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

IRC log for #mojo, 2015-07-29

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

All times shown according to UTC.

Time Nick Message
00:49 kaare joined #mojo
00:59 jberger joined #mojo
01:00 Grinnz i stole the blog example into Mojo::SQLite too ;P
01:47 mattastrophe joined #mojo
01:47 mojoex joined #mojo
01:48 meshl joined #mojo
01:48 mojoex hi, its possible to have 1 file with extend. which has extend another file which has extend other files.. etc right ?
01:59 Grinnz mojoex, are you talking about templates? or subclassing?
02:00 Grinnz question is far too vague :)
02:03 Grinnz Dave, i added the ability to specify bind type in Mojo::SQLite, it actually wasn't too bad... maybe it could be added to the others as well https://metacpan.org/pod/release/DBOOK/Mojo-SQLite-0.005/lib/Mojo/SQLite/Database.pm#query
02:03 meshl joined #mojo
02:29 jberger mojoex: http://trout.me.uk/vague.jpg
02:41 PryMar56 joined #mojo
02:52 noganex joined #mojo
03:29 Grinnz its amazing how seamlessly the blog example works even on a prefork server with sqlite... http://grinnz.com:3000
03:31 Dave thanks Grinnz :) got waylaid by a security issue but I'll look at that soon
03:59 zivester joined #mojo
04:36 PryMar56 joined #mojo
04:37 mattastrophe joined #mojo
05:57 dotandimet joined #mojo
06:03 dotandimet joined #mojo
06:10 dotandimet joined #mojo
06:42 mattastrophe joined #mojo
06:50 arpadszasz joined #mojo
07:08 salva joined #mojo
07:08 eseyman joined #mojo
07:09 Vandal joined #mojo
07:09 mattastrophe joined #mojo
07:17 trone joined #mojo
07:21 AndrewIsh joined #mojo
07:43 ladnaV joined #mojo
07:47 inokenty joined #mojo
07:53 teejay joined #mojo
08:51 berov joined #mojo
09:17 dotan joined #mojo
09:19 nic joined #mojo
09:19 jontaylor joined #mojo
09:29 odc joined #mojo
09:38 punter joined #mojo
10:24 arpadszasz joined #mojo
11:56 jberger Grinnz: why wouldn't it?
12:31 human39 joined #mojo
12:37 hernan605 joined #mojo
12:56 zivester joined #mojo
13:01 batman Grinnz: yeah... are you sure that *actually* works? i've had problems with two processes writing to the same sqlite file
13:01 batman problems = the database gets corrupt
13:03 batman i guess it works, since all the processes use dbd::sqlite
13:03 batman https://metacpan.org/pod/DBD::SQLite#DBD::SQLite-and-fork
13:03 batman i could be wrong
13:07 marcus Something is breaking my @ARGV, but I can't find out what :( It works fine in my BEGIN
13:09 batman marcus: i would just move the BEGIN block under each use statement and see when it breaks
13:09 batman what modules are you using?
13:10 marcus Mojo::JSON, Mojo::Util, Mojo::Base, Carp and POSIX
13:12 marcus oh fuck. Found it. a my $self=shift; had dropped outside of a sub.
13:13 batman ;)
13:13 marcus That was such a pain to track down. Thanks for the rubberducking, batman :)
13:14 marcus or eliminating possibilities even
13:14 batman marcus: i'm here for you mr. :)
13:15 stephen joined #mojo
13:16 Grinnz batman, yes, it works :) i have write-ahead log enabled
13:18 batman Grinnz: is that a dbd thing?
13:18 Grinnz it is a SQLite thing
13:18 Grinnz https://www.sqlite.org/wal.html
13:22 hahainternet joined #mojo
13:26 batman Grinnz++ # https://metacpan.org/source/DBOOK/Mojo-SQLite-0.006/lib/Mojo/SQLite.pm#L72
13:27 PryMar56 joined #mojo
13:28 marcus wow, great progress.
13:28 theo joined #mojo
13:28 marcus time to write more tests even :)
13:31 DaniBunny joined #mojo
13:39 jb360 joined #mojo
13:50 bd joined #mojo
13:51 bd is there a way to get notified before morbo reloads the app? i want to s<ve some in memory data
13:57 pink_mist just a wild guess --- when your app->start() returns? maybe
13:58 bd hm
14:01 bd i am using start_app, and if i log after the call I get "did not return an application object"
14:05 jontaylor joined #mojo
14:06 bd hm, I can overload DESTROY in my application class
14:08 jberger bd: END { } block perhaps
14:08 diegok joined #mojo
14:11 y1mmm joined #mojo
14:11 Ptolemarch joined #mojo
14:12 bd jberger: hm, will I have access to the application object there?
14:12 hahainternet joined #mojo
14:12 jberger eh, I'm starting to rethink that suggestion, it probably is insane for several reasons
14:12 jberger why not just use DBM::Deep for your "in memory" data and be done with it?
14:14 Grinnz yeah, i would not rely on anything persisting between requests, that is just going to cause you problems later when you use a prefork server or you just want to restart the server
14:15 Grinnz DBM::Deep is a nice solution if you dont already have a database to use
14:15 Grinnz (or if you just want the data to stay perlish)
14:16 hahainternet joined #mojo
14:33 ajr_ joined #mojo
14:58 Ptolemarch joined #mojo
15:05 meredith_ i don't think i've ever had sqlite files go corrupt when letting more than one process at them, just started to see long lock delays that aborted transactions in the default setup
15:05 Grinnz_ right, thats what WAL mode avoids
15:05 meredith_ working on local files only, though
15:05 Grinnz_ it should lock it properly normally
15:10 meredith WAL seems to make for an all-around improvement of the sqlite experience, which isn't too bad to begin with
15:13 sh4 joined #mojo
15:18 deserted joined #mojo
15:24 jberger joined #mojo
15:30 jontaylor joined #mojo
15:35 PryMar56 joined #mojo
15:49 disputin joined #mojo
16:25 berov joined #mojo
16:41 da5id joined #mojo
16:42 Lucas1 joined #mojo
16:56 sri joined #mojo
17:17 thowe if a form action goes to one method, but it needs to redirect to another method based on the options provided, what's the best way to pass on the options.
17:18 thowe or should I be using the word route instead of method?
17:18 thowe er I mean params, not options
17:19 meredith you could just have that default action's method catch the odd param early and pass control to the other method
17:20 meredith or, on the client, add an onsubmit handler to the form that checks for the alternate behavior and switches the action before letting the event continue as normal.
17:20 thowe Yeah, that's what I intend, but if I do that will the params be available after control is passed?
17:21 thowe the first one, I mean
17:22 meredith yeah.  you get a Controller instance per request and it holds the 'tx' transaction attribute, so if you just short-circuit a method with `return $c->foo if $c->param('bar');` it will all be on the same controller instance with the same transaction attribute.
17:23 thowe meredith, cool, thanks.  I'll try that out.
17:24 thowe I have this "simple search" thing that needs to dish off to any one of the more complex search methods with simplified data.  That way I just utilize the same search methods regardless of where it came from.  Essentially the simple search is pre-defined basic searches.
17:24 joelb joined #mojo
17:24 thowe At least, that's my intent.
17:25 thowe I may change my mind along the way...
17:25 joelb I'm confused; now that Canary::Stability is in my codebase, what's it supposed to do?
17:25 joelb I thought it might stop me using 5.22.0, but that doesn't appear to be happening
17:26 meredith i thought it threw up a "i don't support this" message with a confirm prompt
17:27 dotandimet joined #mojo
17:27 joelb ye, was looking forward to that
17:27 joelb nm, cpanm has happily uninstalled it, so no permanent effect
17:29 Grinnz it just adds warnings to the install process, which you won't see with cpanm anyway
17:29 joelb cool
17:29 mst joelb: generally, what it's supposed to do is tell you to pick a different perl module
17:29 joelb cpanfile should support an extra option: packages to uninstall after installing the packages you asked for
17:30 mst or you could just not use things that do stupid shit
17:30 joelb mst: ye, it's a puzzle for modern psychology; a person who puts code in the public domain, but really doesn't want you to use it
17:30 mst joelb: the debian perl packaging team referred to it as 'the first cpan darwin award'
17:31 mst really though, JSON::MaybeXS will get you Cpanel::JSON::XS instead of JSON::XS
17:31 joelb (As someone who works with psychologists, I genuinely find myself concerned about his mental health)
17:31 mst and AnyEvent has been actively hostile to CPAN for years now so anybody with any sense has been steadily migrating off it
17:31 mst and I don't think the tiny stuff like Devel::FindRef has the canary bollocks yet
17:32 joelb My current project requires EV, but like I say, the C::S package uninstalls nicely
17:33 mst ahh, yeah, EV is a pain
17:33 * mst grumbles
17:33 Grinnz EV is the only one i still use, and even that's iffy; Mojo::Reactor::Poll is nearly as fast on linux
17:33 mst I keep forgetting about that one
17:33 joelb Grinnz: ah, that's v interesting
17:34 Grinnz (as Mojo::Reactor::EV; dunno about EV used directly)
17:35 joelb btw, I've only just discovered the joy of Lucy (not the drug)
17:36 joelb and I find myself pondering why people don't shout about how wonderful it is
17:36 meredith yeah most of the talk about EV as an issue is everyone looking wistfully at libuv but that's it's own big mess to implement, when i think the big issue here is hostile maintenance. that doesn't require finding a new lib, but people willing to maintain a newly-named fork of the old.
17:37 trone joined #mojo
17:37 meredith i'm not sure we're at quite that "jumping ship" point yet though
17:38 mst if his C code is as full of bullshit dickwaving as his perl code, finding somebody who'd want to maintain a fork would probably be hard
17:39 joelb libuv and libusb are my two big wants
17:40 meshl joined #mojo
17:40 meredith yeah i don't know about the internals there.  coro, on the other hand, i have like zero hope for someone else to try touching  (nor do i really want it continued in that form)
17:41 mst actually, somebody already got most of the way to fixing it
17:41 mst mostly I think to figure out how
17:41 mst I'm willing to bet he won't take the patch though
17:44 meredith Yeah, I saw it was a five line change or so.  I was thinking in the larger "take over maintenance" context though, coro being quite a hack.
17:44 meredith i mean it's impressive, ofc
17:51 Grinnz someone setting up an up-to-date library for libuv would be a start
17:51 Grinnz UV is painfully out of date now
17:51 Grinnz if only i had any idea about C bullshit
17:52 Grinnz unfortunately, "i didn't start developing perl to develop C" is probably a common theme among CPAN devs
17:53 lluad joined #mojo
17:57 Dave Ah the open source ego carousel. I doubt that will ever change :)
17:57 jberger joelb: I like Lucy but it doesn't have the type ahead search capabilities that elastic search has
17:58 jberger Grinnz: re libuv, one of these days ...
17:58 mst Dave: lehmann is very definitely a special case though
17:58 mst Dave: so far along that particular spectrum he's in danger of redshifting
17:58 Dave heh
17:58 jberger mst++
17:58 mattastrophe joined #mojo
17:59 Dave The thing is..."better" or "worse" are not scalar values
17:59 Dave it's a tensor of unknown rank if you really want to compare
17:59 mst Dave++ # math humour
17:59 Dave I've watched open source communities for quite a long time now. :D
18:00 Dave and it's always the same...logical errors turn into emotional dislikes
18:01 Dave and then the fork happens
18:01 jberger If this particular problem started as that, it has long since outgrown it
18:01 jberger Logical error -> emotional dislike
18:01 Dave May I hear your side of it? :D
18:01 Dave if you want
18:01 jberger The guy is darn near paranoic
18:02 Dave lol I understand him already ;) Does he take his paranoia seriously? (If I took mine seriously I'd be holed up in a cave in some mountain)
18:02 jberger Evening that doesn't go exactly his way is clearly an attack on him
18:02 mst I had him filed under '
18:02 meredith the guy has an autoresponder for rt.perl.org submissions telling you to email him, then has a tendency to curse at you when you email  >__>
18:02 jberger His first response to even the most well meaning criticism is accusations
18:02 mst I had him filed under 'just another ego' until he added a hard-coded die() to AnyEvent to stop you using it with IO::Async
18:03 jberger Then that happened ^^
18:03 jberger I thought it was exit though
18:03 mst he also refuses to set his email address as his bugtracker in META.*
18:03 mst while abusing anybody who then uses RT
18:03 Dave oo I remember reading about the die in IO::Async
18:03 meredith then you couldn't read the diatribe!
18:04 jberger mst: I kinda see his point I the bug tracker now
18:04 Dave that stopped me from using any of the modules ;)
18:04 jberger Especially with riba encouraging people to use the rt bug email even if the author sets another bug tracker in meta data
18:05 mst that wasn't until after he refused to do that
18:05 mst also after riba and I tried to help him get an email workflow for RT that he liked
18:05 mst and he then stopped responding, then blamed us for not keeping our promises
18:05 jberger No I mean riba is encouraging that for all modules now
18:05 jberger Mine included
18:06 jberger I set my meta data, I don't check or want to check rt
18:06 jnbek joined #mojo
18:06 jberger Mlehmann,s point, badly delivered as it is, it's that he can't opt out of rt all together
18:07 Grinnz if you want to understand mlehmann, go through his module documentation until you have read all 370 rants he's included
18:07 jberger He then throws out the baby the bath water and most of the rest of civility over it
18:08 Grinnz i mean, not wanting to check RT is perfectly fine, setting the metadata would prevent 99.9% of people from ACTUALLY USING IT for your modules though
18:08 Dave wow
18:08 Dave common::sense
18:08 mst right, and s.c.o and metacpan both use the bugtracker meta-info
18:08 Grinnz common::sense is actually one of his most sensible modules
18:08 meredith yeah i can understand being annoyed with it but when annoyed with something it doesn't get fixed if you just stop all action on your own part and fuss about it.
18:08 mst and rt.cpan.org's web UI will display that there's a different bugtracker
18:09 mst note: shadowcat is responsible for both those sets of changes
18:09 Grinnz :)
18:09 Dave didn't know about half of these features
18:09 Grinnz the RT redirect feature is awesome
18:09 mst so we've given him something much closer to opting out
18:09 mst and he still won't use it
18:09 jberger I didn't know that rt will show the other tracker, that's nice
18:09 mst because waving his dick about it is apparently more important to him than his users
18:09 Grinnz because he doesn't actually care, he just wants to complain on principle
18:09 mst right
18:09 Dave he actually waves that?
18:09 jberger will it still accept bugs for those projects?
18:09 mst which is the point at which, frankly, he can fuck off
18:10 Grinnz jberger, yeah, it jsut tells you at the top to use the other bugtracker
18:10 jberger (note that I am ABSOLUTELY NOT agreeing with him on behavior or anything)
18:10 mst jberger: yes, and that's a feature, because if the bugtracker disappears you want somewhere to file stuff for the next maintainer
18:10 mst remember, the rt.cpan.org queue exists for your users first and you second
18:10 jberger that's riba's point too
18:11 jberger I'm not sure I agree, but I can't come up with another solution for the "what if it goes away question"
18:11 Grinnz why is riba encouraging using the rt email? that seems dense to me
18:11 Grinnz missing the point of the metadata
18:11 jberger I think it doesn't help your users much if they throw bugs into a queue which isn't checked
18:12 jberger (granted, its nice that they at least see a warning on the ui)
18:12 Grinnz "if it goes away" you're going to have more problems than losing the existing bugs
18:12 mst hrm, maybe we should tweak the RT autoresponder as well
18:12 jberger Grinnz: his point is to make it dead simple to report bugs
18:12 mst Grinnz: ok, so, speaking as somebody who's been involved in a fucktonne of adoptions, no, actually, you aren't
18:12 Grinnz being simple is pointless if you're reporting it to somewhere that isn't used
18:12 mst Grinnz: generally rt accumulates the outstanding patches as well
18:13 Grinnz for github dists; generally the github PR queue does
18:13 mst Grinnz: so adoption becomes 'git init last cpan tarball, apply patches from rt, release, victory'
18:13 mst um
18:13 Grinnz i'm not sure i'm following the point here
18:13 mst the entire context of this is IF THE REPO AND TRACKER GO AWAY
18:14 mst mentioning something involving the repo and tracker is, therefore, rather missing the point
18:14 jberger catastrophic adoption
18:14 jberger think why_
18:14 Grinnz right, but there's no good way to solve that except duplicate everything from <external tracker> to RT
18:14 Grinnz which would be kind of annoying
18:14 mst o
18:14 mst no
18:14 * mst cries
18:14 jberger I mean _why
18:14 mst Grinnz: ok, look, speaking from a load of experience
18:15 mst if the original tracker goes away
18:15 mst people start filing on rt instead
18:15 mst so at least at the point of adoption not everything is lost
18:15 Grinnz ok?
18:15 mst it is not perfect, but omg is it better than nothing
18:15 Grinnz right, i'm saying there's no better way than what we're doing now
18:16 jberger Grinnz: which was my conclusion :-P
18:16 mst you started off by saying: "if it goes away" you're going to have more problems than losing the existing bugs
18:16 mst I was specifically speaking to that not really being the case
18:16 mst I agree with 'no better way' - but that isn't what you said :)
18:16 Dave common::sense turns off "use warnings qw(uninitialized)"
18:16 Grinnz well, i was thinking more that it's probably a political problem if that happens
18:17 mst huh?
18:17 Grinnz rather than the technical aspects
18:17 mst nope
18:17 mst usually an author changed jobs or languages or countries or died
18:17 Grinnz none of that will cause a github repo to go away
18:18 mst once you've handled enough adoptions to be given PAUSE admin bits to stop you bugging the other admins all the damn time, you're welcome to tell me if your experience differs to mine
18:19 mst right now, I'm saying "this is what actually happens" and you're theorising randomly and thinking that's a counter-point
18:19 mst it isn't. I'm not telling you what I expect to happen. I'm telling you what has happened.
18:20 Grinnz i make no claims to having experience in the matter, so i concede the point
18:22 Dave wow...the subject of this lehmann guy comes up and debates appear like magic
18:22 Grinnz oh, were we supposed to debate about him
18:23 Dave lol
18:23 Dave I'm just saying the subject seems to have associativity to debate :)
18:25 Dave still, it's probably just the normal "get a bunch of geeks in a room and watch the fun" idea
18:25 da5id joined #mojo
18:25 Grinnz this is #mojo, that's true of pretty much any subject
18:26 Dave this channel is pretty tame as they go tho :D
18:26 Dave at least people here have a brain
18:26 Dave so the debates are usually interesting
18:39 thowe can I do a redirect_to with params and the POST method?
18:40 Grinnz thowe, no. a redirect is just a response to the browser, to tell it where to go, which it will go to with a GET
18:40 thowe hm, then how do I post to a route from another route method?
18:40 Grinnz could you just call a common method internally?
18:42 thowe Perhaps if I describe what I'm trying to do....
18:42 thowe I have a number of fairly comples searches that are almost on the level of full blown query builders.  One for accounts, one for services, etc...
18:42 thowe s/comples/complex/
18:43 marty thowe:  Maybe something like...   $c->redirect_to($c->url_for($c->req->url->to_abs)->query($c->req->url->query));
18:43 Grinnz i have kind of a similar thing. it's implemented as an object, which i can create and call a method on from other routes when i need it
18:44 thowe I also have this simple search window on every page that allows you to select a search type and put in a search string and it does a simple search.  I want that to dish out to the right search method and pass just the params required to do the simple default search.
18:44 thowe so if account is in the dropdown and some string is entered, it will dish to the account search and just do a "account name contains $string" search.
18:45 thowe But the account search accepts GET and POST.  If POST, it does the search and lists the reselts, if GET it displays the form.  So I want to send a POST request.
18:46 thowe I guess I could just look for params on GET and do the search...  but now I am curious how to POST.
18:47 Grinnz that seems like an odd usage of GET vs POST
18:47 Grinnz it sounds like it should be a different endpoint instead
18:47 thowe Is it?  I thought it was a fairly standard pattern...
18:48 Grinnz POST is for creating/updating data, usually
18:48 Grinnz you are just specifying parameters for a search
18:49 thowe So would I have to use the UA?
18:49 Grinnz huh?
18:50 Grinnz using the UA on another route in the same app usually means you should have it abstracted out to a common method instead
18:50 Grinnz (and won't work very well)
18:51 thowe right.  Look, I can work around this.  Now I'm just wondering if there is a way to "redirect" but with the POST method.  It's turned into less of a "how do I make my app work" to "now I have a general curiosity"
18:51 Grinnz no, there isn't
18:51 Grinnz you know how when you hit refresh on a POSTed page, it gives you a warning? that's another reason why; generally, POST means something you don't want users to accidentally do
18:52 thowe Fair enough.  And perhaps I need two separate routes for GET and POST here to make things more clear...
18:53 thowe I think the "same page does get and post" is something I inherited from some of the PHP around here...  and the less I inherit from that the better...
18:53 Grinnz "search parameters form" and "search results" sound like two routes to me
18:53 thowe especially since Mojo makes it easy
18:53 Grinnz it is reasonable for somethings... but doesnt quite fit this
18:54 thowe Yeah, I think you are right.
18:55 mst thowe: there is, actually, technically, a set of circumstances in which user-agents are allowed to preserve the POST data across a redirect, but I've never seen one that does
18:56 thowe My limited experience with web dev has me more interested in accepting advice and less interested in pioneering a new wheel project...
18:58 thowe But isn't the search results a page that would take POST?  Or should it accept GET and POST and just take the params as given?
18:58 mst right, but what I'm saying is that Grinnz' "no, there isn't" can be expanded to "if you read the HTTP spec carefully, in theory, there is, but in practice, no, there isn't"
18:58 thowe where the form page is just GET
18:58 mst I'm not suggesting a different conclusion, just expanding on why what you already got told was right ;)
18:58 thowe mst, gotcha
18:59 thowe This is all stuff I've done before, BTW.  I hope I'm not annoying by thinking out loud in here.  I tend to rethink my assumptions with every new project.
18:59 Grinnz thowe, a search is usually GET territory, in that it doesn't change state of the application
19:00 thowe but, still, the form will POST, no?  I mean, I guess it can just be a GET...
19:00 Grinnz there is one instance where we use POST for a search, but that's because it's through Datatables which tends to send so many parameters that it doesn't fit in the URL
19:00 Grinnz you can set a form method to GET
19:01 thowe I suppose intent is perhaps more important there than convention
19:01 Grinnz http://www.w3schools.com/tags/att_form_method.asp
19:01 thowe I just tend to think "forms should POST" but that's probably an over-simplification
19:02 Grinnz forms usually are used to update things, which is why that's a convention
19:02 thowe right, but in this case it's a vehicle for creating a distilled view of results.
19:02 thowe I think I've been talked into using GET here.  That's really solves the whole thing
19:06 thowe I think I asked this before, but to clarify:  If I redirect_to, are my $c->params preserved?
19:07 Grinnz no, you would have to include the query string in your redirect
19:07 disputin joined #mojo
19:08 Grinnz you can use Mojo::Parameters or Mojo::URL to build the query string properly
19:13 thowe so, can I  $self-url_for('searchdisplay')->query($self-param)  ?
19:13 thowe $self->url_for('searchdisplay')->query($self-param)
19:14 thowe or is that not a parameter object?
19:19 thowe $self->url_for('searchdisplay')->query($self->param)   I keep doing that....
19:20 thowe forgetting the > I mean
19:22 zackiv31_ joined #mojo
19:29 Grinnz_ if url_for returns a Mojo::URL then yeah you could do that to get the url to redirect to
19:31 meredith this is more what i was thinking:  https://gist.github.com/merrilymeredith/a050276a3c190a49dc8b
19:31 meredith is the redirect necessary?
19:32 Grinnz_ meredith: right thats what i was saying before about using an internal method (but it doesn't even have to be a controller method)
19:32 Grinnz_ just something that's commonly accessed
19:32 meredith yeah, the next step in abstraction would be that distinct search class, the controller just passing on to it in all cases
19:33 thowe Well, I'm still trying to figure out how much query prep I need to do beforehand...  But thanks for that example.
19:33 thowe One thing I need to figure out is how to give some "and" and "or" select ability without actually building a full on query builder.
19:34 Grinnz_ basically i have a search object, i set params in it based on request parameters, and get a hashref back which i turn into the JSON output (could also be thrown into a template)
19:34 Grinnz_ i do pass the request object itself for one set of parameters, but other than that the search object doesn't know anything about the request
19:34 meredith It's just really nice there being no particular magic in mojo, being able to use perl OO as you normally would.
19:35 meredith i've been working with catalyst for about four months, forgive me if i get a little emotional here
19:35 thowe Initially I wanted my list (query results) to accept requests for html or json, but we'll see how industrious I get
19:35 Grinnz_ thowe: with the right abstraction, that's easy
19:35 thowe Grinnz, right.
19:36 thowe content negotiation and all that...
19:38 thowe about thew query building, I was just going to allow the "search logic operator" to be "and" or "or" and the params requested would be HASHed or ARRAYed as needed...
19:38 thowe SQL Abstract style
19:39 thowe I also allow "NOT" for just about everything
19:43 thowe it's a fabulous mess
19:54 Dada joined #mojo
20:00 meshl joined #mojo
20:00 berov joined #mojo
20:01 hernan605 joined #mojo
20:13 jnbek joined #mojo
20:20 disputin joined #mojo
20:36 thowe something I don't get here...  I'm doing "$self->redirect_to('list_accounts',  { 'name' => $search }); "   but then $self->param('name') is undef in list_accounts.
20:37 thowe and I have already established that $search is not undef
20:38 thowe I don't seem to be getting any args in the URL....
20:45 thowe guess I need to build a URL...
20:51 thowe shabam.  Working search abstraction.
21:44 jberger right, all of the url_for-like functions take a name and stash parameters
21:44 jberger and stash parameters will not populate a query
21:44 jberger so as you found you have to build those yourself
22:18 Dada left #mojo
22:58 melch joined #mojo
23:18 bpmedley joined #mojo
23:55 ChmEarl joined #mojo

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