Raku Conference Day 2 on Youtube: www.youtube.com/watch?v=BL9-XdC9WYI 🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 14 August 2022.
00:04 duuude left 00:07 reportable6 left 00:09 reportable6 joined 00:38 melezhik joined
melezhik .tell lizmat looks like irc log site has no fresh logs again ... 00:38
tellable6 melezhik, I'll pass your message to lizmat
00:42 melezhik left 00:50 xinming joined
xinming m: class A { method t () { "class".say; } } does role { method t () { "role".say; } }; A.new.t; 00:50
camelia Cannot use 'does' operator on a type object A.
in block <unit> at <tmp> line 1
xinming Any way to apply anonymous role to a class directly? Or what is the right way to add "around" method for existing method in the same class? 00:51
ugexe kind of 01:07
m: (class A { method t () { "class".say; } } but role { method t () { "role".say; } }).new.t
camelia role
tonyo with meta model stuff: 01:13
m: class A {}; A.^add_role(role :: { method t {"role anon".say;}}); A.^compose; A.new.t; 01:14
camelia role anon
xinming ugexe: thanks. :-)
long time not play with raku, and I forgot there is "but"
tonyo: I don't like the way to type more than one statements. 01:15
Prefer oneliners to DWIM
tonyo the but method fails depending on context:
m: class A {} but role :: { method t { "role anon".say; } }; A.new.t;
camelia No such method 't' for invocant of type 'A'
in block <unit> at <tmp> line 1
tonyo ugexe's `but` only applies to that `new` specifically
xinming Yea, just tried, and figured that out, Seems your version is the only way 01:17
tonyo if you're into one liners you could put it all on a line with:
m: class A {}.^add_role(role { method t { "anon".say; } }) or A.^compose; A.new.t
camelia anon
tonyo (the compose bit is important so it updates the mop cache) 01:18
xinming yea. thanks
m: class A { method t { "t".say }}; A.^add_role(role :: { method t {"role anon".say;}}); A.^compose; A.new.t; 01:27
camelia t
xinming tonyo: After adding roles, Seems the class method is still being called.
01:38 Colere left 01:43 bigdata joined 01:53 Colere joined
tonyo are you looking for inheritance rather than roles? 02:11
i do believe that's the correct behavior: 02:13
m: role Z { method t {"role".say;}}; class A does Z { method t {"class".say;}}; A.new.t;
camelia class
tonyo kind of out of the loop on how the mop works since a few years ago when it changed a bit and broke the orm i wrote but you can enumerate roles if you anticipate your class being randomly added to: 02:29
m: class A { method t { "class".say; (|self.^roles).grep({$_.first.^can("t").first}).map(*.t()); }; }.^add_role(role :: { method t { "anon".say; } }) or A.^compose; A.new.t
camelia class
anon
tonyo callsame and callnext seem to not work as they do with inheritance
m: class A { method t { "class".say; callsame; }; }.^add_role(role :: { method t { "anon".say; } }) or A.^compose; A.new.t 02:30
camelia class
xinming tonyo: I don't mean inheritance, I'm looking for a way to add wrapper for a method in the class. 02:46
If you used Moose, something like, sub method { }; around method => sub { ... };
tonyo xinming: i'd probably use .wrap instead of the role.. 02:49
m: class A { method t { "class".say; } }; A.^lookup("t").wrap(sub (\self:D) { "after".say; callsame; "after".say; }); A.new.t;
camelia after
class
after
xinming tonyo: Yea, thanks, I just googled and find something like that. :-) 02:50
but your example has more detailed arg. :-) 02:51
02:53 evalable6 left, linkable6 left 02:54 linkable6 joined, evalable6 joined 03:31 bigdata left 04:11 Sgeo_ joined 04:14 Sgeo_ left 04:15 Sgeo left 04:18 Sgeo joined 04:20 Sgeo left 04:37 Sgeo joined 05:18 abraxxa joined 05:22 abraxxa left 05:23 abraxxa joined 06:04 Xliff joined 06:07 reportable6 left 06:10 reportable6 joined 06:28 bigdata joined 06:57 Sgeo left 07:33 duuude joined
leont Within a class, there was a way to specify an "argument of the same type as self", right? 07:58
08:01 Geth left, Geth joined 08:04 Xliff left
lizmat ::?CLASS ? 08:07
tellable6 2022-08-25T00:38:56Z #raku <melezhik> lizmat looks like irc log site has no fresh logs again ...
lizmat .tell melezhik thanks for the nudge
tellable6 lizmat, I'll pass your message to melezhik
08:14 dakkar joined
leont lizmat: thank, that's the one 08:17
it was one of those "google doesn't work for these super common words" cases 08:18
08:21 tadzik joined
xinming leont: That's why human AI bots are needed here. ;-) 08:33
08:51 masak left, sena_kun joined 08:58 duuude left 09:03 duuude joined 09:04 kylese joined
Geth ecosystem: 01807edceb | (Elizabeth Mattijsen)++ | META.list
Freeze Terminal::ANSIColor

While it is being moved to the zef ecosystem		 09:29
09:29 wafflus joined
xinming releasable6: status 09:39
releasable6 xinming, Next release in ≈1 day and ≈9 hours. 1 blocker. Changelog for this release was not started yet
xinming, Details: gist.github.com/1c956ca151c8ca2397...a771dcc48b
09:52 duuude left, kylese left 09:59 duuude joined 10:01 euandreh left, euandreh joined 10:03 euandreh left, euandreh joined 10:04 euandreh left, wafflus left 10:05 wafflus joined
Geth ecosystem: e5d60d2ab5 | (Elizabeth Mattijsen)++ | META.list
Argh, confused Term::ANSIColor with Terminal::ANSIColor

Term::ANSIColor was effectively removed by the previous commit. Since this module lives in the REA, and it has been deprecated since 2017, I saw no need to restore.
This properly freezes Terminal::ANSIColor until it can be removed from the p6c ecosystem in a few days		 10:07
Nemokosch 😂 10:18
10:19 sena_kun left 10:21 sena_kun joined
jjatria lizmat: looks like the change log for Terminal::ANSIColor kept some bad copy-paste from snip: github.com/tadzik/Terminal-ANSICol...Changes#L1 10:26
tellable6 2022-08-18T20:15:36Z #raku <melezhik> jjatria - I have finished first working version of Cromtit - github.com/melezhik/Cromtit , you can use the one to run cron jobs for PublicSuffix - it has nice web UI, jobs dependencies and other useful features
lizmat Oops :-) 10:27
jjatria Things that I notice while trawling the recent releases :P 10:28
lizmat jjatria: thanks, fix committed, will be in next release
wafflus how do i do this in rakue regex101.com/r/3uBm3m/1 10:39
10:41 Xliff joined
lizmat create an error ? 10:43
10:59 euandreh joined
tbrowder .tell skaji i have a PR for mi6 that may be useful to other mi6 users 10:59
tellable6 tbrowder, I'll pass your message to skaji
tbrowder check it out, testing green
11:13 duuude left 11:18 duuude joined
Xliff gist.github.com/Xliff/c7e4b55598de...64c541ca9a 11:20
11:24 euandreh left 11:26 euandreh joined 11:27 euandreh left
lizmat Xliff: better posted on #cro ? 11:42
11:59 wafflus left 12:06 reportable6 left 12:07 reportable6 joined 12:18 duuude left 12:25 euandreh joined
Xliff lizmat: ALSO posted on #cro. 12:26
lizmat Xliff: ah, ok :-)
somehow it had de-highlighted #cro for me already :-(
tbrowder mi6 users should check my PR. it enables you to specify a second rakudoc file to be converted to Markdown. very useful for more detailed information on your repo for web browsing. 12:27
and you can ping skaji if you like it--i need it :-D 12:29
lizmat tbrowder: only a second?
why not have it convert all .rakudoc files in a doc directory ? 12:30
tbrowder for now. i couldn't easily figure out skaji's code. but yes i would like that, too. this is more of a test
lizmat or something like that ?
tbrowder you are welcome to mod the PR 12:31
lizmat I have too many of those already :-)
but should be able to bring out .48 of App::Rak later today, finally
12:31 jgaz joined
tbrowder but i think this pr can stand alone anyway 12:32
there is one glitch on the pr: i had to comment out one of his tests because it fails on me (even without my pr). testing mi6 involves its "xt" directory and i had trouble creating my own pr's test. 12:35
but my test is working, as are all skaji's tests but one 12:36
note my pr uses the "barename" of the second rakudoc to name the markdown file which placed in the base directory with the README 12:39
*is placed
with my pr, skaji could create a separate doc to tell pr submitters the correct way to test mi6! 12:42
12:44 duuude joined
Kaiepi m: class Coolish hides Nil { method Failure { callsame } }; put Coolish.Failure 12:45
camelia Invocant of method 'Failure' must be an object instance of type 'Cool',
not a type object of type 'Coolish'. Did you forget a '.new'?
in method Failure at <tmp> line 1
in block <unit> at <tmp> line 1
Kaiepi ^ hides not recursing caught me off guard. this is kinda nice for what i'm doing, but this isn't documented. is it supposed to do this?
lizmat not sure :-( 12:46
Kaiepi the implementation doesn't seem to have changed since way back in 2012 github.com/rakudo/rakudo/commit/64...346bef2d65 12:52
bar some role stuff in C3MRO iirc 12:53
xinming Anyone here uses DBIish:ver<0.6.5> getting error 'DBDish::Pg: Can't connect: connection to server at "localhost" (::1), port 5432 failed: server closed the connection unexpectedly' randomly? 12:58
On the postgresql server, I can get error which is "could not receive data from client: Connection reset by peer" 12:59
Is there a way to make this error more verbose for easier troubleshooting?
Kaiepi try wireshark
there's --ll-exception ig 13:00
xinming Kaiepi: were you answering my question?
Kaiepi i say that because it seems a bit like a dns problem
yes
xinming Ok
Kaiepi see if it ever tries localhost on ipv4 13:01
xinming termbin.com/l8t9 <--- This is the ll-exception 13:04
hmm, Let me use a local source instead 13:05
termbin.com/f554e <--- This is version with checkout source from master 13:10
Kaiepi: How do I use wireshark to troubleshooting the issue in cli? 13:12
Kaiepi there's a tshark tui tool 13:13
er
just cli
the connection is handled by libpq. not very familiar with it :( 13:15
seems like hyper/race is involved in some way though. maybe it's not thread-safe?
xinming I just now removed the ::1 line in /etc/hosts, Seems the *thing* is gone. 13:16
I'll leave it for a while.
I don't remove the line, I just commented it out, You said it's probably dns issue, So I thought it's probably ipv4/ipv6 mixing issue 13:17
Kaiepi yeah, that's along the lines of what i was thinking 13:18
13:44 linkable6 left, evalable6 left 13:45 linkable6 joined, evalable6 joined 13:56 bigdata left 14:02 morte_ joined 14:22 morte_ left 14:34 ab5tract joined 14:44 melezhik joined
melezhik o/ 14:44
tellable6 2022-08-25T08:07:51Z #raku <lizmat> melezhik thanks for the nudge
melezhik .tell lizmat you are welcome
tellable6 melezhik, I'll pass your message to lizmat
lizmat .
14:51 melezhik left 15:17 human-blip left 15:18 human-blip joined
xinming Is there a way to make DBIish reconnect to database? 15:27
15:28 wafflus joined
wafflus how do i dof i insert named regexes into lookaheads and lookbehinds 15:30
tonyo xinming: are you using ssl on the postgres connection? 15:37
ah, should've read further and you've solved it 15:38
15:47 ab5tract left 16:04 duuude left 16:10 sena_kun left 16:11 sena_kun joined 16:13 abraxxa left 16:16 duuude joined 16:32 dakkar left
Xliff xinming: You might have to close and reconnect. 16:35
tonyo looked like his issue was using ipv6
Xliff xinming: For timeouts... .do('SELECT 1') on the database handle. 16:36
Ahh.... Ipv6 16:37
Yet another tech with great expectations it didn't live up to.
The dotted quad still lives.
The coloned octal never caught on. 16:38
I mean seriously, if I have to use the word "Colon" ever again...it will be too soon. 16:39
lizmat
.oO( everybody wants the colon :-)
 16:49
xinming Xliff: It's not timeouts, I just found now, when I restart the postgresql daemon, The app needs to be restarted again, So I'm thinking wether DBIish have builtin support for reconnect 17:07
Yea, I found it's ipv6 issue, I tried to comment the 127.0.0.1 line in /etc/hosts directly, the ::1 to localhost will also cause connection refused. 17:09
tonyo xinming: i'd use and just auto-reconnect github.com/CurtTilmes/raku-dbpg/bl...akumod#L85 17:21
17:32 evalable6 left, linkable6 left 17:33 linkable6 joined
Nemokosch Insert joke about v6 coming too late to save the world 17:34
17:34 evalable6 joined
jdv more like second system syndrome:) 17:43
Nemokosch Calling it second system syndrome would be as faulty as calling Unix the second system of DOS 17:48
jdv im not aware unix and dos have a strong relation. plus iirc unix came first, no? 17:53
by a decade or more i think 17:54
xinming tonyo: what is the best way to check the ping for the db connection? before each statement, or periodically 17:58
Nemokosch DOS is a broader term than Unix even 18:02
DOS/360 was published in 1966, for example
also, I don't know what sort of "strong relations" IPv4 and IPv6 have, other than targeting the same problem 18:04
18:06 reportable6 left
the point was rather, that IPv4 did have reasons to be replaced even 20 years ago but people rather hacked around it until the problems became manageable 18:07
jdv oh wow. when i think dos i think ibm and ms. is dos/360 strongly related to the 80s/90s dos?
Nemokosch I frankly doubt it 18:08
jdv so you think different versions of the same protocol are not strongly related? 18:09
Nemokosch anyway, IPv4 wasn't created with billions of devices on one network in mind
I think it's merely a name - unless proven otherwise, of course.
18:09 reportable6 joined
jdv im not sure what were talkin about anymore e cept a pissing contest 18:10
fun stuff:)
Nemokosch I wouldn't assume deep relations between two "versions" of a term as generic as "internet protocol"
well, let me reiterate the point then 18:11
> the point was rather, that IPv4 did have reasons to be replaced even 20 years ago but people rather hacked around it until the problems became manageable
jdv ok. my point was vaguely that ipv6 perhaps went too crazy and divergent. 18:13
thanks for the volley
Nemokosch by the way, you made me curious; IPv4 (then just IP, I suppose) was developed by DARPA in the seventies while IPv6 was developed by IETF in the nineties. I really don't see why they would be iterations of one product 18:15
If anything, we can imply that IETF guys weren't as clever as DARPA 🙂
p6steve design by committee 18:29
Nemokosch aaaand we are back to the Perl jokes 👏 18:31
leont It's not possible to add a COERCE method to a subset, is it? 18:38
18:47 japhb left 18:54 japhb joined 19:04 clarjon_1 joined 19:05 archenoth joined, spacekookie left, samebchase left, andinus left, ab5tract joined 19:06 samebchase joined, spacekookie joined, andinus joined 19:07 clarjon1 left, Oshawott left 19:31 smimp joined, smimp left
tonyo xinming: what kind of app is this? 19:33
19:37 japhb left 19:43 japhb joined 19:46 sena_kun left
leont really wants union types, or something like that 20:01
20:26 duuude left 20:51 ab5tract left 20:52 ab5tract joined 21:00 euandreh left 21:02 euandreh joined 21:05 Kaipei joined, jgaz left 21:07 Kaiepi left
japhb lizmat: Your changes to Terminal::ANSIColor seem to have broken Grammar::Debugger. In particular, switching the uppercase "constants" from exported subs to exported constants broke lines like this one: github.com/jnthn/grammar-debugger/...r.pm6#L116 21:14
21:19 ab5tract left 21:27 wafflus left
japhb Oh, I should be clear: I mention this not because it's not trivially easy to patch Grammar::Debugger, but rather because others might be tripping over the same problem. 21:28
22:02 bigdata joined
guifa leont: I've thought about union types, and wanted to experiment with an implementation but I couldn't figure out how to hack into is repr<> 22:04
or at least how that bit worked 22:05
22:08 Oshawott joined 22:11 archenoth left 22:15 Colere left 22:16 Colere joined
lizmat japhb: meh 22:22
22:25 rba_ joined, [dpk] joined, pierrot joined, ecocode____ joined, ProperN[out] joined 22:26 PotatoGim_ joined, BinGOs left 22:27 PotatoGim_ joined, Name joined, Name is now known as Guest2457 22:28 lucs_ joined 22:30 Juerd_ joined, perlmaros_ joined, clarjon_1 left 22:31 ecocode joined, Scotteh_ joined 22:33 eseyman left, Juerd left, ProperNoun left, ecocode___ left, amenonsen left, corwin left, Scotteh left, mykhal left, pierrot_ left, ecocode_ left, tib left, lucs left, PotatoGim left, rba left, perlmaros left, dpk left, rba_ is now known as rba, perlmaros_ is now known as perlmaros, ecocode____ is now known as ecocode___, PotatoGim_ is now known as PotatoGim, Juerd_ is now known as Juerd 22:34 mykhal joined 22:36 Sgeo joined 22:38 tib joined, amenonsen joined 22:40 corwin joined
lizmat japhb: Terminal::ANSIColor 0.7 now on zef, restored the sub interface 22:45
japhb lizmat: OK, thanks. I'll try my full rebuild 22:54
lizmat well, actually added it back in 22:59
japhb lizmat: OK, that got a lot farther in my build list. Now it's failing on Testo:ver<1.003007>:auth<zef:raku-community-modules> -- No such symbol 'Terminal::ANSIColor::EXPORT::DEFAULT::&colored' 23:18
gfldex leont: no, you can't COERCE to a subset. I tried, I really tried: gist.github.com/337fb1f3d8a8bcec0c...56124ff2b1 23:31
23:32 bingos joined 23:40 irc_user joined 23:47 phogg left 23:49 phogg joined 23:50 phogg left, phogg 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!