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

IRC log for #mojo, 2015-10-07

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

All times shown according to UTC.

Time Nick Message
00:02 howitdo joined #mojo
00:03 howitdo joined #mojo
00:09 mattastrophe joined #mojo
00:34 PryMar56 joined #mojo
00:55 aborazmeh joined #mojo
01:09 cpan_mojo Mojo-Webqq-1.5.3 by SJDY https://metacpan.org/release/SJDY/Mojo-Webqq-1.5.3
01:33 jberger Phil22: of course there are other things you can do
01:34 jberger Especially if it is a rest api used from a non browser client
01:36 mattastrophe joined #mojo
01:36 jberger But in the end I think most of them involve some other exchange of secrets
01:36 jberger And so you come to public key exchange, and now you're getting back to just text over https
01:44 thowe joined #mojo
01:46 mattastrophe joined #mojo
02:16 inokenty-w joined #mojo
02:27 noganex joined #mojo
02:36 kaare joined #mojo
02:37 bc547 joined #mojo
02:46 Phil22 yeah
02:46 Phil22 figure there may have been a standard "shared secret" auth plugin for mojo is all
02:47 Phil22 server side is mojo, client side is mojo :)
02:47 Phil22 had to.. just make my own, no biggie
02:51 pink_mist well, now you can make it into a plugin and put it on CPAN so others can use it :P
03:06 Phil22 hehe, not a bad idea
03:06 Phil22 I should clean this concept up a little
03:14 Grinnz the problem with putting things on CPAN, then you're responsible :P
03:33 preaction yeah...
03:34 preaction and then eventually they start making you responsible for other things, like ether: http://metacpan.org/author/ETHER
03:45 Phil22 haha
03:45 Phil22 man, that's nuts
03:50 melo1 joined #mojo
05:16 disputin joined #mojo
06:01 dod joined #mojo
06:05 dod joined #mojo
06:12 dod joined #mojo
06:13 omega is there a reason Test::Mojo doesn't have an "attr_*" family of helpers? not used enough times?
06:13 panshin joined #mojo
06:14 batman omega: can you explain how you would use it?
06:14 omega batman: is($t->tx->res->dom->at('head link[rel="canonical"]')->attr('href'), 'http://www.
06:14 omega like that for instance
06:14 omega or to check href on a elements
06:15 omega hmm, maybe I could use element exists
06:15 omega if I can use link[rel="canonical" href="something"]
06:15 batman right. i think element_exists() covers that: ->element_exists(q([href="http://example.com"]));
06:15 omega but that doesn't give good diagnostics if the thing is there, but the attribute is different
06:18 omega and you can't ask for two attributes?
06:18 omega or select on two attributes?
06:18 batman yeah, just do q([href="/example"][target="_blank"])
06:19 omega ahh, yes
06:19 batman but it still won't give you good diagnostics, but your is(...) example won't do that either
06:19 omega read the CSS spec now
06:19 omega if teh value is there, but wrong, it tells me "got: blabla expected: balasd"
06:19 omega like if the canonical link is there, but wrong
06:19 batman not really... it would just fail completely if the at() doesn't return an element
06:20 omega 08:19      omega: like if the canonical link is there, but wrong
06:20 omega but yes, it dies if the element isn't there. an attr_is could just say "no such element found"
06:21 omega anyways, was just an idea
06:21 batman another tip: you can do [href^="/foo"] for matching at the beginning... have a look here for more examples about $=, *=, ~=: https://metacpan.org/pod/Mojo::DOM::CSS
06:21 omega yeah, I know
06:22 batman omega: yeah, i like the idea. i've often found myself in the same situation. like element_exists('form[href=...]') failing, which makes me curious, since i know i haven't removed the form element
06:23 batman not sure if it's easy to implement though.
06:24 omega shouldn't be too hard, looking at the code for Test::Mojo
06:25 batman omega: would be cool if you make a draft in a pastebin
06:26 omega ok
06:27 batman i think less confusing feedback from a test would be nice
06:29 batman i often do something like this: ->element_exists(q(form))->element_exists(q(form[action="/login"])), but i wouldn't have to if the error message said: found form, but not with  [action="/login"]
06:29 batman but i don't see how that is easy to do
06:30 omega batman: https://gist.github.com/omega/73afe7b09b140bee3dd4
06:30 omega no, the selector would still either fail or not I guess
06:31 omega but with this, if you specify the wrong attr (it needs extra argument to specify which attr), you still get the diagnostics
06:31 omega or if the attr is wrong value, you get diagnostics
06:32 omega the "no such element" could probably be handled in a better way
06:32 batman i think maybe you want find() instead of at() and the loop through the elements to see if an element match ($attr, $value)
06:32 omega ahh, might be
06:32 batman so you can do ->attr_is("a", href => "http://example.com");
06:33 omega mhm
06:33 batman not sure if i like the method name, but i think it's an interesting idea.
06:33 batman it's good enough for me (personally) to make an issue...
06:35 batman i think maybe i would like to support ->attr_is($selector, %attrs, $desc);
06:35 batman $desc = @_ % 2 ? pop : '...';
06:35 omega makes it more complicated though, what happens if there are conflicting ones?
06:35 omega or if there is more than one match for a given attr => value pair?
06:36 batman it requires all the %attrs to match
06:36 omega like if there are 2 links with an href to example.com
06:36 omega it would be quite different from how text_* works at that point I guess, but that might make sense
06:37 omega Checks text content of the CSS selectors first matching HTML/XML element for exact match
06:37 omega so they document it is "first match"
06:37 batman you filter on each pair. something like this: for my $k (keys %attr) { @e = grep { $_->attr($k) eq $attr{$k} } @e }
06:39 batman yes. most of the tests doesn't care if you have more than one element that match.
06:41 omega mhm
06:51 kivilahtio joined #mojo
07:08 AndrewIsh joined #mojo
07:08 marcusr moin
07:19 CandyAngel Morning marcusr
07:22 eseyman joined #mojo
07:26 trone joined #mojo
07:30 marcus CandyAngel: :)
07:33 batman I don't get this: http://twitter.com/Joab_Jackson/status/651392983112282113
07:44 berov joined #mojo
07:45 berov joined #mojo
08:17 Vandal joined #mojo
08:17 vanHoesel joined #mojo
08:24 marty joined #mojo
08:45 berov joined #mojo
08:47 panshin joined #mojo
09:02 vytas joined #mojo
09:05 meshl joined #mojo
09:09 vytas joined #mojo
09:23 melo joined #mojo
09:29 CandyAngel DBIC is made of magic. That is all.
09:29 Snelius DBIC is world of pain
09:32 gaunt_ joined #mojo
09:34 CandyAngel Hm? The only thing it hasn't been able to do that I can tell (so far) is INSERT INTO () SELECT..
09:40 panshin joined #mojo
09:42 sugar joined #mojo
09:47 meredith since i think that's pretty much three words prefixing a regular select, i'd look to ->as_query on the built-up resultset.
09:48 Snelius Guys, Perl6 is out.
09:48 meredith just nice that up into a method that takes a destination table and tacks the insert bit on the front of ->as_query->[0] and such...
09:48 nic Everyone will remember exactly where they were the moment they heard perl6 was alive
09:49 CandyAngel meredith: It's not that important, just trying to get the conversion to a database done first
09:49 CandyAngel Dependencies will be fun :D
09:50 sri batman: you can't parse html with regex, that's a law of nature
09:50 nic Snelius: Do you have a link?
09:51 nic Is it hush-hush at the mo?
09:51 Snelius nic: http://www.pigdog.org/auto/software_jihad/link/3138.html
09:56 jontaylor joined #mojo
10:05 sri for the record, perl6 is not out
10:05 * nic was about to say that
10:09 cosimo joined #mojo
10:14 sri omega: that attr test stuff doesn't seem very elegant
10:28 sri and for core it has to be elegant, you can still make it a cpan modle though, with jberger's role stuff
10:28 sri +u
10:36 panshin joined #mojo
10:40 batman sri: you can't parse html with _a_single_regex_ ..? i mean, Mojo::DOM::HTML seems to use regex...
10:48 mattastrophe joined #mojo
10:50 sri batman: yes, Mojo::DOM::HTML only uses a regex to tokenize HTML, not to parse
10:51 MartinR joined #mojo
10:51 batman aha. thanks for the clarification :)
10:51 sri you just have to look at the long lists of special cases to see why a single regex could never ever work
10:52 sri https://github.com/kraih/mojo/blob/master/lib/Mojo/DOM/HTML.pm#L49-L101
10:56 sugar joined #mojo
10:57 MartinR it may sounds odd but is it possible to run two mojolicious apps on same machine ?
10:57 MartinR I would like to test user::agent
10:59 melo1 joined #mojo
11:00 Snelius sure, it's a possible
11:06 MartinR yeah, -l in morbo did the job
11:06 MartinR thanks
11:07 neilhwatson joined #mojo
11:19 absolut_todd joined #mojo
11:33 sugar_ joined #mojo
11:34 meshl joined #mojo
11:44 sugar_ joined #mojo
11:57 absolut_todd joined #mojo
12:02 vytas joined #mojo
12:05 cpan_mojo Mojolicious-Plugin-ProxyPassReverse-SubDir-0.02 by MURATA https://metacpan.org/release/MURATA/Mojolicious-Plugin-ProxyPassReverse-SubDir-0.02
12:06 panshin joined #mojo
12:07 marcusr sounds lovely
12:07 panshin joined #mojo
12:07 marcusr ...
12:07 jberger omega / batman: that handles an element with the the attribute, but what if I want a descendant with an attribute?
12:08 batman jberger: "a" can't have a decendent, but... you could do things like ("div.foo div", class => "foo")
12:09 batman jberger: nothing special about that really
12:10 jberger And that would only match the class on the second div correct?
12:11 batman yes. that is the only element that would match. nothing special about that. just the way the css selectors work.
12:12 batman "the only" = "not the parent"
12:12 jberger div.foo div [class="bar"]
12:12 batman that would match whatever-element which is the child of "div.foo" and then "div", because of the space.
12:13 jberger Yes that's what I'm asking about
12:13 batman what you wrote is not the same as element_with_attr("div.foo div", class => "foo") though
12:13 batman or whatever the method should be clled
12:13 jberger I guess you just go back to element_exists for that then
12:13 Zoffix Snelius, just an overly eager author. Only the developer release 6.Birthday is out. The first release of Perl 6—version 6.Christmas—is still scheduled for Christmas.
12:13 * Zoffix is disappointed seeing so many people get confused over that same article.
12:15 sri "confusion" sums up the entire perl6 process pretty well
12:16 Zoffix I'm pretty sure it's about to get worse.
12:16 * sri is really disappointed with the new perl6 site
12:16 Zoffix sri, why?
12:16 * Zoffix made it
12:16 marcus muhaha
12:16 Zoffix :)
12:16 sri few years back someone made a really nice looking one with my butterfly logo, but it wasn't accepted
12:18 sri Zoffix: oh
12:18 Zoffix sri, are you still going to make a Perl 6 Mojolicious? I saw a tweet or something from a few years back, where someone else asked and your response was "when it gets better IO" or something
12:19 MartinR huh it may be hard to maintain both
12:19 sri Zoffix: maybe
12:20 sri if i can get payed doing perl6 stuff sure
12:21 MartinR sri, "volunteers" is the most popular word around perl6 ;p
12:22 jberger At least a few weeks ago when I looked there still wasn't an async useragent available
12:22 jberger Or not that I could find
12:22 jberger You'd think that'd make a good example for the baked in async features of p6
12:26 sri Zoffix: if i was going to make a new perl site i'd aim for something clean like https://www.python.org
12:28 sri of course there's only so much you can do with the current logo...
12:28 Zoffix sri, I have a different vision. Something very light; on white bg.
12:28 sri like the new ruby site?
12:29 sri seems a bit bland
12:29 MartinR perl6.org is on what framework ?
12:29 Zoffix MartinR, it's a static site generated by Perl 5's App::mowyw
12:29 Zoffix (or however you spell it)
12:30 MartinR mhm
12:30 Zoffix But... baby steps. I'm glad the atrocious typography of the old site is not there any more.
12:31 Zoffix Which is what motivated me.
12:31 MartinR do you like changes made in perl6 ?
12:31 Zoffix MartinR, the language?
12:31 MartinR yes
12:31 Zoffix I don't know Perl 6.
12:33 MartinR I have heard some opinions that perl 6 compares to perl5 like C++ to C
12:33 sri in the beginning mojolicious actually had a rule that new features should be designed with a perl6 port in mind
12:33 panshin joined #mojo
12:34 MartinR I also heard it's much slower than perl5 but I've never tested it myself
12:34 Zoffix MartinR, it is. But there's no production version yet.
12:34 Zoffix So saying it's "slower" is ridiculous :)
12:34 meredith the navbar and footer greens on perl6.org feel disconnected from the rest of the basic black/white/blue/grey of the main content
12:35 Zoffix meredith, https://github.com/perl6/perl6.org
12:35 Zoffix :) PRs welcome
12:36 sri <insert rant about sticky navbars>
12:37 vytas joined #mojo
12:37 Zoffix <insert retort about sticky navbars>
12:39 sri but seriously, sticky navbars are the worst
12:40 y1mmm joined #mojo
12:40 Zoffix No, they're the best!
12:40 Zoffix What's wrong with them?
12:40 sri waste of screen real estate in our mobile world
12:41 sri if it's an actual web app, like gmail, sure
12:41 MartinR http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#Large-file-upload --> This example involves 2 machines, what if I would like to redirect user upload request (file on his HDD) to another mojo server
12:41 sri but if you want people to read your pages, then you better make them unobtrusive
12:41 batman Zoffix: i think i would re-use the colors from the camelia logo. there's too many colors which isn't really grounded to anything
12:42 Zoffix batman, that was my first go. And it looked atrocious.
12:42 sri Zoffix: start by making a color palette
12:42 Zoffix Then I tried about 6 different palettes. And then I gave up.
12:42 batman it also feels a bit weird that some of the elements are flat and others not.
12:42 Zoffix batman, it's the damn Bootstrap3 update!
12:42 meredith yeah unfortunately camelia doesn't give the best base palette for the brand ;)
12:42 Zoffix I hate the non-flat stuff too
12:43 batman meredith: sure, but at least it would be consistent.
12:43 bpmedley MartinR: I'm confused.  Would the user be making the upload request in a browser?
12:43 MartinR bpmedley, the scenario is: User send upload request to ->  Mojo server, mojo server pass it to another one -> Another mojo server
12:44 bpmedley MartinR: Why not redirect before the upload starts to happen?
12:44 MartinR the point is that the data should flow from user computer to that third server, bypassing the master machine
12:44 MartinR bpmedley, definitely before the upload start to happen but I am not sure how to accomplish it
12:45 MartinR never used useragent before :>
12:45 bpmedley MartinR: Give me a few minutes.
12:45 MartinR ;)
12:50 MartinR but master machine has to decide where to pass that upload request so there will be really three machines involved (user, master, destination server)
13:00 nic MartinR: redirect_to ?
13:00 stephen joined #mojo
13:01 nic One of my favourite areas in Mojolicious is url_for + url_with
13:02 nic With all this talk of server-to-server, it might be a great plugin where your master server has the route structure(s) of the satellite servers and gives you an equivalent of those helpers
13:02 nic I mean specifically the lookup between 'easy name' and 'proper url'
13:02 bpmedley MartinR: Gaw, I had forgotten that the POST would redirect to a GET.   Did I say that correctly?
13:02 nic cos the lookup mapping changes over time
13:03 MartinR hmm I dunno, both will be POST probably
13:04 nic MartinR: Once the master server has identified the appropriate satellite it redirects there, and then the first post takes place, user->satellite
13:05 nic kind of similar to what retail sites do when identifying which payment partner to delegate to
13:06 MartinR nic, but it should happen in a way invisible for the user, so basicly user is constantly on site.com/upload and even sees progress bar of ulpoad
13:06 MartinR on the background all the magic happend
13:07 MartinR and his data is actually send to another machine
13:07 sri bpmedley: 307 redirect
13:07 sri or even 308
13:07 MartinR with rediret_to ?
13:08 hernan605 joined #mojo
13:08 sri it's a bad idea though, since the entire upload will be sent to the wrong server first, before being sent again to the right one
13:09 MartinR sri, no
13:09 MartinR the idea is to send it once
13:09 bpmedley sri: thanks
13:09 MartinR maybe I said it wrong and you misunderstood me
13:10 MartinR but when master mechine gets upload request it only decide where it should pass it until that no upload is being done
13:10 jberger MartinR: browsers don't work that way
13:11 MartinR hmm
13:11 jberger The upload is part of the initial request, whether you want it to be or not
13:12 jberger You could build some preflight request to get the actual target
13:12 MartinR jberger, I think about it as a pre_request
13:13 jberger Yes, that's what I'm saying too, only it had to be two separate requests
13:13 pink_mist you might be able to write a flash application that would work like that...
13:13 pink_mist but normal http requests from a browser won't
13:13 * jberger runs screaming
13:14 sri pink_mist: then you might as well upload through a websocket
13:14 pink_mist good point
13:14 sri which is actually not a bad idea, since websockets can be compressed, so you get compressed uploads
13:15 sri permessage-deflate
13:16 stephan48 you could figure out the target server and then point the action="" of the form towards it. afterwards you do a redirect from there back to the master server
13:17 jberger I had a chucked websocket transfer mechanism in the works for Galileo
13:17 jberger Been sitting deep in my github since then
13:17 MartinR is there any example for that ?
13:17 pink_mist long ago in a branch far far away?
13:17 jberger repo far far away
13:18 jberger https://github.com/jberger/GalileoSend
13:18 jberger No idea if I recommend any part of that btw
13:19 pink_mist wow, not even the same repo heh
13:19 jberger I've come a long way as a programmer since then
13:19 MartinR ok, thanks for help guys ;)
13:20 nic MartinR: In finance it's typically done with an iframe
13:23 MartinR is it a good attempt $c->res->code(307); $c->redirect_to(ip/upload); ?
13:23 gryphon joined #mojo
13:25 bpmedley MartinR: As sri stated, won't that upload the entire file twice?
13:26 MartinR yes probably,
13:26 MartinR hmm, I have to sit back and think about it in details - no clean solution at this time ;/
13:30 jberger GET /api/whereshouldmyfilego
13:30 jberger POST //some/target/from/response
13:31 MartinR jberger, its correct but the question is got ro redirect it from first get to that post on satelite server
13:31 MartinR how to*
13:32 pink_mist MartinR: the GET request there shouldn't redirect. it should show the person the upload form. that should then simply be posted to the correct place once the user has filled it in
13:36 jberger MartinR: we've told you what isn't possible and ideas to work around it, if those won't suffice I don't know what will
13:36 jberger But wishing won't make redirect help you here
13:37 MartinR pink_mist, that's right the upload form shown by master server, and now user select file and when click 'Upload' button the upload should go to right server (satelite)
13:38 pink_mist MartinR: yes, there's no need for a form post target to be the same server that served the form.
13:38 pink_mist just post it directly to the right place
13:38 pink_mist no redirect needed
13:40 MartinR but how do you expect to control upload status that way ?
13:40 MartinR let's say progress of upload
13:40 MartinR one way would be to save json formated data on that server and read it
13:41 pink_mist uh what?
13:42 MartinR if the upload is performed by satelite server and upload form (progress bar etc) is on master mechine you have to somehow controle the upload process right ?
13:44 pink_mist if you want a progress bar, supply it from the target server. if you can.
13:44 MartinR that's what i was thinking about
13:46 pink_mist seems like you need to put more logic on your target server then
13:47 pink_mist well, seems like you might need to redirect after all, but then it'd be GET /masterserver ->redirect_to-> GET /targetserver/uploadform ... and then the user would fill in the form an POST it to the target
13:48 pink_mist not sure that's necessary though ... dunno what restrictions there are on js communicating with other servers than what it's served from these days
13:49 MartinR no doubt about it, that POST request (right to satelite server) should be like this: POST /satelite_ip/upload , is it right ?
14:02 bpmedley MartinR: https://bitbucket.org/snippets/bpmedley/rL9L9 <-- This may work for you
14:03 MartinR bpmedley,thanks I will give it a try and let you know later;)
14:04 jberger bpmedley: there's lots of cruft in there, no?
14:04 jberger /action isn't used
14:04 bpmedley doh
14:04 jberger post / is the thing we've been saying this whole time, it redirects and thus the upload is done twice
14:05 jberger plus it has an unused $url
14:05 bpmedley jberger: Doesn't my example use an ajax upload?
14:05 aborazmeh joined #mojo
14:05 bpmedley Let me update, sorry
14:05 jberger oh, then the post / handler is unnecessary too
14:06 bpmedley Yes, sorry.
14:06 bellows joined #mojo
14:08 bellows Just started using Mojolicious::Lite but not actually finding any docs how to provide the auth token from "Under" to Mojo ?
14:08 bpmedley bellows: What kind of auth are you wanting to do?
14:09 Grinnz bellows, you can store things in the stash
14:09 Grinnz it will persist for that particular request
14:12 bellows Grinnz: let me rephrase it, I got to localhost:3000 where to type the authentication ? I found a little reciepe that i adapted here : http://code-tricks.com/build-a-simple-restful-api-using-mojolicious/
14:12 jberger bpmedley / MartinR: http://pastie.org/10466164
14:13 Grinnz bellows, that example shows it taking the auth token as a query parameter "token"
14:13 bpmedley jberger: My example has a fancy bar graph and can support multiple uploads and other doodads via the file input plugin.
14:14 bellows grinnz: how send that query parameter ?
14:14 jberger bpmedley: sure, add whatever javascript niceties that you want
14:15 jberger but the idea is that the server that serves the forms is presumably this master server and it should just put the target into the form at template time
14:15 Grinnz bellows, ?token=xxxx
14:15 bpmedley jberger: It's just that I thought I read that MartinR wanted a bar graph and wanted to upload across a CORS boundary.
14:16 zackiv31 joined #mojo
14:16 jberger bpmedley: sure that's all the niceties, but the real problem is that he wants server A to tell the client to upload to server B
14:17 bellows grinnz thanks , that helped me a lot
14:18 jberger but (s)he (assumed he from Martin, sorry) that the process of A passing to B had to happen after the upload, while I think most of us are trying to say that the "redirect" should come before the actual upload somehow
14:18 jberger bpmedley: your example did that too, but it did so by hardcoding in the javascript, which I suppose is fine if there is only ever to be one server B
14:19 bpmedley Hrmm.  Perhaps I should update the example to use url_for in the template..
14:19 jberger the most difficult example I can think of would be if some other form option determines (via some server-side algorithm) determines which server is to be the recipient of the file, but even then, I think the javascript should make some pre-flight request of server A to determine the target
14:20 bpmedley I agree with the pre-flight request.
14:20 ashimema if your not using mojo sessions.. is there still a need to change the secret?
14:20 ashimema used anywhere other than cookies>?
14:20 jberger ashimema: I don't believe so, other than to silence the warning
14:20 jberger but I could be wrong
14:21 ashimema thanks jberger.. I thought that may be the case.
14:23 bpmedley jberger: Out of curiosity, were you able to try the Minion UI?
14:23 jberger bpmedley: sorry, I really haven't had time
14:23 bpmedley It's cool, thx
14:24 jberger I haven't worked on mine either if that tells you anything
14:24 bpmedley :-0
14:32 PryMar56 joined #mojo
14:37 vytas joined #mojo
14:49 sri re postgresql, i love how easy stuff like this is http://dba.stackexchange.com/questions/90482/export-postgres-table-as-json/90495#90495
15:02 sugar__ joined #mojo
15:03 batman http://twitter.com/rjbs/status/651487524880302080
15:11 sugar joined #mojo
15:12 go|dfish 2017 ;_;
15:16 panshin joined #mojo
15:16 sri NOOOOOOO
15:17 batman :(
15:19 sugar joined #mojo
15:28 * jberger weeps
15:28 meshl joined #mojo
15:30 Onigiri Having the db do json itself is kind of odd.
15:31 Onigiri I fully get that having the validation when storing it could be valuable, but producing it from random things?
15:35 jberger so those are json documents by line, can it do a json array of the results instead?
15:36 jberger Onigiri: always best to have the db do as much of the work as possible
15:37 jberger if the webapp just has to pass out a string rather than json encode a bunch of results, that's more efficiency from your server (assuming you do the query such that it doesn't impede the server)
15:37 sri that makes a perl6 port more realistic again
15:38 sri even without funding
15:38 sri 2017 is a pretty long time
15:39 jberger and even then, if mojo core was going to use signatures it would mean killing all support for previous perl versions
15:40 jberger changing the documentation over is easy, adding it to Mojo::Base based on version is not so bad, but using it in core would be costly
15:40 sri even using it in docs, don't think i could defend it as long as it's still experimental
15:48 sri considering how hard it was already for the frontpage example
16:01 sri perhaps we can handle a perl6/perl5 dual-project like what play framework did with java and scala
16:01 jberger I'd be interested in that
16:02 jberger but I don't expect that any 1-1 port would be possible
16:02 jberger but overall the structure is probably similar
16:03 bwf joined #mojo
16:03 * sri nods
16:07 Grinnz_ i imagine the async will be quite different
16:08 sri much less necessary, since there's real threads
16:09 pink_mist react { whenever .. { .. } whenever .. { .. } } is how you write async in p6 ... react is the loop
16:10 pink_mist the 'whenever' block can be written anywhere, but needs to be called from within the react loop to ... react
16:11 jberger pink_mist: that's some rather recent sugar, no?
16:11 pink_mist somewhat
16:12 pink_mist you can see jnthn giving a talk about it here ... very very interesting: https://youtu.be/JpqnNCx7wVY
16:12 sri https://gist.github.com/jnthn/a56fd4a22e7c43080078
16:12 pink_mist err, not sure why that seems to link 11 minutes into the talk 0_o
16:13 pink_mist (or is that just for me for some reason?)
16:13 sri just you
16:13 sri youtube remembers
16:13 jberger I'd expect, though certainly not require, that a project of the scope of a mojo port to p6 would use less sugary interface
16:13 jberger primatives
16:15 jberger hmmm, as I read that, maybe my comment was too hasty
16:15 jberger these are basically just constructors/tappers
16:15 sri the json example with await is especially impressive
16:17 sri just my $foo = start from-json(slurp $some_file); my $bar = start from-json(slurp $other_file); my ($result1, $result2) = await $foo, $bar;
16:17 pink_mist that's not async though
16:18 sri i'm assuming start creates a thread and returns a promise
16:18 pink_mist yeah, it's parallel
16:18 sri yea
16:18 sri so async is much less important
16:18 sri you can wrap anything in a thread and it's reasonably cheap
16:19 sri you want a differet app server architecture when you can do that
16:19 sri async only really matters for the parts that have to deal with super high concurrency
16:19 sri like the server i/o bits, but for the http client it's meh
16:20 pink_mist still think async is the right solution to a lot of problems some would use threads for though
16:21 sri if only i had more time :(
16:22 pink_mist just last week they added so you could get a Supply from an IO::Handle ... before that it was impossible to read from a filehandle asyncly
16:23 pink_mist sockets already had support for it though
16:23 sri bet java/scala and Go will provide a lot of inspiration for perl6 frameworks
16:27 ZoffixWork joined #mojo
16:29 disputin joined #mojo
16:31 * sri would start by porting Mojo::DOM anyway
16:32 sri it's reasonably isolated and simple, so a good project for learning
16:33 sri and then move on to a simple web server, for testing/benchmarking different architectures
16:34 sri and from there move on to everything else
16:35 sri web server to routing and actual framework, then user agent and testing
16:36 sri in my head protocol abstraction is still the trickiest problem
16:36 sri http/1 next to http/2
16:38 sri without the need for an actual event loop the low level stuff gets much easier
16:39 sri if perl6 was reasonably stable, i bet it could be production ready within 2 years
16:47 hernan605 joined #mojo
16:54 hernan605 joined #mojo
17:07 trone joined #mojo
17:26 mattastrophe joined #mojo
17:33 lluad joined #mojo
17:40 Kogurr joined #mojo
17:44 sri *crickets*
17:52 disputin joined #mojo
17:56 bpmedley I'm personally very excited about the what I just read.  I didn't realize perl6 had native threads and it seems that something as useful as await/async can be realized.  I'll have to wait several hours to read through and view the youtube, though.
17:56 sri the youtube video is really good
17:57 sri and so are the perl6 primitives for concurrency/parallelism
18:01 ajr_ joined #mojo
18:13 btyler_ the moar implementation is still a little fragile, but that'll get better
18:13 btyler_ some bits that should be caught at the p6 level are MVM exceptions, or occasionally segfaults (double closing an async socket, for example)
18:14 btyler_ but that's where having a JVM backend pays off
18:14 neilhwatson joined #mojo
18:16 good_news_everyon joined #mojo
18:16 good_news_everyon [mojo] kraih tagged v6.23 at 7babee1: http://git.io/vCesy
18:16 good_news_everyon left #mojo
18:17 good_news_everyon joined #mojo
18:17 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vCeG3
18:17 good_news_everyon mojo/master 7004bb5 Sebastian Riedel: bump version
18:17 good_news_everyon left #mojo
18:18 sri bpmedley: so, did your minion experiments show any holes in the api?
18:19 btyler_ but yeah, lots of joyous dancing at the prospect of writing socket code like this: http://fpaste.scsys.co.uk/499994 :)
18:19 sri job listing was the part i'm most interested in
18:21 btyler_ even without doing multithreaded/promise stuff, single thread concurrency is pretty nice: http://fpaste.scsys.co.uk/499995
18:21 btyler_ (ok, done, end of topic drift)
18:24 sri btyler_: i like how it completely avoids callbacks
18:24 sri whenever is a pretty great idea
18:26 btyler_ it felt nice for those small examples, for sure. I don't have a ton of intuition if it'll be substantially better for structuring larger things
18:26 bpmedley sri: Not yet, although I haven't gotten very many feature requests and I still need to make some changes.
18:26 sri btyler_: it's just like await, but for observables, so i have a very good feeling about it
18:27 btyler_ ah, cool, yes, good way to think about it
18:27 sri bpmedley: just thought of something, you could use the started/finsihed timestamps to show jobs/s numbers for each job
18:28 bpmedley Awesome, I'll try and implement that in the next few days.  I also need to add support for scrolling and searching jobs.
18:29 panshin joined #mojo
19:04 dave perl6 looks like it would be awesome for writing large scale games
19:05 preaction depends, performance is at a premium for that. the more time you spend in game logic, the less you can spend on pretty
19:05 dave that's what a client is for ;)
19:07 Grinnz_ after having grazed the surface of some MMO codebases, thats an industry i never really want to code for
19:07 Grinnz_ the client, maybe, but the server basically has to be C, and ridiculously parallelizable if you want it to scale at all
19:08 preaction we had a good mmo engine just about ready, but yeah, it's more about the product, and less about code quality
19:08 preaction nah, we were using python, it was fine
19:08 preaction i could do it better in perl now. if you scale horizontally, it's fine
19:08 Grinnz_ for millions of users?
19:08 dave C codebases like that are stuff of nightmares
19:08 dave I think nowdays, for millions of users .. yes
19:08 preaction i didn't run any metrics, but depending on how the sharding worked out, i see no reason to believe we couldn't
19:09 dave the problem is, game devs cant afford "nowdays" :)
19:09 Grinnz_ yeah the database is a big choke point as well
19:09 btyler_ Grinnz_: there are pretty successful game backends in erlang, actually
19:09 dave also, the "game dev" stigma is in effect
19:09 preaction if you can only have 50 users in the same "area" at a time, it's fine, you can scale that forever
19:09 Grinnz_ i'd believe it
19:09 btyler_ call of duty, if memory serves
19:09 dave Most game devs think you can't write software if you have a degree
19:10 preaction and once you have more than that, you have to worry about bandwidth on the clients. each additional uuser in an "area" is an explosion of more traffic for each client to handle
19:10 dave this makes MMO events really laggy ;)
19:10 preaction our plan was some coordinating servers (brokers) negotiating sockets to backend "area" servers
19:11 dave that old network O(N**2) problem
19:11 dave what we need is IP multicast ;)
19:11 Grinnz_ yep... having "channels" for each area makes things a lot easier, then you essentially have your open world instanced
19:11 Grinnz_ transparently
19:11 preaction you can do fun things with invisible players. they're all in the same "area" (instance of the area), but yoou only show the client the players that the client might be interacting with, like party members, friends, similar players, and otherwise
19:12 Grinnz_ its kind of weird for the user but it worked out in a lot of games like tera
19:12 Grinnz_ just gotta present the UI right
19:12 preaction but yeah, you can "instance" any individual area. the old MMO SubSpace had a lot of interesting tech to make 100+ players in an arena work on 33.6 modems
19:13 Grinnz_ and then there's Eve Online, which sort of "instances" deep space mission areas, but doesn't do any instancing of even the most popular station hubs, and it has to do it that way because there's fleet battles of thousands of ships
19:13 preaction yeah, eve has time dilation, which is an interesting solution to the problem
19:13 dave Tera lol
19:14 dave that animation tho
19:14 Grinnz_ before they added that, the server would just lag forever lol
19:15 preaction iirc eve online uses python on the backend
19:15 Grinnz_ partially, but mostly on the client i think
19:16 dave instancing happens on all the modern games now
19:16 dave easy to spawn an instance thread to handle an area
19:17 preaction sure, because that thread doesn't need to be on the same physical machine
19:23 Onigiri joined #mojo
19:39 Kogurr joined #mojo
19:52 melo joined #mojo
20:03 disputin joined #mojo
20:07 jberger sigh: https://news.ycombinator.com/item?id=10347772
20:07 jberger if its obvious to the outsiders, why isn't it obvious to the developers
20:07 jberger this is the argument I've been making for years
20:13 sri sigh
20:14 Grinnz_ the name is forever going to be a huge problem
20:14 Grinnz_ it's unfortunate
20:14 genio yep.
20:14 genio use v5.98;  # sigh
20:15 Grinnz_ damn, now that's modern
20:15 genio heh.  now if they could just adopt Net::FullAuto's versioning ideology.  We could have v5 forevvvvvvver
20:16 Grinnz_ technically, we can anyway; the dotted format allows arbitrarily large integers, it just doesn't convert nicely
20:16 jberger we could do like TeX and MetaFont, is there a universal constant who's integer is 5?
20:24 jberger mostly I want a perl 5 interpreter for whom "use strict;" is no longer necessary
20:24 jberger for the new users
20:24 jberger the users who need it most don't get strict because they don't know to enable it
20:32 jberger on a lighter note: https://blog.twitter.com/2015/usgs-twitter-data-earthquake-detection
20:42 stephan48 my old boss once asked me to tell my trainees to remove "use strict; use warnings;" once a script was finished. i needed quite a bit of time to find a answer to that stupidity
20:43 ZoffixWork joined #mojo
20:43 neilhwatson Please remove the safety from your weapon.
20:45 ZoffixWork jberger, well, it *is* obvious for the devs. But there are good reasons for P6 to keep the name: https://github.com/perl6/faq/blob/master/answers.md#why-not-renaming-to-something-other-than-perl
20:45 stephan48 "and carry it pointing towards your head afterwards please pull the trigger. thank you for making the world a better place."
20:45 ZoffixWork It's just part of our history now. We should embrace it.
20:46 jberger I disagree with each and every point in that list
20:47 ZoffixWork The brand is a very strong point.
20:47 jberger but if you can figure out a way to make small breaking changes in the default behavior of the perl (5) interpreter without it, then so much the better
20:47 jberger the brand is a terrible point
20:47 jberger everyone outside of perl hates perl
20:47 jberger the name has been trashed
20:47 jberger I could argue in large part due to the perl 6 effort
20:48 ZoffixWork #perl6 has about 100 new users today.
20:48 Grinnz_ i wouldnt say i disagree with each point, but that none of them are good enough
20:48 Grinnz_ the brand name is the entire issue with keeping the name
20:48 stephan48 and due to bad coding practices which where "valid" for quite some years
20:48 Grinnz_ perl 5 still not being able to be perl 6 in the future is not really a problem
20:48 disputin joined #mojo
20:48 jberger they still are "valid" and I would love to make them "invalid unless you ask for them (no strict)"
20:48 Grinnz_ the problem is entirely perception of those not familiar with the situation
20:49 jberger there really is no reason not to say that perl6 was a development codename and announce "From the mind of Larry Wall comes the next great language, the successor to Perl, Language X"
20:50 ZoffixWork "successor to Perl"?
20:50 jberger then we release perl 5.26 as perl 7 with signatures and strict by default
20:50 ZoffixWork No one cares about "Language X". They see "Perl 6" and go "oohhh. I haven't heard about that for ages"
20:50 bpmedley In order to win new users I believe we need to focus on the Modern Perl movement.  It's excitement; versatility; and productivity.
20:50 stephan48 i would maybe use fork in that context
20:50 Grinnz_ ZoffixWork: right, and that's cashing in on the very problem that it creates for perl5
20:51 jberger bpmedley: how's that been working for you, other than slowly
20:51 bpmedley Hrmm.  I should evangelize more.  Right now, I'm using the osmosis effect.
20:52 jberger "dear new user, who likely won't find this until its too late, please read these best practices before doing anything else, thanks, Love perl 5"
20:52 ZoffixWork Grinnz_, I'm a bit blurry on the problem P6 creates for P5.
20:52 jberger ZoffixWork: :o
20:53 Grinnz_ ZoffixWork: the link that lead to this discussion
20:53 meredith maybe just showing people code that is Modern Perl(R), not a best practices sheet
20:53 jberger perl6 exists now so perl5 is dead
20:53 Grinnz_ or worse, that perl6 is an updated perl5
20:53 genio "Mu is the base type from which all others are derived."  maaaan.  who's naming things?
20:53 jberger of course that's true because the defaults are so bad that new users just find an unusable language and go back to python/ruby/etc
20:56 ZoffixWork jberger, I find that argument is charged more with emotion than rational thought. You said yourself above that everyone hates Perl, then why P6 a problem? P6 folks are on board with differentiation between P5 and P6 and pointing out that P6 is not an upgrade to P5. I think P5 can coast on P6's publicity actually.
20:57 ZoffixWork People who haven't touched P5 code in decades can now be alerted to some changes. And then realize there are two Perls now.
20:57 Grinnz_ it can't, because the first assumption is that p5 is done
20:58 Grinnz_ as the perl6 homepage clearly points it out, and this is still coming up, it's clear willingness to differentiate when asked isn't going to suffice
20:59 ZoffixWork "as the perl6 homepage clearly points it out" <-- points out what?
20:59 Grinnz_ media is still going to call it a new version of perl, people are still going to assume that and not look into it further
20:59 Grinnz_ that perl6 is not a new version of perl5
20:59 ZoffixWork I don't see that happening.   <ZoffixWork> #perl6 has about 100 new users today.
20:59 ZoffixWork And working Perl 6 hasn't even been released yet.
21:00 Grinnz_ see what happening?
21:00 ZoffixWork That people will assume it's a new version of Perl and not look into it further.
21:00 Grinnz_ i'm not sure what the rest of what you said has to do with that
21:03 sri ZoffixWork: by that logic we should rename mojolicious to catalyst2
21:03 jberger look, for the 15 years that perl 6 has been announced and NOT RELEASED its been hard enough to convince people that perl 5 development continues, what makes anyone think that perl 6 being available is going to make that any better?
21:03 sri so we can all benefit from each others marketing efforts
21:04 Grinnz_ yeah, and i bet the dancer folks love it too..
21:04 sri catalyst3!
21:04 Grinnz_ i mean, since they already have dancer and dancer2 which are simultaneously being developed :P
21:04 jberger that said, oooh, its been on the front page of hacker news all day and there's 100 new users on #perl6, how do you know that if you had said "great new language that makes typing and async easy, like perl but better" it wouldn't be 1000
21:04 ZoffixWork Grinnz_, Larry "unveiled" Perl 6. The #perl6 channel's population increased by 150%. Based on tweets and HN comments people are excited. Hence the assumption that media lauding the "next" release of Perl will make people not look into it further is not supported by evidence.
21:05 Grinnz_ ZoffixWork: i didn't mean "into Perl further"
21:05 Grinnz_ i meant "into whether perl5 is actually still there"
21:05 ZoffixWork sri, Perl is a much stronger brand name than Catalyst.
21:06 jberger ZoffixWork: Perl is a weak brand, made weaker by years of erosion by a vaporware product that confused external potential users
21:06 jberger Perl 6 released in 2005 might have wanted to capitalize on the name, in 2015 I really think it would do better to pick something else
21:06 ZoffixWork Well! If that's your attitude then of course Perl will die.
21:06 ZoffixWork Blame the boogieman.
21:07 sri you can actually verify that "cool new async language" hypothesis by looking at elixir
21:07 jberger sri: exactly
21:07 jberger and go, dear gods that language is odd and has still taken off
21:07 Grinnz_ perl 6's strengths are what it reinvented, it should capitalize on that
21:07 sri perl6 marketing has been a disaster
21:08 meredith yeah we had a guy who did a talk to a polyglot meetup and called it Rakudo and avoided all the dumb perl remarks and drop in interest. he explaiined it at the end.
21:08 sri literally everything about it, i can't think of many things to make it worse
21:08 meredith what exactly is our brand making people think.
21:09 ZoffixWork Well.. I haven't seen much of Perl 5's marketing, TBH.
21:09 ZoffixWork In fact, I didn't even know what logo to give to Rollebar guys when I implemented their API
21:09 jberger "so we have this new language, its like perl ... wait wait, come back, its different, you'll like it I swear"
21:09 ZoffixWork backwards compatibility is not a selling point.
21:09 sri ZoffixWork: that is due to lack of leadership though
21:10 jberger I love perl (5), but most people who have had to use it, especially in legacy, want nothing to do with it
21:10 ZoffixWork ¯\_(ツ)_/¯
21:10 ZoffixWork (gotta go catch my bus)
21:10 jberger well, we aren't going to solve anything here anyway
21:10 dave no but I bet it was fun to talk about ;)
21:11 jberger can you tell I get a little worked up about it?
21:11 dave I passed that phase in 2007 sadly
21:11 dave no sense in getting worked up about it IMO
21:11 jberger well I didn't start perl until 2009, so I guess that puts me in the "defend your perl" zone eh?
21:12 jberger dave: because you have given up
21:12 sri letting off some steam every now and then is good for your health
21:12 dave perl 6 is a gorgeous language ... as usual the smarter people see that and marketing is ineffective vs smart people
21:12 dave I haven't given up
21:12 jberger I want to release a perl-5-like interpreter that uses the defaults that we all use
21:12 dave I'm deciding at the moment whether to write a game engine in perl 6
21:12 jberger so that newcomers get a fighting chance to do things right
21:12 sri personally, i'm looking for an answer to the question why i should pick perl6 over ruby for a mojolicious port
21:13 dave er
21:13 jberger mixins
21:13 dave is that a troll? lol
21:13 sri i like ruby
21:13 jberger dave: I doubt it, sri learned a lot from ruby I know (as he just said)
21:13 dave I liked APL too when I learned it at 12
21:13 sri in fact, i've taken more ruby jobs recently
21:14 dave I haven't taken the time to look more deeply at ruby, it looked cool when I looked at it
21:14 jberger sri: I think the best things about p6 are the ease of creating closures
21:14 dave but you moving mojo to perl6 is a statement
21:14 jberger my brain thinks in closures and perl5 makes them easier than other languages
21:14 Grinnz_ copying, not moving
21:14 dave since you are concerned about marketing :)
21:14 dave copying, yes
21:14 jberger p6 makes it even easier still
21:16 sri we keep saying perl6 is an entirely different language, but i'm getting the impression that's not really how we see it
21:16 dave I submit that you need to use it on a few projects first
21:16 sri most would blindly support a perl6 port, but a ruby port raises eyebrows ;p
21:17 jberger sri: perl5ers keep saying it, I've never been convinced that perl6ers believe it
21:17 dave I raise an eyebrow because you think marketing is important :)
21:17 btyler_ jberger: sheesh, the tribalism :P
21:17 dave If you are being purely technical, then port mojo to the languages that make technical sense to do so
21:18 jberger btyler_: I repeat the question, is it a separate language or not, if so please give me my major version number back so that we can correct a few things that need correcting
21:19 btyler_ going to back away slowly from that debate
21:19 dave If you are concerned about marketing, I daresay Mojo is more associated with perl than ruby and you porting to ruby markets that you are abandoning perl
21:20 btyler_ but I'm not "them". I write perl 5 for money. I'm writing a personal project in mojo right this moment
21:20 btyler_ both of those things generally make me happy. I also write some p6, which also (usually) makes me happy
21:20 btyler_ some occasionaly wtf-ing and re-training aside
21:24 jberger anyway, I should also clarify, I'm also excited for perl 6, and I hope to do some "for fun" coding in it soon (as I have already done (see /Programs/Test/perl6 in my system))
21:25 dave which perl6 do you use?
21:25 pink_mist ssh jberger ls /Programs/Test/perl6 # ssh: Could not resolve hostname jberger: Name or service not known
21:25 pink_mist =(
21:26 sri wow, the elixir slack channel alone has 400 people (with 2400 registered)
21:26 jberger ah, well see, I forgot the ~ in ~/Programs/Test/perl6
21:26 sri re success of new async languages
21:27 pink_mist jberger: I don't quite think that'll help me see it in your system though :D
21:27 dave so success is measured by how many people there are in the channel? :D
21:27 sri today it is
21:27 dave and tomorrow?
21:27 sri i'll make up something else
21:28 dave by that metric PHP is extremely successful ;)
21:28 dave more jobs for that language
21:28 sri i wouldn't disagree at all, php is pretty damn successful
21:28 dave so why is it that the worst tools are the most successful again?
21:29 dave java, javascript, php, ... I'm sure you can name more
21:29 sri java was taught everywhere, javascript is everywhere on the client side, php is everywhere on the server side
21:30 btyler_ elixir is awesome, not surprised it is picking up so quickly
21:30 btyler_ really fun to write, very accessible in its functional-ness
21:32 jberger elixir has no brand recognition, if p6 is as great as we all hope it is, let it stand on its own, honestly I think thats better at this point, I think the collective conciousness about perl is a hinderance to it
21:32 jberger that's not my call I know
21:32 sri i guess this is what a language website should look like http://elixir-lang.org/
21:33 jberger and it does seem self-serving given my other stances, but I do believe it
21:33 Grinnz_ i totally agree jberger
21:34 berov1 joined #mojo
21:34 Grinnz_ the async and threading stuff looks neat and i'd love to see it under its own banner
21:36 dave atoms look neat in elixir
21:36 btyler_ elixir isn't exactly free of brand: the primary dev was a major rails contributor, and through that they got a prominent rubyist (Dave Thomas) to write a book
21:41 ajr_ joined #mojo
21:41 jberger "From the mind of Larry Wall ..."
21:41 hernan605 9~/win 2
21:44 sri lets all just switch to swift
21:44 jberger hahahaha
21:44 jberger tempire might be on board :-P
22:24 sri this perl6 thread is not very positive at all https://news.ycombinator.com/item?id=10345728
22:24 sri compare that to the latest elixir thread https://news.ycombinator.com/item?id=8335360
22:25 sri Zoffix: ^
22:25 sri the brand is no advantage
22:26 sri (not quite latest)
22:28 sri but all release announcement threads look like that
22:29 sri btw. what i find least appealing about elixir is that it's a compiled language
22:29 sri phoenix actually has a build step for compiling templates to bytecode
22:35 * CandyAngel gets the feeling she's doing something wrong >.>
22:35 CandyAngel On the plus side, I'm using Mojolicious in my VR app :P
22:37 melo joined #mojo
22:37 berov2 joined #mojo
22:38 CandyAngel I'm sure initialising GLUT in Class::Tiny's BUILD is fine.. <.<
22:39 CandyAngel >.>
22:47 disputin joined #mojo
22:48 Grinnz_ wait, the ternary operator in perl6 is ?? !! not ?:
22:48 Grinnz_ all i have to say about that is ??!!
22:49 ZoffixMobile joined #mojo
22:56 * jberger mumbles "not the same language" and goes off to watch the cubs game
23:05 meshl joined #mojo
23:12 ZoffixMobile sri, fair enough. The now-drunk me can stipulate that Perl 6 could've benefited from a name change earlier in the game. But I think a wider point I want to communicate to the Perl 5 community is to stop treating Perl 6 as the boogieman that destroyed Perl 5 or *is* destroying Perl 5. I believe the P5 and P6 communities should join together and market *Perl*. One product. Two flavours. This infighting won't help us. *At this point in time*, P6 changing
23:13 Grinnz_ it's not one product though
23:13 Grinnz_ and it's not infighting, just frustration
23:13 pink_mist *At this point in time*, P6 changing <-- your message got cut off
23:14 ZoffixMobile Grinnz_, wait 'till you see custom unicode operators! :) I'm still finding time to release a P6 version of JSON::Meth that will use a unicode "subscript j" for encoding/decoding :p
23:14 Grinnz_ obviously perl6 did not destroy perl5, but it's not helping either
23:14 ZoffixMobile pink_mist, ugh this client sucks
23:15 ZoffixMobile ...its name won't do any good to P5. We have to grow up and deal with the fact that there are two flavours of Perl. Instead of sighing at posts asking about Perl 5's EOL, there should be excited responses of "Hey, we actually have two flavours of Perl! Isn't it awesome?" I think I know how jberger is feeling. A year ago I felt the same: people telling me P5 is dying amd what not. The dread. But getting involved in P6 community alleviated that. I reali
23:15 pink_mist P6 community alleviated that. I reali
23:16 ZoffixMobile .. I realized P6 was not the boogieman. As I expounded earlier, Perl 5th plight is trying to be back-compat at all costs. It tries to pass that off as a selling point, but it isn't, which is why no one cares. And that's precisely why I feel Perl 6's release can be good for Perl 5. But the Perl 5 community has to accept the reality and instead of grumbling behind the curtains, come forward and discuss Perl together with the Perl 6 folks.
23:16 pink_mist that finally looks somewhat complete :D
23:16 Grinnz_ it is a selling point actually
23:16 pink_mist it at least ends in a period :P
23:16 Grinnz_ many sysadmins care
23:16 ZoffixMobile Yeah, that is it. (/me checked on online log)
23:17 Grinnz_ i have nothing against the perl6 community except that the continued use of the perl6 name is confusing people
23:17 jberger All good points, and if I could convey them to outsiders in one word I would
23:17 jberger Hmmmm what weird could that be?
23:17 ZoffixMobile Grinnz_, those sysadmins run "software". They aren't programmers. It's not a selling point for programmers to choose a language to program in
23:17 pink_mist I suggest "blarf"
23:17 Grinnz_ it can be
23:17 jberger s/weird/word/
23:18 Grinnz_ many programs are written for the purpose of being low maintenance
23:18 jberger ANY OTHER NAME AT ALL
23:18 lluad I'm not entirely sure the perl6 folks have any interest in perl any more.
23:18 pink_mist lluad: the perl 6 folks have interest in perl 6 ... surely
23:19 lluad perl6, sure. But not perl (aka perl 5).
23:19 pink_mist well why would they?
23:19 pink_mist it's a different langauge
23:19 pink_mist *language
23:20 ZoffixMobile lluad, if they didn't the perl6.org page would still read "the next version of Perl". They don't care about using it, but they care enough not to stomp on its community.
23:20 lluad So there's probably not that much to talk about with them. Not much more than, say, talking to the go people or the ruby people.
23:20 * ZoffixMobile finds humour in perl6.org currently being generated by Perl 5
23:20 jberger after a long discussion with riba on why perl 5 can't change and this discussion I'm about ready to just say that perl 5 dies on Christmas and move on
23:20 Grinnz_ http://design.perl6.org/Differences.html#foreach_becomes_for has some misunderstandings... that splice is completely unnecessary (should just assign from the array), and foreach/for aren't separate constructs
23:20 pink_mist ZoffixMobile: is xtatic build on mojolicious even? =)
23:21 pink_mist ZoffixMobile: ... or wasn't it xtatic that you put in there?
23:21 ZoffixMobile lluad, there is one thing to talk to them about: the confusion about the namea
23:21 ZoffixMobile pink_mist, put where?
23:21 pink_mist ZoffixMobile: I just noticed your demo on the xtatic.org domain
23:21 pink_mist ZoffixMobile: so I assumed it was actually made with ztatic
23:21 pink_mist *xtatic
23:21 Grinnz_ ZoffixMobile: but there's nothing that can be done about the confusion except for one of them changing the name
23:22 dave just call it perl++ and be done
23:22 * lluad looks at the perl6.org page
23:22 ZoffixMobile pink_mist, nah. It's a static site generated by Perl 5's App::mowyw
23:22 pink_mist ZoffixMobile: oh
23:22 lluad Perhaps it should be renamed to "spunky".
23:23 ZoffixMobile Grinnz_, but there is! Instead of rolling our eyes and sighing, we can calmly explain that there are two flavours of Perl.
23:23 ZoffixMobile Name won't change.
23:23 Grinnz_ and we do that, and we have been doing that
23:23 Grinnz_ and here we are
23:23 ZoffixMobile Good! :) Then we're all set :)
23:23 Grinnz_ clearly not
23:23 * lluad gestures vaguely at python and python3
23:23 ZoffixMobile Why not?
23:24 Grinnz_ because the names are still confusing
23:24 Grinnz_ you're mistaking my objection with an emotional one
23:24 ZoffixMobile Some guy asked a question on HN about Perl 5's EOL, while simultaneously inferring that P5 and P6 are now different languages? I don't see too much of an issue
23:25 ZoffixMobile Grinnz_, perhaps. I *am* currently wasted and am sitting at a bus stop :p
23:26 Grinnz_ python/3 situation is different, as python 3 is the actual successor and python 2 is no longer maintained
23:27 ZoffixMobile To play the Devil's advocate: why won't P5 change its name? :)
23:27 Grinnz_ i did say "one of them changing the name", not "perl6 changing the name"
23:27 * ZoffixMobile smells the can of worms they've just opened
23:28 Grinnz_ and that possibility has been tossed around before
23:28 lluad I've a nasty feeling it'll end up with perl5.86.0 being the actual language and perl6.beta.2018 being no longer maintained.
23:28 ZoffixMobile Oh, sorry. I missed that
23:28 Grinnz_ but perl5 has been perl for eons
23:30 ZoffixMobile lluad, doubt it. Perl 6 has nifty stuff (e.g. grammars). If it can shake off the "looks like line noise" feel, it can become widely uses.
23:30 ZoffixMobile *used.
23:38 Grinnz_ the last example here also seems wrong: http://design.perl6.org/Differences.html#References_are_gone_(or%3A_everything_is_a_reference)
23:38 Grinnz_ surely "$bar is rw" does not create a reference to a reference, as \$bar does?
23:39 Grinnz_ a better comparison would be modifying things by alias in @_
23:39 bpmedley jberger: Thank you for pointing out that my evangelizing of perl5 could use improvement.  As a specific example of trying something different, I created a google ads campaign for Mojolicious.
23:40 ZoffixMobile bpmedley++
23:52 sri Zoffix: it often does feel like the perl6 folks are stomping on the perl5 community
23:53 sri i like perl6 the language, but somehow it doesn't feel right to get invested
23:55 jberger bpmedley++
23:55 sri perhaps to me perl languages now feel like they're not made to last
23:55 sri bpmedley++ # btw
23:56 bpmedley Thx

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