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

IRC log for #mojo, 2014-10-01

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

All times shown according to UTC.

Time Nick Message
00:06 woz joined #mojo
00:09 sri hah!
00:09 sri found the blog post that helped me really get joins http://blog.codinghorror.com/a-visual-explanation-of-sql-joins/
00:09 sri also used venn diagrams :)
00:13 sri looks like i might be buying the art of sql
00:13 sri "The relational model is thus named, not because you can relate tables to one another (a popular misconception), but as a reference to the relationships between the columns in a table."
00:13 sri WOOOOSH
00:14 sri already learned something
00:14 mst it taaught me a lot about how to think about databases
00:14 mst a lot of the specific advice was more oracle-ish, but the thinking patterns were way useful
00:15 sri the chapter titles are a bit annoying
00:16 mst I read it cover to cover as stream of consciousness from somebody who was better at thinking about databases than I was
00:16 mst on that basis, it worked really well for me
00:16 mst beyond that, YMMV etc.
00:23 sri ah
01:03 genio I've just never found basic SQL to be complicated.  joins, group by, having, triggers, views, etc.  I get more hung up on the other details, proper backups, replication/sharding, etc.
01:04 genio When the whole nosql thing started becoming a thing, it honestly baffled me that it was needed
01:06 mst right
01:06 mst I mean, DBIC is more 'sql metaprogramming' than anything else
01:06 mst SQL is good at what it does, FFS
01:07 woz joined #mojo
01:36 noganex joined #mojo
01:47 dvinciguerra_ joined #mojo
02:04 D4RK-PH0ENiX joined #mojo
02:06 noganex_ joined #mojo
02:08 woz joined #mojo
02:12 r0b3rt How about a link on mojolicio.us asking for user quotes like hapijs.com does?
02:54 disputin joined #mojo
03:00 real_escape_string joined #mojo
03:01 basic6 joined #mojo
03:03 sri r0b3rt: once there are enough to have a decent looking page
03:04 sri which imo would be between 10-15, with maybe 1-2 better known companies
03:05 r0b3rt I was thinking that having the text "Using Mojolicious? Let us know" with a link to the wiki would be useful now.  And when you have enough, add the logos/quotes to front page.
03:06 sri i don't think that would work
03:07 sri if you're already using mojolicious odds are you're not reading our frontpage regularly
03:10 woz joined #mojo
03:11 fhelmber_ joined #mojo
03:13 r0b3rt Not sure.  I think the toolbar on the front page is good place to start digging into the docs if I need to find something.  But maybe that is not typical.
03:27 preaction joined #mojo
03:31 real_escape_string joined #mojo
03:33 ceej__ joined #mojo
03:37 sri btw. the art of sql didn't work for me, it's just too wordy
03:38 sri i don't have the necessary attention span reading english text
03:43 jberger Yer should speak English dern it! Murica!
03:48 sri her majesty's english you mean!
04:00 basic6_ joined #mojo
04:11 woz joined #mojo
04:23 Zx3 joined #mojo
04:29 zivester joined #mojo
05:12 woz joined #mojo
05:19 sujithm joined #mojo
05:20 KCL_ joined #mojo
05:24 sujithm joined #mojo
05:24 real_escape_string left #mojo
05:26 d4rkie joined #mojo
05:35 jnbek MURICA!!
06:04 Vandal joined #mojo
06:13 tempire The idea that there's pushback to learning sql and rdms is so strange to me.
06:13 tempire Kids these days.
06:13 purl Get off my lawn!
06:13 * tempire waters his lawn
06:13 woz joined #mojo
07:02 basiliscos joined #mojo
07:12 fhelmber_ joined #mojo
07:15 woz joined #mojo
07:18 denis_boyun joined #mojo
07:24 Lee joined #mojo
07:25 trone joined #mojo
07:27 meredith joined #mojo
07:39 sugar joined #mojo
07:47 fhelmber_ joined #mojo
08:13 Dandre joined #mojo
08:14 Lee joined #mojo
08:16 woz joined #mojo
08:29 alanminter joined #mojo
08:30 alanminter Just had to search for what is this thing Mojolicious Twinkle
08:30 alanminter :P
08:31 alanminter batman: If you have some SemTEX, you can blow up the 'aminter' convos user -- he's stuck in an infinite cycle
08:32 alanminter During registration, the idiot copied the example port number and inserted his convos credentials
08:33 alanminter Apparently now every time he logs in, it says there's a problem and offers (with a button) to take him back, but loops
09:09 trone joined #mojo
09:17 woz joined #mojo
09:23 arthas joined #mojo
10:18 woz joined #mojo
10:19 dp_ joined #mojo
10:24 d4rkie joined #mojo
10:24 dod joined #mojo
10:37 d4rkie joined #mojo
10:45 sujithm_ joined #mojo
11:20 woz joined #mojo
11:37 neilhwatson joined #mojo
11:46 batman alanminter: SemTEX?
11:46 purl hmmm... SemTEX is a pretty good pain medication
11:47 alanminter It's like LaTEX but more explosive
11:47 alanminter http://en.wikipedia.org/wiki/Semtex
11:48 batman alanminter: i think it's ok now. let me know if it's not
12:10 lipizzan joined #mojo
12:21 woz joined #mojo
12:41 ladnaV joined #mojo
12:58 Vandal joined #mojo
12:59 sugar joined #mojo
13:12 zivester joined #mojo
13:13 ryozi joined #mojo
13:23 woz joined #mojo
13:53 Akron joined #mojo
14:03 arthas joined #mojo
14:13 jeff joined #mojo
14:15 arthas joined #mojo
14:16 trone joined #mojo
14:24 woz joined #mojo
15:02 sujithm joined #mojo
15:25 woz joined #mojo
15:45 sh4 joined #mojo
16:11 preaction joined #mojo
16:18 dsterne joined #mojo
16:21 dsterne hello, I'm curious if there's a way I can get a list of the loaded/available templates prior to rendering (my goal is to automatically set up routes depending on templates in a specific path)
16:26 sri interesting
16:26 sri http://www.mikeperham.com/2014/10/01/the-path-to-full-time-open-source/
16:26 woz joined #mojo
16:29 sri especially for perl that might be a viable strategy, considering the lack of high quality choices in many areas
16:32 GillesM joined #mojo
16:36 sri dsterne: that list actually exists internally, but is not exposed with a public api
16:36 basiliscos joined #mojo
16:37 dsterne sri - bummer! thank you for the help. I'll figure out a different way to do it.
16:40 sh4|2 joined #mojo
16:52 sri dsterne: you can always propose it as a new feature
17:03 disputin joined #mojo
17:19 KCL joined #mojo
17:28 woz joined #mojo
17:45 dvinciguerra_ joined #mojo
17:49 * sri wonders how to handle postgres notifications
17:50 sri i could handle them only when the event loop is running, or check for them too every time $sth->execute has been called... hmm
17:51 sri as far as i understand the implementation in DBD::Pg it is very cheap, because notifications are stored client side and the pg_notifies method only checks that queue
18:13 batman i wonder what is most secure: http://user:pass@domain/ or http://domain/verylongprefixthatishardtoguess/
18:13 batman s/http/https/
18:15 mst a lot of browsers will refuse to use user:pass@ links anymore
18:17 batman that's fine. i'm thinking more M2M
18:19 mst M2M?
18:19 purl it has been said that M2M is not a relationship, you need to test against the linker table (the has_many relation) or machine to machine, meaning technologies to exchange informations between machines
18:19 mst right, see, I can only parse that as either many-to-many or some sort of gay hookup scene slang
18:20 mst do you have an expansion that doesn't involve SQL or penis?
18:20 batman "Machine to machine"
18:21 mst so, do you actually mean "HTTP auth versus URL-based auth" ?
18:21 mst and there's never actually a "real" user:pass@ URL involved?
18:22 batman i supposed both are part of the URL, so i have a hard time seeing the difference
18:22 batman but i might miss out on something
18:22 batman not sure what you mean about ""real""
18:22 mst user:pass@ means "when you send this over the wire, turn it into HTTP basic auth"
18:22 bobkare joined #mojo
18:23 mst so the user+pass are in the HTTP headers at that stage
18:24 batman but it doesn't have any additional value..?
18:25 batman i mean the basic authorization isn't actually secret in any way.
18:26 mst I'm not even sure why you're asking this question
18:27 * batman will try something
18:28 sri what looks more correct to you? trying to hide a resource behind an andom public path, or using an actual authentication mechanism to make a resource inaccessible?
18:29 woz joined #mojo
18:29 sri a/n / r/
18:29 sri argh!
18:29 sri s/a/s/
18:30 batman the latter. doesn't make i more secure though (i think)
18:30 * batman need to stop hacking and start watching television...
18:30 sri seriously... use HTTP like it's supposed to be used
18:31 sri get the RESTful web services cookbook or so
18:31 bobkare joined #mojo
18:31 sri you're inventing problems
18:31 mst batman is a certified master of the 'fire, ready, aim' development process
18:31 mst it's like agile but web scale
18:31 mst I hear 10gen use it for MongoDB too
18:34 batman i'm already using this here: http://home.thorsen.pm/private/tree/notverysecret/
18:35 batman i got other directories with a more "random" string than "notverysecret" which i use to share other files.
18:35 batman it's quite convenient...
18:36 sri that still doesn't make it an authentication mechanism
18:36 batman i didn't say i wanted an authentication mechanism.
18:37 denis_boyun joined #mojo
18:37 batman but i guess the question is rather silly. i suppose both mechanisms are just as good when it comes to protecting a resource, without the need for users
18:38 sri if you decide to ignore all http semantics and their implications...
18:39 mst and let caches store your live content
18:39 mst and all sorts of other shit
18:39 sri caching and whatever else is built on top of assumptions about how http works
18:39 mst basically: use basic auth, unless you have a specific reason not to
18:39 mst AFAIK the -only- reason for the /notverysecret/ style is because of the slow demise of browser support for user:pass@
18:39 mst at which point, for server to server stuff, there's no reason at all to use that
18:40 batman that is true
18:41 sri http://www.amazon.com/RESTful-Web-Services-Cookbook-Scalability/dp/0596801688/
18:41 sri lots of wisdom
18:41 basiliscos joined #mojo
18:43 sri (of course i'm a fan of the cookbook style)
18:44 phillipadsmith joined #mojo
18:44 berov joined #mojo
18:44 jalie joined #mojo
18:47 batman https://gist.github.com/jhthorsen/250ed38d80679869accb # i'm thinking about adding a super simple cpanserver for work
18:48 batman it need to be protected because of in house code :(
18:51 jzawodn_ interesting idea
18:51 jzawodn_ we've toyed with that too
18:52 batman i was rather surprised how little code the server required
18:54 batman jzawodn_: do you have a lot of servers?
18:55 jzawodn_ batman: we have a fair number, yeah.  but the main reason is that our build env is firewalled off and cannot directly hit CPAN so getting new modules into our world is annoying.  with an internal CPAN we could make it quite a bit easier
18:55 batman ok
18:56 batman the code above + minicpan + "cpanm --mirror-only --mirror ..." would work
18:57 mst jzawodn_: this is what CPAN::Mini and/or Pinto are for
18:57 mst that or just say 'fuck it' and rsync from a cpan mirror
18:57 batman i think i need to fork ansible-modules-extras and add "--mirror-only" support so i can use ansible to install the modules as well
18:58 mst right, cpanm without --mirror-only is an error for production installs
18:58 disputin joined #mojo
18:59 * batman tries to open an issue first
19:03 jzawodn_ mst: yeah, Pino is probably overkill for us, but the rsync a mirror option is likely what we'd do.  (I've agrued for hosting a public mirror too)
19:03 mst sure, I've never found Pinto to be the right thing for me, but if you want a mixed private+public mirror sometimes it's a good answer
19:04 mst we're going rapidly towards "mst hates everything but hasn't come up with a replacement he hates less yet" territory though :)
19:04 genio hater
19:04 jzawodn_ hah
19:05 jzawodn_ if we didn't alredy have our own home-brewed system for handling internal modules...
19:06 jamesaxl joined #mojo
19:09 batman jzawodn_: i hope i can build the internal system on App::git::ship so i *have to* open source it
19:10 batman we already have a system, but it's not...excellent...
19:10 jzawodn_ ours is just highly specific to our environment
19:10 jzawodn_ ditto with our monitoring system
19:10 batman yeah, no idea why such a system should be specific to any environment :P
19:11 batman the one i've built for work is also specific btw
19:11 fhelmber_ joined #mojo
19:11 jzawodn_ a lot of our stuff pre-dates most of the good open source options, which is a blessing a a curse
19:12 batman hehe
19:12 cfedde Of course it
19:12 purl Of course it is probably the correct one current context would have started drinking kiwi_uk
19:12 cfedde bah.
19:26 jalie joined #mojo
19:26 phillipadsmith joined #mojo
19:28 ua_ joined #mojo
19:28 bjoernfan joined #mojo
19:30 woz joined #mojo
19:34 bobkare joined #mojo
19:47 basiliscos joined #mojo
19:48 jzawodn_ wat
20:10 basiliscos joined #mojo
20:13 Viktor joined #mojo
20:18 bjoernfan joined #mojo
20:31 woz joined #mojo
20:32 batman tempire: grumpify is in it's way back into the App Store :)
20:33 batman No idea how long it takes though...
20:43 batman I will make a tweet when i can find it.
20:47 ua_ joined #mojo
20:50 disputin joined #mojo
20:50 stryx` joined #mojo
20:51 Zx3 joined #mojo
21:14 marty joined #mojo
21:31 denis_boyun joined #mojo
21:32 woz joined #mojo
22:22 disputin joined #mojo
22:29 zgarnog joined #mojo
22:34 woz joined #mojo
23:09 basiliscos joined #mojo
23:11 fhelmber_ joined #mojo
23:26 d4rkie joined #mojo
23:27 D4RK-PH0ENiX joined #mojo
23:35 woz joined #mojo
23:42 davido__ joined #mojo
23:46 zivester joined #mojo
23:57 sri hmmm
23:57 sri postgres notifications are complicated
23:58 sri looks like they are not distributed out of band, but somehow piggyback on other operations

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