The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2015-11-08

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

All times shown according to UTC.

Time Nick Message
00:34 absolut_todd joined #mojo
00:46 aborazmeh joined #mojo
01:18 espent_ joined #mojo
01:33 vervain joined #mojo
01:38 espent joined #mojo
02:02 espent_ joined #mojo
03:32 ZoffixMobile joined #mojo
03:34 jontaylor joined #mojo
03:41 ZoffixMobile joined #mojo
03:47 lluad joined #mojo
05:24 kaare joined #mojo
06:09 irqq joined #mojo
06:37 Kogurr joined #mojo
06:53 stephen joined #mojo
06:58 aborazmeh joined #mojo
07:07 CIAvash joined #mojo
07:37 * nic bijects to the left
08:23 Vandal joined #mojo
09:06 bjakubski joined #mojo
09:48 cpan_mojo Minion-3.03 by SRI https://metacpan.org/release/SRI/Minion-3.03
09:52 vervain joined #mojo
10:07 trone joined #mojo
10:43 asm35 joined #mojo
10:56 absolut_todd joined #mojo
11:28 Shaeto joined #mojo
11:28 robinsmidsrod joined #mojo
11:29 janus joined #mojo
11:33 Shaeto did someone see any example with reactjs frontend and mojolicious as backend ?
11:37 CIAvash joined #mojo
11:49 vervain joined #mojo
11:57 punter joined #mojo
12:02 sdp joined #mojo
12:50 vervain joined #mojo
12:55 moritz joined #mojo
13:03 asarch joined #mojo
13:08 asm35 joined #mojo
13:11 sri postgres 9.5, where are you?!
13:19 Zoffix joined #mojo
13:19 CIAvash joined #mojo
13:27 sri 25% of all websites run on wordpress :O http://w3techs.com/technologies/details/cm-wordpress/all/all
13:29 CIAvash_ joined #mojo
13:33 Zoffix That's pretty scary.
13:34 Zoffix And based on that diagram, 17% of them never upgrade.
13:35 Zoffix So that's about 43,000,000 websites that have potentially known vulnerabilities.
13:36 ajr_ joined #mojo
13:39 sh4 joined #mojo
13:51 vervain joined #mojo
14:19 absolut_todd joined #mojo
14:26 asm35 joined #mojo
14:27 vervain joined #mojo
14:41 punter joined #mojo
15:04 * jberger dreams of a Galileo successor
15:06 sri the market for content manegement systems is so huge, hard to believe perl doesn't really have a serious competitior
15:10 * jberger starts a sketch
15:10 * jberger has a little free time as it happens :s
15:10 Craftsmanship joined #mojo
15:18 PryMar56 joined #mojo
15:22 Zoffix It likely has something to do with deployment. I'd imagine people who use CMSs don't wanna fiddle around installing a bunch of modules from CPAN.
15:22 sri then have a tarball with everything bundled on your site
15:23 Craftsmanship mkdir vendor/
15:23 Zoffix But it doesn't compare with "clicking a few things" on my hosting provider's account page
15:24 sri silly argument
15:24 Zoffix A computer illiterate person can pay $10 to 1and1.com and have a site they built themselves.
15:24 Zoffix No it isn't.
15:24 sri "you're not already dominating the market, so might as well not join"
15:24 sri that's what you said
15:24 jberger Zoffix: that was the whole purpose of galileo
15:25 Zoffix Never even heard of galileo
15:25 jberger fully cpan deployable
15:25 jberger my cms, such as it is
15:25 Zoffix There's Bee comething. That's a competitor
15:25 jberger a good proof of concept, but limiting
15:25 jberger mine was before that
15:25 jberger PearlBee
15:25 sri does galileo have a pretty site yet?
15:25 jberger (the "a" offends me)
15:25 jberger galileo has the same old stie
15:25 jberger site
15:26 jberger it doesn't lend itself to theming, which is one thing I hope to correct in a successor
15:26 Zoffix See, there are Perl CMSs, they're just not the 2 world-leaders.
15:26 moritz hi all; I get this error: http://perlpunks.de/paste/show/563f698d.1fbb.321 any ideas besides trying a different perl?
15:26 jberger galileo-cms.herokuapp.com
15:27 moritz I checked that the template file is valid UTF-8
15:27 sri even googling "galileo cms" doesn't lead me to that site
15:27 jberger sigh
15:27 jberger oh, it isn't responsive, I bet the great mobile recalculation on google killed me
15:28 Zoffix That would only affect mobile searches
15:28 jberger I've been meaning to port to bootstrap 3/4, but it really doesn't seem worth it
15:28 Zoffix I get it as 8th result. The first one being https://github.com/jberger/Galileo
15:28 batman jberger: use a predefined theme. preferably built on top of some framework to make it easy to change.
15:28 jberger I know there are people who use it
15:29 jberger because I get reports etc
15:29 sri of course you'd want to at the very least be listed when googling for "perl cms"
15:29 Zoffix moritz, do you possibly have some code to look at?
15:30 jberger moritz: you should look at the full error stack
15:30 jberger Mojo::Exception is just a wrapper for $@
15:30 chansen moritz: If you can produce a minimal reproducible case, be sure to report it to p5p
15:30 jberger but it has a stack trace embedded
15:36 sri really too bad movabletype crashed and burned, if it wasn't for the mismanagement at the time it could have been market leader
15:38 moritz Zoffix: not really; I can't give you the full repo (not open source), and I don't know what causes it
15:38 moritz jberger: I have no idea how to get the full error stack; the error page only shows that one line from Mojo::Exception::trace
15:38 moritz and even with -MCarp::Always I don't get a backtrace
15:45 jberger moritz: it seems you are still getting error log output, perhaps you could change your log formatter to spit out the trace?
15:48 jberger that might be a handy plugin actually
15:56 gryphon joined #mojo
16:32 cpan_mojo Mad-Mapper-0.06 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mad-Mapper-0.06
17:12 irqq joined #mojo
17:15 jberger any thoughts on doing user passwords in pg via pgcrypto vs say Mojolicious::Plugin::Bcrypt?
17:21 batman i don't know anything about pgcrypto, but bcrypt is what you should do these days.
17:21 jberger batman: yes I like bcrypt and will do that unless I can (safely) push it into the db
17:24 batman jberger: looks like pgcrypto can do many different things, amongs "bf" (blowfish)
17:24 batman which is the only sane crypting algorithm in that list...
17:25 batman at least if the options are bf, md5, xdes, des
17:25 jberger yeah
17:25 * jberger tries to figure out how to tell if the extension is enabled
17:25 jberger the question is moot if I can't tell the user early that the extension is required
17:26 batman i mostly copy/paste the Mojolicious::Plugin::Bcrypt code into my own module and try to tweak the settings
17:28 batman the reason is simply that i don't see why i want to have bcrypt in my controller
17:29 batman (i want it in my model)
17:29 batman https://github.com/Nordaaker/convos/blob/batcode/lib/Convos/Core/User.pm
17:42 kyshtynbai joined #mojo
17:44 jberger I have often thought the same thing about the bcrypt plugin
17:44 jberger but if I can push the hashing into the db, that's a long running db operation and a non-blocking query is a win
17:47 batman indeed. you don't want the hashing to be instant.
17:47 batman if it's instant, then it's not secure, hehe
17:50 Zoffix Won't that block the db?
17:50 Zoffix Um, never mind me actually
17:51 sri block as in shared lock?
17:52 sri anyway, if you want to learn about that stuff here's a good read http://www.postgresql.org/docs/9.4/static/explicit-locking.html
17:53 sri the postgresql process model is preforking btw. like an old apache
17:53 jberger What does checking a password have to do with locking?
17:53 sri so, anything you do in your db session stays in the one worker
17:53 Zoffix It's just the guy who was presenting at the Mojo talk said even when you use Mojo::Pg in a non-blocking way, it blocks the connection, so if you're doing a long operation and your code attempts to access the db, it'll have to wait until the first blocking operation completes
17:53 sri unless database level locks are involved
17:54 jberger Zoffix: yes, but you have more than one connection
17:54 Zoffix Do I?
17:54 Zoffix Oh, there's one connection per worker?
17:54 jberger In mojo pg you do
17:54 Zoffix Is that what you mean?
17:54 sri every connection has a dedicated process on the postgresql side, which can only perform one operation at a time
17:55 jberger As many as you want (and are available)
17:56 sri locks in postgresql are very well done btw.
17:56 sri the whole mvcc thing makes it very elegant
19:26 ChmEarl joined #mojo
19:33 PryMar56 joined #mojo
19:35 mspo joined #mojo
19:35 mspo is there a Mojo::UserAgent that supports timings?
19:36 Grinnz what do you mean?
19:37 mspo I want to add timers like HAR or curl has (network connect, ttfb, download, etc)
19:38 Grinnz oh like print out durations?
19:38 mspo yep
19:38 Grinnz could probably be done, but i dont know of anything built for that so far
19:39 mspo okay thanks
19:39 mspo is there like a big list of everything Mojo:: ?  there's a lot of stuff :)
19:40 Grinnz https://metacpan.org/pod/Mojolicious::Guides#REFERENCE has all the built-in stuff
19:40 Grinnz there's of course many external CPAN modules in the ecosystem
19:40 mspo ohokay
19:43 Zoffix joined #mojo
19:46 * sri doesn't know why Grinnz always links to metacpan
19:46 sri the class list looks nicer on the site http://mojolicio.us/perldoc#REFERENCE
19:47 Grinnz i search metacpan for every perl module
19:47 Grinnz ¯\(°_o)/¯
19:48 Zoffix I don't like mojolicio.us perldoc either :P
19:48 sri nested lists look really awful on metacpan
19:49 sri Zoffix: (╯°□°)╯︵ ┻━┻
19:49 Grinnz i like it, i just am hardwired to do a few certain keystrokes in chrome to get to module docs :P
19:51 Zoffix I don't like the URLs on mojo site. I often copy-paste module names either from code to metacpan URL or off metacpan page and into code.
19:51 Zoffix And the latter doesn't work well on mojo site 'cause of spaces
19:52 Zoffix Plus, I'm used to metacpan :) I still remember resisting using metacpan because I was used to s.c.o lol :P
19:52 sri huh?
19:52 Grinnz http://mojolicio.us/perldoc/Mojolicious::Guides works
19:52 sri http://mojolicio.us/perldoc/Mojo::Pg
19:52 sri ;p
19:52 Grinnz i found that out a while back by accident ;)
19:54 sri think it works because the module name is captured with a wildcard placeholder and we run s/\//::/g against it
19:57 Zoffix I think I'm just used to going to, say, https://metacpan.org/pod/Mojolicious, then clicking some module name in the dist (e.g. https://metacpan.org/pod/Mojolicious::Renderer) and then just going CTRL+L -> End -> Shift+Ctrl+Left arrow a few times to copy/paste module name
19:57 sri only addition i would still like for the pod browser is something metacpan doesn't have either, which is better navigation tools
19:57 Grinnz sri, http://graphemica.com/%F0%9F%8D%BB i wonder if we could embed an emoji font in the perldoc somehow :P
19:58 Grinnz that site works even though my browser normally cant show that symbol
19:58 sri ????
19:58 ivan_ joined #mojo
19:59 sri there's this http://twitter.github.io/twemoji/
20:01 Grinnz ooh, that uses JS to replace the characters with images
20:02 Grinnz probably dont want to rely on JS for pod though :P
20:02 absolut_todd joined #mojo
20:03 sri the fallback is your system fonts
20:03 sri that makes it pretty nice
20:04 Grinnz could just specify the images directly, like they have "this snippet will be translated into the following HTML"
20:04 pink_mist how does selecting the "text" (for copying) work if it's replaced by images? :P
20:04 Grinnz it has alt="the original character"
20:05 pink_mist I've never known that to work in copy/paste
20:05 Grinnz i wasnt saying for copy/paste, i meant for fallback
20:06 Zoffix It does work for copy-paste too
20:07 pink_mist it does? 0_o
20:07 * pink_mist boggles
20:08 Grinnz ???? ???? ???? ???? ???? ???? ???? ???? ???? ✨
20:08 Grinnz apparently, yep
20:14 Zoffix Yeah, alt="" text getting copied worked for ages in browsers. Really annoying when you're trying to copy a large amount of text that has some pics and the author stuffed useless alt="" text into it
20:18 Grinnz sri, it works for all of them except comet and snowman :o
20:18 Grinnz http://twemoji.maxcdn.com/16x16/1f37b.png
21:05 sri WANT! :O http://www.thekeyboardwaffleiron.com/
21:05 Grinnz now thats a waffle
21:06 * Zoffix too
21:06 Zoffix Now I want waffles :/
21:14 Grinnz Zoffix, when doing a =for html, there's no way to get non-html pod to end up on the same line, is there?
21:15 Zoffix Grinnz, why does it need to be on the same line?
21:15 Grinnz i guess that would be at the mercy of the pod renderer anyway :/
21:15 Zoffix Yeah
21:15 Grinnz i'm playing with inserting the emoji images at the start of the lines here https://metacpan.org/pod/Mojolicious#CODE-NAMES
21:16 Zoffix Ahhh
21:16 Grinnz but if i put the whole thing in the =for html, it's completely missing after pod2man
21:16 sri lol
21:16 Zoffix Right, you'd need to versions =for html and =for... text I guess
21:16 sri that doesn't seem worth it at all
21:16 Zoffix s/to/too
21:17 pink_mist s/too/two/
21:17 Zoffix heh right :)
21:17 Grinnz lol
21:27 Grinnz ok this seems to work for pod2html and pod2text/pod2man
21:30 bc547 joined #mojo
21:31 Grinnz so then the question is what to do with  ☃ and ☄
21:32 Grinnz i think font support is pretty good for them but its not like i have a lot of systems to test on
21:32 marcusr o/
21:42 mattastrophe joined #mojo
21:51 ajr_ joined #mojo
21:56 Grinnz http://grinnz.com/stuff/test_twemoji.html
21:57 * Zoffix learns of the draggable="" attribute
21:57 * Grinnz didnt actually look up what that does yet :P
21:58 Grinnz i guess so you dont accidentally drag the image when trying to select text
21:58 Zoffix My guess would be it doesn't let you drag the image with the mouse
21:58 sri draggable="" is almost certainly an error
21:58 Grinnz draggable="false"
22:00 Grinnz https://grinnz.com/stuff/test_twemoji.html also work well under https
22:00 sri if it's a flag you'd use just "draggable" or "draggable=draggable"
22:00 Grinnz i was expecting chrome to complain about cross-domain BS
22:01 Grinnz sri, dunno, its just what that javascript lib was doing
22:01 Zoffix The spec shows to use "true"/"false": http://www.w3.org/TR/html-markup/global-attributes.html
22:01 * sri thinks it is actually draggable=true/false
22:01 Grinnz makes sense
22:03 Grinnz https://github.com/twitter/twemoji#attribution-requirements
22:03 * Grinnz is not sure the best place to put an attribution
22:04 Zoffix Grinnz, there should be a section that mentions jQuery copyright.
22:05 Grinnz yeah, but jquery is bundled
22:05 Grinnz these are just sort of linked
22:06 Zoffix ¯\_(ツ)_/¯ potato pot-ah-to
22:11 Grinnz https://grinnz.com/stuff/test_twemoji.html#Twemoji-Graphics perhaps
22:12 Grinnz ???? 6.0, Clinking Beer Mugs (U+1F37B)
22:12 Grinnz yep copy paste works :P
22:13 sri lol, talking about locking in postgresql, this just popped up in my twitter stream http://blog.nordeus.com/dev-ops/postgresql-locking-revealed.htm
22:13 Grinnz heh
22:16 good_news_everyon joined #mojo
22:16 good_news_everyon [mojo] Grinnz created twemoji (+1 new commit): http://git.io/v80RO
22:16 good_news_everyon mojo/twemoji 3e8bb80 Dan Book: use Twemoji for code names when rendered as HTML
22:16 good_news_everyon left #mojo
22:16 Grinnz git didn't line up that diff very well
22:19 sri "that doesn't seem worth it" -- sri, one hour ago
22:20 Zoffix :P
22:29 Grinnz just wanted to see if i could get a working implementation, i think it looks nice
22:33 sri hmm, there is an inconsistency in minion with automatically retrying jobs
22:33 sri $minion->repair can also transition jobs from active to failed if the worker went missing
22:34 sri but there we can't automatically retry, and they get stuck in the failed state until a human intervenes
22:39 * sri wonders if that's big enough to remove attempts again
23:10 jberger I don't think that's the same scenario
23:10 jberger though, perhaps not enough to care about
23:11 jberger perhaps not DIFFERENT enough
23:14 sri one might argue that a missing worker is more severe than an ordinary job failure... but i don't think that's really true
23:15 sri only real difference here is that one failure happens during processing, and the other during queue maintenance
23:21 sri whole thing would be easier if auto retries were triggered by $minion->repair too
23:21 sri but then the delay would get too big
23:40 sri hmm, i guess there's a solution
23:40 sri the retry logic could be moved into the backend
23:51 sri hmm, i do have a working patch, but the code gets less elegant

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