»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 May 2018.
00:08 w_richard_w joined 00:09 raynold joined
Kaiepi when you retrn a prmoise from an async function the return values not a promise right? 00:22
meaning like
00:23 markoong left
Kaiepi m: sub foo { Promise.start({ 1 }) }; sub bar(Promise $b --> Int) { bar(foo: 1) } 00:23
camelia ( no output )
Juerd Kaiepi: Er, several things are wrong with that line. You're never calling any of the functions, and you're calling bar with an invocant while it's a regular sub. 00:25
Kaiepi yeah ik i fucked up that
lemme write an example at works 00:26
Juerd Well, the latter doesn't actually happen because it's in the definition of bar, which is never called.
I don't understand your question either
I think it's missing at least a verb.
lookatme :) 00:27
Kaiepi yeah probably 00:28
ok so the quesiton is
in node functions return promises 00:29
except async functions, who return the awaited result
so when i return a promise from an async function, does that also resolve the promise?
raschipi There's no "async functions". 00:30
Kaiepi ohhhh right
i forgot about that
lookatme what's async function
Kaiepi good riddance tbh
timotimo lookatme: a function you're allowed to use "await" in
Kaiepi async functions in node nad python lock you ino using await and async only there 00:31
raschipi wut?
Kaiepi yeah
end result: async function spaghetti 00:32
being able to use async stuff however you please is a godsend after fighting with node's ideas of async stuff 00:33
lookatme m: sub foo { Promise.start({ 1 }) }; sub bar(Promise $b --> Int) { await bar(foo) }; say bar; 00:34
camelia 5===SORRY!5=== Error while compiling <tmp>
Calling bar() will never work with declared signature (Promise $b --> Int)
at <tmp>:1
------> 3ise $b --> Int) { await bar(foo) }; say 7⏏5bar;
lookatme :) oh I am foolish
00:35 ohmysummmer joined
lookatme m: sub foo { Promise.start({ 1 }) }; sub bar(Promise $b --> Int) { await $b; }; say bar foo; 00:35
camelia 1
00:42 ohmysummmer left 00:47 lizmat left
b2gills There's a syntactic shortcut for `Promise.start({…})`, and it is `start` 01:02
m: sub foo { start 1 }; sub bar(Promise $b --> Int) { await $b; }; say bar foo;
camelia WARNINGS for <tmp>:
1
Useless use of constant integer 1 in sink context (line 1)
Kaiepi one thing i love about promises in perl 6 is how they're threaded 01:04
b2gills m: sub foo { my $ = start 1 }; sub bar(Promise $b --> Int) { await $b; }; say bar foo; # the above shouldn't have warned
camelia 1
Kaiepi still have flashbacks of hunting down obscure race conditions form node's event leep
plus mutexes for when all else fails 01:05
b2gills If you create a Promise by calling .new, they are not threaded
Kaiepi i was usuing Promise.start
actually wait i've been using it where single threaded work would be what i need 01:07
b2gills There is never a reason to use Promise.start, just use start
Kaiepi do they differ or is it just a style thing?
b2gills m: sub f () { start sleep(1), 42 }; say await f(); say now - $*INIT-INSTANT 01:08
camelia (Nil 42)
1.0720169
timotimo i'd like to advocate against conflating promises with the tasks they represent 01:09
Kaiepi wdym?
timotimo well, "Promise.start" really just gives you a promise object that will at one point be kept or broken by some piece of code scheduled on a scheduler 01:11
Kaiepi yeah
timotimo that's like saying "run this array for me, please" when you call a function that pushes values into the array
the magic isn't the array, the magic is the function that does something to it
Kaiepi i'm aware 01:12
b2gills m: sub start ( &c ) { my $p = Promise.new; $*SCHEDULER.cue({ $p.keep( c() ) }); $p }; say await start( { sleep 1; 42 })
camelia 42
Kaiepi my minn's still suck on how node handles promises 01:13
which tbf isn't the most well designed way of dealing with async the way it was handled 01:14
Zoffix What's a better designed way? 01:15
Kaiepi perl6's ;)
Zoffix Ah :)
Kaiepi node's leads to endless promise chains and async/await spagehetti
timotimo but async/await is supposed to get you away from chains and such? 01:16
i have yet to try what node offers, fwiw
Kaiepi yes and now
async functions tend to force you to use other async functions since there's no way to await a promise outside them 01:17
timotimo oh, that was the whole topic of the conversation above, wasn't it
Kaiepi the point is yeah timitimu
honesly node isn't *awful*, but it does have a lot of quirds 01:19
s/quar(d)s/quar\1ks/
01:20 comborico1611 left
Kaiepi node let's you make linked lists 01:20
the caveat is you need to hack around with classes until you end up with something the npm might praise 01:22
raschipi Why would that be any different in Perl 6?
Kaiepi because what you do is you overwrite the prototype of three classes so they end up becoming chained together 01:23
along with some constructor fuckery 01:24
it's normally stuff that's heavily discourages practive, but hey js has linked lists
raschipi That doesn't sound pleasant.
And "classes" are hashes, isn't it? 01:25
timotimo sorry, i don't understand :)
Kaiepi classes are functions in a dress
timotimo yeah, it's a cool concept to use closures for that purpose
raschipi Perl 5 is that way, using one thing as another, and TimToady explicitly rejected it for Perl 6. 01:26
Zoffix yey, modafuka... I got logs mapper working 01:35
Kaiepi i need to learn perl 5 at some point 01:36
Zoffix Check it out. What does this commit point to? github.com/perl6/doc/commit/792e1f...5d543513e4
irclog.perlgeek.de/perl6/2018-05-23#i_16195729 eh?
irc.perl6.party/perl6/2018-05-23#i_16195729 then this is it
raschipi Zoffix++ 01:37
🎉🎉🎉 01:38
Kaiepi do supplies work sort of like event emitters? 01:39
i've worked channels and promises, but i was a little confused reading about them 01:40
raschipi A Supply is like a Seq, but you don't know when the values wwill come. 01:45
Zoffix In fact, if you add 69.164.222.157 irclog.perlgeek.de to /etc/hosts and then add browser exception for the cert, simply clicking on the links will lead to right place: irclog.perlgeek.de/perl6/2018-05-23#i_16195729 01:47
01:55 MasterDuke_ joined 01:57 MasterDuke left 01:59 MasterDuke_ is now known as MasterDuke 02:01 wamba left
Kaiepi how do they differ from channels raschipi ? 02:09
raschipi In a Supply, everyone listening will get a notification. In a Channel, just one of the readers will get the message. 02:11
Zoffix New blog post: "How To Make Old #perl6 IRC Log Links Work": rakudo.party/post/How-To-Make-Old-...Links-Work 02:13
Kaiepi ok i think i understand now 02:14
thanks
02:29 hythm_ joined 02:39 uzl left 02:45 Zoffix left 03:00 Xliff joined 03:05 MasterDuke left 03:06 hythm_ left 03:18 MasterDuke joined 03:29 raschipi left 03:47 eliasr left 04:05 molaf joined 04:12 kurahaupo_ joined 04:13 kurahaupo left, kurahaupo_ left, kurahaupo joined
buggable New CPAN upload: Magento-0.0.2.tar.gz by SACOMO modules.perl6.org/dist/Magento:cpan:SACOMO 04:34
04:38 Xliff left 04:43 [Sno] left 04:50 molaf left 04:52 |oLa| joined 04:57 curan joined 05:05 espadrine left 05:10 ufobat_ joined 05:17 epony left 05:21 jmerelo joined 05:28 robertle left 05:29 domidumont joined 05:34 epony joined 05:36 domidumont left 05:37 domidumont joined 05:46 [Sno] joined
Geth doc/master: 4 commits pushed by (JJ Merelo)++ 05:51
jmerelo squashable6: status 05:54
squashable6 jmerelo, Next SQUASHathon in 29 days and ≈4 hours (2018-07-07 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
05:58 Kaiepi left 05:59 lizmat joined 06:04 Kaiepi joined 06:08 mcmillhj joined 06:11 kurahaupo left 06:12 lizmat left 06:13 mcmillhj left 06:17 xtreak joined 06:18 domidumont left 06:23 Xliff joined, kurahaupo joined
Xliff \o 06:23
yoleaux 30 May 2018 05:31Z <moritz> Xliff: you might be interested in the cro HTTP client: cro.services/docs/intro/http-client
Kaiepi i never had to use grammars for anything up to now and damn 06:28
never thought i'd parse an entire protocol with regeex 06:29
moritz it's more than just regex :-) 06:30
06:35 robertle joined 06:38 HaraldJoerg joined 06:42 darutoko joined 07:12 Xliff left 07:32 wamba joined 07:36 zakharyas joined, mephinet left, mephinet joined 07:47 zakharyas left, zakharyas joined 07:48 scimon joined 07:53 dakkar joined 07:56 psychoslave joined 08:02 wamba left 08:17 lizmat joined 08:23 xtreak left 08:24 xtreak joined 08:28 xtreak left 08:30 domidumont joined 08:37 epony left
lookatme Is the macro feature available now ? 08:44
08:45 xtreak joined
moritz not really 08:48
lookatme oh, do you know what's the plan ? 08:51
moritz, :)
gfldex .seen masak 08:52
yoleaux I saw masak 5 Jun 2018 12:48Z in #perl6: <masak> to be honest this feels much better
gfldex lookatme: I don't know what the plan is but I can tell you it is hatched here: github.com/masak/007/ 08:54
lookatme hmm, I know this project
thanks anyway
Geth doc/master: 4 commits pushed by (JJ Merelo)++ 08:57
09:00 rindolf joined 09:01 zakharyas left
AlexDaniel Zoffix++ 09:05
awesome temporary solution 09:06
09:08 sftp joined 09:09 mcmillhj joined, xtreak left 09:13 mcmillhj left 09:33 eliasr joined 09:37 xtreak joined, w_richard_w left 09:52 jkramer left 09:57 llfourn left 10:03 wamba joined, zakharyas joined 10:07 parv joined 10:08 parv left 10:44 epony joined 10:54 Kaiepi left 10:59 zakharyas left 11:00 sacomo joined 11:14 lizmat left 11:17 lizmat joined
jmerelo This is an issue that has arisen a couple of times in StackOverflow stackoverflow.com/questions/507231...-in-perl-6 11:28
Is there no way to declare an CArray with a particular size? I have seen you can allocate after declaring it, but not when declaring... 11:37
11:40 raschipi joined 11:50 itaipu joined 12:06 Hor|zon left 12:07 jkramer joined
jkramer m: constant %foo = asd => 123, yxc => 456; 12:08
camelia 5===SORRY!5=== Error while compiling <tmp>
Type check failed in constant declaration of %foo; expected Associative but got List (List)
at <tmp>:1
------> 3constant %foo = asd => 123, yxc => 4567⏏5;
expecting any of:
pos…
jkramer m: my %foo = asd => 123, yxc => 456;
camelia ( no output )
12:08 mcmillhj joined
jkramer Why doesn't this assignment work with constant? 12:08
12:08 shlomif joined 12:09 rindolf left 12:10 mahafyi joined 12:11 Zoffix joined
Zoffix AlexDaniel: what do you mean by temporary? What's the permanent solution, given the original logged data isn't available? 12:11
12:12 shlomif is now known as rindolf
Zoffix jkramer: it will in a month, but for now you need to explicitly give it a Hash 12:12
El_Che Zoffix: do you intend to offer a solution for the historical data or for future logs? Some people expressed reservations about the second
Zoffix m: constant %foo = %(asd => 123, yxc => 456); dd %foo
camelia Hash % = {:asd(123), :yxc(456)}
Zoffix El_Che: we already have a solution for future logs. It was in place for 13 years 12:13
colabti.org/irclogger/irclogger_log...06-07#l245
El_Che i heard for the first time here 12:15
12:15 n1ge left
Zoffix Which is how my mapper is able to map virtually all #perl6 log links 12:18
El_Che great job
12:23 APic joined
raschipi The channel will continue to be public and logged throught colabti, that isn't supposed to change. 12:27
El_Che public and logging are not related at all 12:28
raschipi Right, that's why I specified both. And being public means logged in practice, because someone will log it. Being a "logged" channel means that the logs are published. 12:29
jkramer Zoffix: Thanks 12:34
12:35 n1ge joined
El_Che raschipi: that does not makes any sense to me 12:35
raschipi Sorry, I probably can't explain myself very well. 12:36
12:36 Kaiepi joined, n1ge left
El_Che you can log and publish private channels 12:36
or mine the information
12:36 n1ge joined
El_Che and you can have public interactions without begin logged/filmed/etc 12:37
they public and monitoring can be combined, but one does not depend on the other
mahafyi hello. I have some questions that are probably off topic. I would like to start using a database with some scripts. since i have to choose a database, i see strong recommendations for postgresql over mysql (mariadb in debian). there is no specific application just exploring it. Also, whatever be the DB, i am starting from scratch, so feel free to recommend your choice :).i would like to end... 12:41
...up building a web application at some point. 12:42
El_Che mahafyi: make your app db agnostic? 12:43
mahafyi El_Che : ok. Heh, I am first of all trying to 'get' sql injection which seems to be a wide spread mistake. Can you point me to some example where i can see what a DB agnostic application looks like? 12:45
El_Che mahafyi: whatever uses this and don't use db specific features should do: github.com/perl6/DBIish 12:46
Zoffix mahafyi: PostgreSQL is the best. There's a much better module to interface with it than DBIish: modules.perl6.org/dist/DB::Pg:cpan:CTILMES 12:48
jast to avoid SQL injection, never just include variables inside your query 12:49
12:49 xtreak left
jast well-designed DB libraries (almost all of them nowadays) allow you to put placeholders in your query and pass the values for those as separate parameters 12:50
and then the library will automatically protect you from SQL injection
raschipi "confirm SSL cert exception, because I couldn't figure out how to make it work with proper SSL" Zoffix, it will never work, that's the king of thing SSL is supposed to make impossible in the first place.
Just turn SSL off, better than adding exceptions, which is not recommended. 12:51
El_Che are you recommending to turn ssl off?
if the lib does not support ssl it's better not to use it 12:52
it's 2018
it was bad advise in 1994
Zoffix mahafyi: and don't bend over backwards for "db agnostic". In my experience, if you have such a drastic change that you're forced to change the type of database is very rare and changing a few lines of db code will be a drop in the bucket compared to all the other things you'll likely be forced to do
raschipi Yes I am recommending turning TLS off, because it's part of best practices.
El_Che Zoffix: mahafyi: it depends where your app is supposed to run 12:53
Zoffix I can't turn it off, because the URLs are https:// URLs
raschipi Using self-signed or invalid certs is worse than turning TLS off.
El_Che if it's a personale app for your vps it does not matter
jast agreed, DB agnostic code is really only worth it for software you're writing for the greatest common denominator (stuff random people will install in random environments)
El_Che if other people have to run it it does. A lot of places, pick 1 db and offer a high available setup for the devs
jast if you control the environment, pick a good DB and use its features :)
12:53 xtreak joined
El_Che (where I work we have mysql and oracle) 12:54
you can run your own db, but you'll be playing dba on tv during its lifetime
mahafyi Zoffix : El_Che: jast: very good, postgres it is... does the DB::Pg connect over TLS also when db server is remote? 12:55
El_Che and I agree about posgresql, but having other people doing the admin job trumps that
mahafyi: if you want to learn about dbs, Zoffix advice makes sense as it's less abstraction 12:56
jast from an admin POV I don't like opening DB ports to the public, I tend to tunnel through an encrypted VPN and use unencrypted DB traffic inside that
mahafyi right , seeing that will start from select * from table where column = $var, lol. 12:57
jast but of course TLS easily beats nothing :)
El_Che tls helps against the enemy inside the walls :)
mahafyi jast: oh ok, that i see is a good way
Zoffix mahafyi: you got SQL injection right there :) What if $var has `''; drop table` in it?
El_Che you'll have smaller backups? 12:58
jast a.k.a. the gitlab database compression technique :)
mahafyi Zoffix : what did jast mean by using a placeholder ? can you please show how it works for a simple query to start with? 12:59
Zoffix mahafyi: it'd be `'select * from table where column = $1', $var` The `$1` is a placeholder and the module will automatically perform all the required quoting of data in the $var, so there's no injection
mahafyi maybe i should read up first ...
jast the classic placeholder is '?' (in some libraries), so you're not tempted to think it's a standard variable expansion
pseudocode example: query("select * from foo where name = ?", $name) 13:00
Zoffix mahafyi: that's with DB::Pg module. DBIish uses `?` symbol
for placeholders
jast and the lib will make '$name' safe and insert it where the placeholder was
mahafyi ah ok great learning. i suppose all of this is available to read, but getting perl6 example direct is much better 13:01
jast sure, there are more articles about this than you can read in a lifetime ;)
most of them saying the same things over and over, of course
we're on the internet after all 13:02
Zoffix mahafyi: here's an example with DBIish module:
mahafyi: github.com/perl6/alerts/blob/maste...m6#L34-L40
mahafyi Zoffix : ty
Zoffix mahafyi: notice the place where'd you normally place values just have ? ? ? in them. And the actual values are given in the next statement.
13:02 zakharyas joined
mahafyi yes, i see that , i mean, i get the syntax. 13:03
last question (for some time at least, heh), why is the .prepare desirable as oppsed to just running the query (based on the doc in DB:Pg) 13:04
jmerelo mahafyi: you save some time on SQL interpretation.
mahafyi: might be important in complicated queries.
Zoffix mahafyi: actually, if you're just learning DBs and making small programs for yourself, I have an alternate suggestion: SQLite. Alas, I'm not aware of DB::Pg-quality Perl 6 module for it, so you'd be using DBIish
13:05 jmerelo left
Zoffix mahafyi: it's not really desired in that code. It's just DBIish has a godawful interface (or docs, if I missed an easier way to do it) 13:05
jnthn has been happily using DB::Pg in production...seems pretty solid as well as a nice API :)
mahafyi Thanks. i will try Postgres itself 13:06
Zoffix & 13:07
13:07 Zoffix left
jast the main reason to do a prepare is if you're going to reuse the query (even with different values for the parameters) multiple times during the lifetime of the program, because it saves the database a lot of query parsing and optimizing 13:09
(but you have to make sure to prepare it once, rather than re-preparing it each time, otherwise you don't actually win anything) 13:10
AlexDaniel Zoffix: well, if irc logging is illegal, then colabti will go down also. If it's not, then we can bring the original website back up. So in my mind current redirection is temporary 13:16
but I'm assuming that the logs were not deleted, which I hope is the case…
I have a backup of pretty much the whole thing though 13:17
13:19 raschipi left 13:20 curan left 13:21 ufobat_ left, raschipi joined
raschipi Not using TLS is recommended over using invalid or self-signed certificates. And offers the same level of protection. 13:26
jast except if you have a way to verify the certificate independently 13:27
mahafyi to run a html page, we use a webserver - example apache. is there a tutorial for how to run perl6 scripts via browser? is there a web server for serving a perl6 output or we have to use apache , ngnix etc? 13:28
13:29 skids joined
Juerd raschipi: TLS with a self signed certificate provides much, much better protection than using the same protocol with clear text transfers. 13:29
raschipi: However, in order to verify that the protection is at all valid, you'd need to check the fingerprint of the certificate. 13:30
mahafyi instead of trying something from the cli (using debian), i thought i can try to open in browser itself.
Juerd raschipi: And once you're pinning the fingerprint, it offers *superior* protection compared to regular CA signed certificates.
13:30 Zoffix joined
jast wasn't there a way to record your fingerprint in DNS these days? 13:31
raschipi Right, using it like PGP does work, but that's very different from asking users to add exceptions.
Zoffix AlexDaniel: it's not illegal and moritz doesn't want to deal with any potential claims and legal fees. So neither of your alternatives will occur.
raschipi jast: DANE
Juerd raschipi: Even with exceptions it's slightly better than ditching TLS altogether
raschipi It's not because it teraches users to ignore warnings.And with a self-signed certificate, anyone can sign an equivalent certificate and intercept it. 13:32
AlexDaniel Zoffix: but then someone else can host it, right?
jast now I just need to enable DNSSEC in my zones :)
raschipi teaches*
Juerd raschipi: Because even having to add a new exception when there's a new certificate because of MITM, that does slightly increase the chance of the MITM being noticed. And of course, the effort that one has to put into MITM or wiretapping is increased much just by having encryption, even if no verification is used.
Zoffix mahafyi: check out Cro cro.services/ it has a built-in server 13:33
mahafyi Zoffix ; thanks
Juerd raschipi: It definitely depends on the use case, and the kind of user.
raschipi Noty for normal users, they will just repeat the procedure. For someone who actually understands TLS, sure, but not the case here.
Juerd raschipi: But with regular nerds as the target audience, opportunistic encryption is better than no encryption. 13:34
Zoffix AlexDaniel: "it" is what? Someone else's logs? Moritz doesn't want to give the data from the perlgeek logs
13:34 scimon left
raschipi No, regular nerds have no change of getting it right, I know plenty of them. 13:34
Juerd TBH, I think it's wrong for browsers to favor unencrypted over badly encrypted.
jast not nerdy enough
"no true scotsman" fallacy deployed
13:34 scimon joined
Juerd I'd be fine with them being treated the same if unencrypted got the same big warnigs. 13:34
jast vendors are planning that AFAIK 13:35
AlexDaniel Zoffix: interesting. Well, I have the data, but I wonder if it can be used…
jast but honestly I'd find that way too annoying
*personally* I don't mind visiting unencrypted websites for a wide range of purposes 13:36
raschipi Your argument is that the average nerd knows how to check fingerprints and pin certificates. While in my experience the vast majority don't even know how to check SSL/TLS versions and block unsecure options.
Zoffix mahafyi: and you can then setup apache to reverse-proxy to your app; to whatever port it's running on. Just google for how to do it. Here's a sample of config file for port 80 that has a thing for certbot that'd enable forced-SSH and autogen the config for SSL port : gist.github.com/zoffixznet/4f2a284...096ef54a3b
mahafyi Zoffix , thanks 13:37
raschipi Anyway, if I can be a curmudgeon for a second, please do not teach users to ignore TLS warnings!
Juerd I'll never teach anyone that 13:38
AlexDaniel only up to 2018-04-28 though
Zoffix raschipi: I'll just ignore your warning. 13:39
or request, I mean
raschipi Please do, it's a case of tragedy of the commons after all.
Zoffix It's more of a case of letting people think for themselves. You're not the King of Internet to tell people what they should do with TLS warnings. 13:40
raschipi I'm certainly not, but I can ask for it.
Juerd If I'm going through the trouble of self signing a certificate, I'll also take some time to instruct users to check the fingerprint. But just going with letsencrypt is much easier.
Zoffix: Then who is? 13:41
jast sure, but telling people that it's a great idea to ignore TLS warnings is basically disinformation
Zoffix It's not a self-signed cert. The cert is for 6lang.party. I can't get a cert for irclog.perlgeek.de, because I don't own that domain, duh
jast: which isn't what I'm telling people
Juerd You don't have to own the domain. You only have to successfully mitm it until the certificate is awarded to you.
jast yeah, I wasn't trying to imply that you were
skids The warnings presented need to be reworded to be less technical so nontech users understand the stakes. They need to be (re)written by nontechnical users until they get something they like that technical users can approve of.
Juerd skids: +1 13:42
jast the average person doesn't read the warning in the first place
Juerd skids: And they should include the damn fingerprints so there's some chance of them being checked
Zoffix Nor are my instructions for average persons.
jast case in point: average person tries to use the copier. copier says "hardware fault detected, please contact tech support ASAP". average person mashes "OK" button.
Zoffix Why am I even in this discussion? Don't like my instructions, write your own. 13:43
Juerd skids: I once called my bank, the tax office, and my municipality, to try and verify their certificates. None could transfer me to anyone who understood what a certificate fingerprint was.
Zoffix Nerds.
13:43 Zoffix left
Juerd Did they just call people in a programming language channel "nerds"? How rude. 13:43
:D
jast rudeness is healthy. ;)
13:45 xtreak left
raschipi Any, it was just a PSA, anyone should do what they think is best. 13:45
jast no that's wrong, do what you think is worst :P 13:46
Juerd I might, but only because worst is Dutch for sausage. 13:47
jast boerenmetworst
Juerd worst case => worst kaas => sausage cheese 13:48
skids Zoffix: I meant such instructions in general. I actually have no idea what in particular is being taled about here :-) 13:49
13:50 itaipu left
Juerd skids: Zoffix left 13:50
jast same here... but Zoffix left
13:50 itaipu joined
El_Che Zoffix never really leaves 13:51
Juerd El_Che: What does that mean?
El_Che that his bots ping him
and he joins when desired
Juerd Oh. Not much point in parting then
Or perhaps just for displaying some drama 13:52
13:52 kaare_ joined
El_Che I am an screen+irssi guy, but whatever works for him :) 13:52
Juerd I used to run irssi but the da^W^Wweechat had cookies.
skids No he just leaves routinely. He's entitled to his workflow preferences.
Juerd I see 13:53
jast tmux+irssi > screen+irssi. tools war go! :)
Juerd jast: tmux doesn't do everything screen does, and I can't be bothered to learn to work with both
jast: But I do know that screen is abandonware (as confirmed irl by one of its authors) 13:54
jast good to know (the latter, I already knew the former)
tmux fits my use cases better so I've phased out screen for myself
Juerd Had a hard question and he said he wouldn't answer because screen was considered abandonware anyway. But then I showed hem the antique terminal and the hacks we had applied and he did help me with hacking the remaining things. :) 13:55
jast that's pretty nice of him
Juerd Couldn't do the same stuff with the antique terminal with tmux :D 13:56
skids
.oO(good terminal code is becoming a lost art form)
 13:59
14:00 raynold left
El_Che does tmux fix the horrible edit mode in screen (aka I can't use pgup to scroll back output)? 14:07
14:09 wamba left 14:11 markoong joined
jast you have to press the hotkey once first 14:11
AFAIK the problem is that scrolling is a thing handled by your own terminal (emulator)
so your normal scroll function can't scroll a virtual terminal inside your terminal
El_Che same as screen then 14:12
(there goes my switching use cae)
case
jast in fact it doesn't even make sense if you consider that many terminal emulators combine pgup/dn scrolling with a scroll bar
to even address this at all someone would have to invent a terminal scrolling protocol. maybe someone has already done so, I have no idea. :) 14:13
14:13 zakharyas left, zakharyas joined
Juerd It exists but only program to terminal: a program can instruct the terminal to scroll a part of the screen efficiently 14:14
That's way too limited to implement a scroll bar :D 14:15
jast yeah, that's the extent of what I'm aware off
(also, basic language is failing me. too warm here etc.)
El_Che jast: are you an AI needed more fans? 14:16
jast I am classified as a regular human being, your assumption is incorrect 14:17
I am typing this with my human fingers
El_Che sorry to have hurt your human emotions 14:18
raschipi As human as Mark Zuckerberg, in fact.
El_Che data was one of the most human character in star trek :)
jast you can deinstantiate your failure by making me the overlord of your^Wmy fellow humans 14:19
14:20 mahafyi left
El_Che jast: I voted for the other guy 14:20
14:21 mahafyi joined 14:34 itaipu left 14:36 scovit joined 14:39 kurahaupo left 14:40 kurahaupo joined 14:48 troys joined 14:56 Schepeers left 14:59 Schepeers joined 15:09 robertle left 15:14 raiph joined 15:15 wamba joined 15:18 n1ge left 15:19 n1ge joined 15:20 kurahaupo_ joined 15:21 kurahaupo left 15:22 kurahaupo_ left, kurahaupo joined 15:26 kurahaupo left, kurahaupo joined 15:30 jmerelo joined 15:34 domidumont left 15:41 Luc__ joined, n1ge left, n1ge joined 15:42 mcmillhj left, Luc__ left, mcmillhj joined 15:46 psychoslave left 15:47 mcmillhj left 15:49 Schepeers left 15:51 Schepeers joined 16:29 mcmillhj joined 16:32 scimon left
buggable New CPAN upload: IConv-0.0.2.tar.gz by ALOREN modules.perl6.org/dist/IConv:cpan:ALOREN 16:34
16:40 hlafarge joined 16:42 dakkar left 16:45 molaf joined 16:47 hlafarge left, stee3 joined 16:50 stee left 16:52 stee joined 16:55 stee3 left 16:58 raiph left 17:04 robertle joined 17:06 zakharyas left, zakharyas joined 17:26 Khisanth left 17:30 domidumont joined 17:37 [Sno] left 17:40 Khisanth joined 17:57 Tison joined 18:02 espadrine joined 18:04 zakharyas left
Kaiepi does perl let you use grammars within other grammars? 18:06
timotimo you can either use a rule from another grammar like a regular rule by putting <Grammar::Name::rule>, or you can use other grammar stuff inside code blocks, just make sure $/ is writable in the right place 18:07
Kaiepi ah, perfect 18:08
thanks
timotimo if you literally OtherGrammar.parse inside a code block in your grammar, you'll surely want to use your current match's position as a start point or something like that
b2gills a grammar is just a special type of class, which means you can use inheritance, and compose in roles with tokens
timotimo that's usually more useful, yeah 18:14
Kaiepi wait you can compose in roles with tokens? 18:23
timotimo yup 18:24
Kaiepi that might be really useful for what i'm doing 18:25
18:25 sauvin left
moritz there's, like, a whole book that discusses this kind of stuff :-) 18:27
b2gills In Perl 6 a [regex|token|rule] is just a sub type of method written in a domain specific sub language
timotimo which also means you can use methods written in regular old perl6, but it's interesting
Kaiepi i've heard they're really powerful, but i haven't had the chance to use them since i'd been working on nativecall stuff before 18:28
b2gills My go-to example for what a grammar can look like is JSON::Tiny::Grammar github.com/moritz/json/blob/master...Grammar.pm 18:30
mahafyi in DB:Pg my $pg = DB::Pg.new; # You can pass in connection information if you want. 18:31
But i need an example here please.. how to pass the connection info.
18:32 lizmat left
raschipi b2gills: Is that a complete grammar for JSON? 18:32
b2gills There may still be some edge cases it doesn't properly handle, but yeah it is complete. 18:34
raschipi wow
Kaiepi what i'm doing is writing a telnet library
jnthn mahafyi: It needs a connection string passed as the conninfo parameter, here's an example from one of my apps that constructs it from env vars: gist.github.com/jnthn/b5168f761165...f3f51e8790 18:35
b2gills Kaiepi: Note the use of `proto token` in the above grammar. I don't know if it would be helpful for what you are doing, but it is a very nice feature. 18:36
jnthn dinner, bbl 18:37
mahafyi jnthn: perfect. i was reading www.postgresql.org/docs/current/st...CONNSTRING but couldnt figure how to pass it
Kaiepi it definitely will
jnthn mahafyi: Yeah, the docs could probably use a better example; I thought to send a PR but was in a bit of a hurry with the thing I was building. :)
mahafyi: I'm sure a pull request with a doc improvement there, if you're intersted to do it, would be welcomed ;) 18:38
*:)
jnthn really goes to eat :)
Kaiepi 16 commands, ~200 subcommands, and even more at a lower lever in some places
it'd make parsing much less hell
18:39 darutoko left
mahafyi jnthn : ok i will try and do it (also requires learning how to do that, in the first place). 18:40
b2gills Note that `( "a$_" with $foo )` checks if `$foo` is defined. If it is it does the first part. If it isn't the whole expression returns Empty (empty slip)
Basically with the example provided above it doesn't provide a password if there isn't a DB_PASSWORD in the environment. 18:43
18:47 lizmat joined
Geth doc: 918f4820b4 | 陈梓立++ (committed using GitHub Web editor) | README.zh.md
Update README.zh.md
 18:50
mahafyi ah ok, i think i figured out how to do a pull request now. 18:58
and added one, to add the example given by jnthn 18:59
19:01 Khisanth left 19:07 domidumont left 19:13 imcsk8 left, imcsk8 joined 19:15 jmerelo left 19:16 Khisanth joined 19:24 mcmillhj left 19:29 [Sno] joined 19:33 mcmillhj joined 19:47 zakharyas joined 19:50 mahafyi left 19:52 dct joined
moritz $ docker pull croservices/cro-http 19:54
Using default tag: latest
Error response from daemon: manifest for croservices/cro-http:latest not found
what am I doing wrong? if I understood github.com/croservices/cro/blob/ma...loyment.md correctly, that should work, no?
19:55 mcmillhj left
jnthn moritz: Ah, we only publish versioned ones so far 19:57
No :latest
The example has a version in there and "Always use a versioned base image rather than :latest 19:58
"
Is there somewhere else to clarify it that'd help?
moritz jnthn: that's under "Tips", which I understood to be "you *can* use latest, but you *shouldn't*" :-) 19:59
jnthn Ah :)
Yeah, it's not even an option at the moment I guess
Perhaps we should push a :latest as well
moritz or add a comment that there is no :latest, and instead link to [the list of available tags](hub.docker.com/r/croservices/cro-http/tags/) 20:00
jnthn *nod*
github.com/croservices/docker/blob...-images.p6 would be the script to patch to push a latest
(If you're sufficiently keen to patch it, that is... :))
We should really make links to that page anyway
moritz I'll add some 20:01
20:01 mcmillhj joined
jnthn moritz++ # thanks! 20:02
I was about to make an issue to not forget it (gotta go afk in a moment), but if you'll sent a patch that's even better :)
moritz github.com/croservices/cro/pull/71 20:06
20:10 uzl joined
rindolf moritz: hi, sup? 20:28
jnthn: sup/
20:28 kurahaupo_ joined 20:30 kurahaupo left
uzl Hi to everybody! 20:33
AlexDaniel hello!
uzl yesterday I posted about an example of a JSON grammar in Think Perl 6 and in the next section, the author adds an actions class to it and it's not working as expected for me.
the example in the book: shorturl.at/HRWX8 20:34
20:34 psychoslave joined 20:35 kurahaupo_ left, kurahaupo joined
moritz uzl: token string needs a pair of parenthesis for $0 to work in the string action method 20:36
token string { \" ( <[ \w \s \- ' ]>+ ) \" }
with that modification, I get this output:
{dob => {day => 9, month => 11, year => 1934}, lastName => Sagan, name => Carl, occupations => [atronomer author lecturer]}
uzl moritz: Wow, that was so simple! 20:37
moritz: Thanks!
moritz uzl: you're welcome 20:39
uzl moritz: does that mean that $0 in the method string is making a reference to the capture group in token string?
20:39 wamba left
moritz uzl: correct 20:40
20:40 wamba joined
uzl moritz: awesome! 20:40
moritz uzl: if you want an even deeper dive into regexes and grammars, there's smile.amazon.com/dp/1484232275/ 20:43
20:46 damnlie left 20:47 damnlie joined
rindolf uzl: hi 20:47
20:49 comborico1611 joined, Possum joined 20:52 comborico1611 left, zakharyas left
uzl moritz: I'll take a look at it. Thanks! 20:53
rindolf: hello!
21:02 skids left 21:06 mcmillhj left 21:09 Xiti joined, Xiti left 21:11 Xiti joined, Xiti left 21:16 raschipi left 21:24 rindolf left 21:31 uzl left 21:41 mcmillhj joined 21:43 epony left 21:46 mcmillhj left 21:52 Xliff joined
Geth doc: kjkuan++ created pull request #2094:
Fix missing nodejs dependency in the Dockerfile.		 21:54
22:00 wamba left 22:02 molaf left 22:06 dct left 22:13 psychoslave left, psychoslave joined 22:19 psychoslave left 22:25 raynold joined 22:26 robertle left 22:32 mcmillhj joined 22:37 mcmillhj left 22:43 HaraldJoerg left 23:06 mcmillhj joined 23:10 mcmillhj left 23:14 markoong left 23:18 comborico1611 joined 23:19 kurahaupo left, kurahaupo joined 23:20 mcmillhj joined 23:25 mcmillhj left 23:30 [particle]1 joined 23:32 mcmillhj joined 23:33 [particle] left 23:37 mcmillhj left 23:44 Kaiepi left 23:59 mcmillhj joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!