri: morning, early bird :)
05:38 Leandr joined #mojo
05:54 * tempire crows
05:57 sri good morning fellow ponycorns
06:12 GitHub172 joined #mojo
06:12 GitHub172 [mojo] kraih pushed 1 new commit to master: http://git.io/2Ek0Vg
06:12 GitHub172 [mojo/master] improved regex formatting in routes command - Sebastian Riedel
06:12 GitHub172 left #mojo
06:53 baton8 joined #mojo
06:54 GitHub61 joined #mojo
06:54 GitHub61 [mojo] kraih pushed 1 new commit to master: http://git.io/fDdSrQ
06:54 GitHub61 [mojo/master] removed a few easter eggs - Sebastian Riedel
06:54 GitHub61 left #mojo
07:01 sri no more shagadelic
07:05 tempire it had its time
07:05 tempire it was good
07:06 fhelmber_ joined #mojo
07:09 snailfromhu joined #mojo
07:18 batman joined #mojo
07:23 Htbaa joined #mojo
07:26 spleenjack joined #mojo
08:12 Vandal joined #mojo
08:14 hugh joined #mojo
08:17 mattastrophe joined #mojo
08:20 AmeliePoulain joined #mojo
08:25 trexy left #mojo
08:29 GitHub75 joined #mojo
08:29 GitHub75 [mojo] kraih pushed 1 new commit to master: http://git.io/q4BwoA
08:29 GitHub75 [mojo/master] small optimizations - Sebastian Riedel
08:29 GitHub75 left #mojo
08:29 sri tempire: would you mind sneaking in a quick followup release? i accidentally a command in 2.81 :/
08:30 sri no tweet
08:30 stephan48 oh no!
08:31 * sri really needs to be more careful before we run out of numbers :o
08:32 sri on the plus side, we are now also below 10k lines of code in sloccount
08:33 sri 9998 lines in sloccount and 9844 lines in cloc
08:37 Htbaa about the versioning, are 1.00 and 2.00 considered stable, and the rest are development releases?
08:38 sri Htbaa: nope
08:38 sri dev releases would follow the cpan convention
08:38 Htbaa so every normal release is considered stable then?
08:38 sri 1.0, 2.0 and 3.0 are major releases that are allowed to break things, maint releases in between need to deprecate
08:39 sri it's all in the coding guidelines
08:40 Htbaa on this system I had Mojolicious 2.69 installed. When I did a mojo version it said "Thanks for testing a development release, you are awesome!". The 2.81 release says "This version is up to date, have fun!"
08:40 Htbaa ah
08:40 Htbaa will have a look
08:40 sri oh
08:40 sri version command broken?
08:41 Htbaa no clue, that's why I thought only 1.00 and 2.00 were considered stable :P
08:41 sri doesn't look like it, maybe it was in 2.69 -.-
08:41 sri i just changed the version here to 2.69 and it said "You might want to update your Mojolicious to 2.81."
08:42 Htbaa I'll check my laptop as well, but I believe that one was already up to date
08:43 * sri shakes tempire
08:47 Htbaa laptop already had 2.81
08:47 Htbaa but thanks for the clarification about the version numbers
08:48 sri you should only see the development release text when you install from github
08:49 sri it makes an api request against metacpan to see what's the latest cpan version
08:50 Htbaa so if my version is outdated it does give that message as well?
08:50 sri then it tells you to update
08:50 Htbaa so far I've never seen that message, only that it's a dev release. And I've never installed from Github
08:51 sri what does it tell you for 2.81?
08:51 Htbaa for 2.81 it says it's up to date
08:51 sri https://github.com/kraih/mojo/blob/maste​r/lib/Mojolicious/Command/version.pm#L52
08:51 sri up to date is right
08:51 sri those are the possible options btw
08:52 Htbaa hmm, I really don't recall installing a dev release though
08:52 sri maybe it was just broken back then
08:53 Htbaa installed 2.71 and it does give the proper message
08:53 sri or the metacpan indexer skipped a release
08:53 sri metacpan is known to have hiccups every now and then
08:57 Htbaa if I see it again I'll report it again. It's all fine now
09:12 Leandr joined #mojo
09:13 bird joined #mojo
09:17 yko joined #mojo
09:42 yko joined #mojo
10:06 memowe SHAGADELIC!
10:06 memowe wait...
10:06 memowe :/
10:06 memowe sri: \of (gets better!)
10:07 sri \o/
10:07 memowe \of
10:07 * memowe looks like LaTeX.
10:08 daxim joined #mojo
10:19 xoma joined #mojo
12:26 azawawi joined #mojo
12:47 asarch joined #mojo
13:03 mire joined #mojo
13:08 sri hmm
13:09 sri think i'm finally going to deprecate waypoints
13:09 d4rkie joined #mojo
13:09 sri crab won't like it but i don't think they can be saved
13:15 jnap joined #mojo
13:22 ver joined #mojo
13:22 memowe \of
13:23 Foxcool joined #mojo
13:33 CromeDome joined #mojo
13:38 sri i've tried fixing format detection bugs with waypoints, but it's just impossible
13:38 GitHub69 joined #mojo
13:38 GitHub69 [mojo] kraih pushed 1 new commit to master: http://git.io/V-s9og
13:38 GitHub69 [mojo/master] deprecated waypoint support in Mojolicious - Sebastian Riedel
13:38 GitHub69 left #mojo
13:40 d4rkie joined #mojo
13:47 sri and while we are at it, i'll take care of relaxed placeholders too :o
13:56 sri good thing tempire was sleeping :)
13:56 sri 2.82 will need a tweet after all ;p
14:02 GitHub197 joined #mojo
14:02 GitHub197 [mojo] kraih pushed 1 new commit to master: http://git.io/Kw-emA
14:02 GitHub197 [mojo/master] deprecated relaxed placeholders "/(.foo)" in favor of "/#foo" - Sebastian Riedel
14:02 GitHub197 left #mojo
14:04 * sri pokes tempire, marcus and crab
14:12 Leandr joined #mojo
14:17 sri i'm not completely happy with the # placeholder character, but imo it's good enough
14:18 sri choices were ? and #
14:19 stephan48 why you done it then?
14:19 sri to bring semantics in line
14:19 sri all placeholder work exactly the same now
14:19 sri no special cases
14:20 sri once semantics are in line changing the character itself is absolutely trivial in an upcoming major release
14:21 sri and we will have feedback if people don't like #
14:29 GitHub21 joined #mojo
14:29 GitHub21 [mojo] kraih pushed 1 new commit to master: http://git.io/RVE1Hw
14:29 GitHub21 [mojo/master] small optimizations - Sebastian Riedel
14:29 GitHub21 left #mojo
14:34 sri hmm
14:35 sri now that waypoints are gone we can also consider format constraint inheritance again
14:35 stephan48 was was the problem with waypoints?
14:35 sri as in "under [format => ['html', 'txt']];"
14:35 sri they just didn't fucking work
14:35 stephan48 k
14:36 sri format handling was unfixable for waypoints
14:40 sri should anyone care enough to actually try and fix it, i will gladly explain the gory details
14:47 sri hmm
14:47 sri maybe /?foo is better than /#foo
14:47 sri since it's a regex quantifier like *
14:50 GitHub136 joined #mojo
14:50 GitHub136 [mojo] kraih pushed 1 new commit to master: http://git.io/ujzYoQ
14:50 GitHub136 [mojo/master] small documentation tweak - Sebastian Riedel
14:50 GitHub136 left #mojo
15:24 ispy_ joined #mojo
15:24 memowe and like :
15:24 memowe ...
15:25 sri ;p
15:25 memowe \of
15:25 sri looking at both, i think i prefer /#foo
15:25 * memowe too. Looks better.
15:25 memowe More "boxy".
15:26 sri ideal candidate would be /+foo but + is too common in paths
15:27 sri actually, even if i could choose a char freely, i'm not sure which one would be better
15:28 sri all the good ones are too common in paths
15:28 plu joined #mojo
15:48 athenot joined #mojo
15:52 batman joined #mojo
15:53 Psyche^ joined #mojo
16:01 mercutioviz joined #mojo
16:16 sri guess i'll release on my own then :o
16:17 garu sri: I'm here :D
16:18 sri basically, waypoints are completely broken and imo not repairable
16:18 garu sri: I never really understood waypoints anyway
16:18 baton8 joined #mojo
16:18 garu but I know they've been around forever
16:18 sri yea, nobody really used them, and the bugs kept piling up with every new routes feature
16:19 garu right
16:19 sri now that format detection actually works there are some bugs that just can't be fixed
16:20 garu yeah, having a feature with so many exceptions (it works, except for this and this and this...) doesn't seem right
16:20 sri relaxed placeholders were supposed to get their own character for a long time
16:20 garu really? that's the one that intrigued me the most
16:20 garu I mean, why replace . for #?
16:20 sri so they can have the exact same semantics as the other placeholders
16:20 tempire as long as relaxed placeholders continue to work through 2.x
16:21 tempire 'cause that would invalidate mojocast 2
16:21 tempire :(
16:21 garu tempire: yeah, I'll have to patch a bunch of code as well
16:21 sri make a counter proposal
16:22 sri just keep it a special case forever?
16:22 garu we also need to make sure docs are updated... even keeping the old information around older blog posts, for example, might confuse newcomers
16:23 sri how is that different from any other change?
16:23 garu I don't know... what do you mean by "so they can have the exact same semantics as the other placeholders"?
16:23 garu I thought they did already :X
16:23 tempire it's hard to say without having an idea of what the new routes are going to be.
16:23 garu ah
16:23 sri relaxed placeholders have a special exception that makes them require quoting
16:23 garu new routes coming up, I see
16:23 snailfromhu joined #mojo
16:23 sri actually...
16:24 sri i'm doing all this to prevent a routes rewrite
16:24 tempire maybe I should look at the routes code, then.
16:24 garu what about all the people that use '#' in their paths? won't the routes get confused by this?
16:25 sri ok, tell me which one looks different in this line "/:foo", "/(.foo)", "/*foo"
16:25 sri and now this line "/:foo", "/#foo", "/*foo"
16:25 sri garu: think about that for a moment :)
16:26 sri # in their paths?
16:26 garu oh, I agree that if you have to add parenthesis each time then it needs to be changed
16:26 garu yeah
16:26 sri # is an invalid character
16:26 sri it does not work and is completely safe to use in routes
16:26 garu maybe I'm just confised, having just woke up
16:26 garu twitter comes to mind
16:26 sri that's not a path
16:26 sri that's a fragment
16:26 garu ah
16:26 garu right
16:27 sri # and ? are safe characters to use
16:27 garu hmm
16:28 garu yeah, I got it... I was a bit confused but it's like saying asking for "foo.html" would confuse the "." placeholder
16:28 sri breaking the mojocast is unfortunate, but i really want to fix routes and prevent the redesign
16:29 tempire just so the relaxed placeholders continue to work in the short term.  eventually the mojocasts will have to be updated - I'm ok with that. just not right at the moment.
16:29 sri at least 3 months
16:30 sri more on request
16:31 sri it's a really trivial deprecation, so i'm not in a hurry to remove it https://github.com/kraih/mojo/commit/ae31f​33280e2518218c922f65e9c841a1236d73e#L2R206
16:31 sri waypoints are much much worse
16:32 garu sounds good to me, it will give us time to skim through some "official" posts/examples and make sure they're updated as well. Regarding "3 months, more on request", how about keeping them on throughout 2.xx (if it takes longer than 3 months, I mean)?
16:32 garu (with the deprecation warning)
16:33 sri we'll run out of version numbers before that ;p
16:33 garu yeah, I figured :)
16:34 xaka joined #mojo
16:34 sri all core examples are already updated
16:34 sri and i doubt it's used very commonly
16:35 sri if only for it's quoting requirement
16:36 tempire I don't personally care about waypoints
16:36 tempire never used them
16:36 sri crab might be the only one :D
16:36 tempire most folks thinks they're confusing.
16:36 garu granted, I for one use :foo a lot more than (.foo), but still, that's not to say I won't have to patch a lot of code :)
16:36 sri because they are
16:37 sri you could instead just use two routes
16:37 * garu never even understood waypoints completely
16:38 sri waypoints were actually not designed by me, but added on request
16:38 tempire that sounds out of character, for so early in the mojo dev phase
16:39 sri happens a lot actually
16:39 sri i accept a lot more stuff early on
16:39 tempire I thought you had changed that policy with mojolicious as to prevent stuff like detach
16:40 sri for stuff i knew how to do, but routes were a new concept
16:42 sri you might remember the wonderful /(((foo))) placeholders :D
16:44 sri /(.foo) was actually not bad in the beginning, since /(*foo) also had the quoting requirement
16:45 sri tempire: speak up now if you want to veto the /#foo change
16:46 * tempire wonders if it will mess with syntax highlighting
16:46 sri # is not a special character
16:47 sri you're already using it for css3 selectors :)
16:47 tempire ah, right.  since it's within a string.
16:50 * sri wonders where marcus hides
16:52 sri basically, my point of view is that if you want to keep "/(.foo)", it should be forever, and you should have a good explanation for its semantics we can present to beginners
16:54 sri it's the middle placeholder, it kinda sucks that it has to be special
16:54 sri if it was the first or last in line explanations would be easier
16:56 sri main reason relaxed placeholders were not explained in the lite tutorial and put at the end of the routing guide
16:56 sri you had to search for examples
17:00 inokenty joined #mojo
17:13 * sri does notice that nobody actually agreed with the change
17:14 GitHub126 joined #mojo
17:14 GitHub126 [mojo] kraih pushed 1 new commit to master: http://git.io/RJIL6Q
17:14 GitHub126 [mojo/master] reverted deprecation for Mojocasts - Sebastian Riedel
17:14 GitHub126 left #mojo
17:15 sri oh well, i suppose the only way to change routes is a complete redesign
17:17 sri maybe deprecate and remove it if there's no solution everybody is happy with
17:40 tempire sheesh
17:40 tempire I'm ok with the change.
17:41 tempire my silence is implied acquiescence.
17:42 sri :S
17:43 GitHub133 joined #mojo
17:43 GitHub133 [mojo] kraih pushed 1 new commit to master: http://git.io/c30mYw
17:43 GitHub133 [mojo/master] Revert "reverted deprecation for Mojocasts" - Sebastian Riedel
17:43 GitHub133 left #mojo
17:44 sri http://i.qkme.me/35n5rb.jpg
17:49 GitHub117 joined #mojo
17:49 GitHub117 [mojo] kraih pushed 1 new commit to master: http://git.io/6w0BPQ
17:49 GitHub117 [mojo/master] documentation tweak - Sebastian Riedel
17:49 GitHub117 left #mojo
17:50 sri tempire: your release
17:53 eamonnw ls
17:54 ispy_ joined #mojo
18:27 kmx sri: FYI just released IO::Socket::SSL 1.65 has NPN support (requires Net::SSLeay 1.46+ and openssl 1.0.1)
18:28 sri \o/
18:28 sri kmx++
18:37 kmx sri: my sample code was not included into pod doc but the usage is very simple - check https://gist.github.com/2400559
19:06 GabrielVieira joined #mojo
19:09 sri already seen the example, really easy
19:09 Leandr joined #mojo
19:23 xoan joined #mojo
19:34 marcus hey
19:34 sri IT'S ALIVE!
19:35 marcus sri: missed me? :)
19:35 sri maybe a little :)
19:36 sri now be a good viking and read the backlog ;p
19:36 marcus I read the part that said 'where is marcus hiding'.
19:37 marcus and I figure you've been hacking on the routes
19:37 sri yea, big deprecations
19:38 sri release 2.82 if you agree ;)
19:40 sri routes should be pretty consistent now
19:45 GitHub37 joined #mojo
19:45 GitHub37 [mojo] kraih pushed 1 new commit to master: http://git.io/eNtmkA
19:45 GitHub37 [mojo/master] added relaxed placeholder recipe to tutorial - Sebastian Riedel
19:45 GitHub37 left #mojo
19:46 sri last time the tutorial had a relaxed placeholder section was when i still used 4 space indent :o
19:47 marcus $r->get(/foo#bar') ...
19:47 marcus hmm
19:47 marcus $r->get('foo#bar')->to('foo#bar') even
19:47 sri it's not like there are alternatives
19:47 sri that's atually valid :)
19:48 marcus yeah, I know.
19:48 marcus I guess & can be used in paths as well_
19:48 marcus ?
19:48 sri yea
19:49 sri or actually
19:49 * sri checks again
19:49 sri perl -Mojo -E'a("/foo&bar" => {text => "works :("})->start' get '/foo&bar'
19:50 marcus :-/
19:50 sri that doesn't actually look good either
19:51 sri i don't really mind the double foo#bar
19:51 sri aside from the impression of repeating yourself it's not ugly
19:52 marcus $ is reserved ?
19:52 sri perl -Mojo -E'a("/foo\$bar" => {text => "works :("})->start' get '/foo$bar'
19:52 marcus http://en.wikipedia.org/wiki/Percen​t-encoding#Types_of_URI_characters
19:53 sri that doesn't mean that much
19:53 sri chars that get through vary from server to server
19:54 marcus mm
19:54 sri and we decode % encoded ones too
19:54 marcus guess it's only reserved in certain contexts of the url
19:54 sri that too
19:54 sri for path i believe ? and # are the only safe options
19:55 sri my first choice would have been /+foo
19:55 sri but + is common
19:56 sri $ % and @ would be annoying to escape btw
19:56 marcus yeah, true
19:56 marcus I just don't like how overloaded with meanings # is
19:57 sri overloaded?
19:57 marcus controller / action separator, client side url separator
19:57 marcus and now relaxed placeholder
19:58 sri and perl comments ;p
19:58 marcus that too
19:58 marcus are you sure you're actually solving a potential problem with this change?
19:58 marcus or is it only for your sense of asthetics?
19:58 sri i'm solving a problem
19:59 marcus ok
19:59 sri special casing relaxed placeholders makes custom routes patterns impossible atm
19:59 sri you may have noticed that patterns have attributes for tokens
20:00 marcus I have to say I haven't :(
20:00 sri if semantics are cleared up you could have a plugin that defines shortcuts with custom patterns
20:00 marcus that sounds vaguely useful to me.
20:01 sri $r->myroute('/$foo/@bar')...
20:01 marcus hmm
20:01 sri https://github.com/kraih/mojo/blob/mast​er/lib/Mojolicious/Routes/Pattern.pm#L6
20:02 sri you could go more crazy with custom parsers
20:02 marcus we've been building up routes using my $foo=$r->route(...) and then $foo->route(...) and so on a lot
20:02 marcus think this would be better?
20:02 sri unrelated i think
20:02 marcus ok, maybe I don't fully understand this. it seems complicated.
20:03 sri and the old relaxed semantics make it even more complicated ;)
20:04 marcus mkay
20:04 marcus let's kill it then
20:05 sri like i said before, even if i had free choice for relaxed chars, i wouldn't know a better one than #
20:05 marcus guess that is a good sign.
20:07 sri or does this line look right to you? "/:foo", "/(.foo)", "/*foo"
20:08 sri compared to "/:foo", "/#foo", "/*foo"
20:12 inokenty "/(.foo)" is easy to understand that it captures dots, I think
20:12 inokenty And "/#foo" looks more clear and modern
20:13 inokenty But meant nothing about dots
20:13 sri that's not so important
20:13 inokenty Okay
20:14 sri :foo is not obvious either unless you read the docs
20:14 xoan left #mojo
20:16 sri that you consider /#foo modern is interesting
20:16 sri wonder if that perception is common
20:17 sri browser heavy apps abusing fragments for client side paths might cause it
20:17 sri another argument for using /#foo :)
20:18 sri $r->get('/#controller/#action/#id');
20:19 marcus phear
20:20 sri vs $r->get('/(.controller)/(.action)/(.id)');
20:22 sri vs $r->get('/+controller/+action/+id');
20:22 sri that was my ideal choice, and it doesn't look much better
20:22 sri vs $r->get('/:controller/:action/:id');
20:22 sri it's good : is the default placeholder
20:23 GabrielVieira joined #mojo
20:23 sri vs $r->get('/:controller/:action/#id');
20:23 sri i suspect those will be the most common
20:23 sri vs $r->get('/:controller/:action/(.id)');
20:25 sri vs $r->get('/:controller/:action/?id');
20:25 sri vs $r->get('/?controller/?action/?id');
20:27 sri maybe it's just my consolas font, but # looks pretty good in context
20:38 sri http://mojolicio.us/perldoc/Mojol​icious/Lite#Relaxed_Placeholders
20:38 sri example looks good too
20:43 * sri will wait a little and then release 2.82 if nobody vetos
20:43 sri marcus: btw. # is also a css3 selector ;p
20:46 inokenty I agree with the new changes. $r->get('/:controller/:action/#id') looks better than others
20:53 sri oh
20:53 sri there is of course another option
20:53 sri $r->get('/:controller/:action/%id')
20:54 sri vs $r->get('/%controller/%action/%id')
20:54 athenot_ joined #mojo
20:54 D4RK-PH0ENiX joined #mojo
20:54 graf joined #mojo
20:54 sri since routes are unescaped already
20:54 hide_ joined #mojo
20:55 ask joined #mojo
20:55 sri hmm
20:55 tempire joined #mojo
20:55 robinsmidsrod joined #mojo
20:56 ashleydev joined #mojo
20:56 sri perl doesn't interpolate a hash
20:56 sri so it's an option too
20:58 inokenty But we cannot use dots in subroutine name. Or you want to replace ":" by "%"?
20:58 sri wat?
21:00 inokenty %action (instead of :action) means that action can contain a dots, right?
21:00 * sri doesn't follow
21:01 sri anyway
21:01 sri tempire, marcus: what about %?
21:02 sri downside, there might be syntax highlighting issues with "/%foo"
21:02 sri textmate for example doesn't know perl won't interpolate those
21:03 marcus I say go for #
21:03 sri word
21:03 marcus it seems to be the worst alternative apart from all the other alternatives.
21:04 sri lets do this
21:12 xaka joined #mojo
21:19 GitHub70 joined #mojo
21:19 GitHub70 [mojo] kraih pushed 1 new commit to master: http://git.io/cfSiTA
21:19 GitHub70 [mojo/master] fixed small bug in cpanify command - Sebastian Riedel
21:19 GitHub70 left #mojo
21:28 sri phew, finally over
21:43 JerJer joined #mojo
21:43 lukep joined #mojo
21:46 JerJer1 joined #mojo
22:10 brambles joined #mojo
22:53 GabrielVieira joined #mojo

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