|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
01:00
lizmat joined,
p6bannerbot sets mode: +v lizmat
01:54
lucasb left
03:34
AlexDaniel left
04:16
lizmat left
06:45
domidumont joined
06:46
p6bannerbot sets mode: +v domidumont
06:52
domidumont left
07:08
domidumont joined
07:09
p6bannerbot sets mode: +v domidumont
07:34
robertle joined
07:35
p6bannerbot sets mode: +v robertle
07:47
huggable left
07:53
domidumont left
08:00
buggable left,
buggable joined
08:01
p6bannerbot sets mode: +v buggable
08:08
domidumont joined
08:09
p6bannerbot sets mode: +v domidumont
08:47
zakharyas joined
08:48
p6bannerbot sets mode: +v zakharyas
09:37
lucasb joined,
p6bannerbot sets mode: +v lucasb
09:38
lizmat joined,
p6bannerbot sets mode: +v lizmat
10:54
domidumont left
11:27
Ven`` joined
11:28
p6bannerbot sets mode: +v Ven``
11:32
lizmat left
11:46
brrt joined
11:47
p6bannerbot sets mode: +v brrt
11:49
zakharyas left
12:06
brrt left,
brrt joined
12:07
p6bannerbot sets mode: +v brrt
12:14
Ven`` left
|
|||
| nwc10 | good *, #moarvm | 12:14 | |
|
12:15
Ven`` joined
12:16
p6bannerbot sets mode: +v Ven``
|
|||
| brrt | good * nwc10 | 12:46 | |
|
12:51
domidumont joined
12:52
p6bannerbot sets mode: +v domidumont
|
|||
| nine | lucasb: what exactly is is you want to achieve by creating mod_moarvm that cannot be done with available solutions like cro or CGI? | 12:58 | |
|
13:00
domidumont1 joined
|
|||
| robertle | but let's not recommend CGI on perl6 to annyone, please | 13:01 | |
|
13:01
p6bannerbot sets mode: +v domidumont1,
domidumont2 joined,
domidumont left
13:02
p6bannerbot sets mode: +v domidumont2
13:03
lizmat joined,
p6bannerbot sets mode: +v lizmat
13:04
zakharyas joined,
domidumont1 left
13:05
p6bannerbot sets mode: +v zakharyas
|
|||
| nine | robertle: why? | 13:09 | |
| lucasb | hello o/ | 13:15 | |
| I didn't had the chance to learn and use Cro yet. I know I should do it, and I'll will :) | |||
| I never tried running perl6 as CGI either. I wouldn't expect it to be anything performant, right :) | 13:17 | ||
| I suggested mod_moarvm as an experiment for the sole convenience of Apache users. I wouldn't recommend mod_moarvm (if it existed) to anyone who isn't *already* an Apache user | 13:23 | ||
|
13:23
domidumont2 left
13:24
domidumont joined
13:25
p6bannerbot sets mode: +v domidumont
|
|||
| lucasb | perl6 as CGI is supposed to work. I never cared to try. mod_moarvm would just be a "slightest faster" CGI (for now :) | 13:25 | |
| If I was to consider the platform for something serious and modern in P6, sure I would consider Cro. But, as a desktop computer user, I don't see myself migrating from my personal Apache instance with CGI+mod_php enabled that has being serving me well for the past years. | 13:34 | ||
|
13:36
lizmat left
|
|||
| lucasb | We can all agree the target audience for mod_moarvm is very small (maybe just me? :) | 13:36 | |
|
13:40
AlexDaniel joined,
p6bannerbot sets mode: +v AlexDaniel
|
|||
| nine | perl6 has quite a bit of startup overhead, especially compared to perl. However, it's in the range of ~ 100ms, so it's not actually that bad. Our module loader is pretty fast (once stuff is precompiled) which is why Perl 6 programs may even load faster than Perl 5 ones. | 13:49 | |
| So especially for rarely used scripts, CGI may still be a viable option. For higher usage counts, Cro seems like the way to go. | |||
| moritz | by comparison, ruby has about 80ms startup time | ||
|
13:50
lizmat joined,
p6bannerbot sets mode: +v lizmat
|
|||
| tadzik | my 160loc work time tracker takes about 880ms to do the simpliest thing :/ | 13:51 | |
| (it uses no modules) | 13:52 | ||
| nine | Put it into a module, and create a wrapper script that just uses the module. Instant speedup | ||
| tadzik | hmmm | 13:53 | |
| can a module export MAIN into the importer? :) | |||
| nine | yes | 13:54 | |
| tadzik | (it's a lot of multi mains) | ||
| ok perfect then :) | |||
| nine | zef is just:#!/usr/bin/env perl6 | ||
| use Zef::CLI; | |||
| tadzik | nine: yeah, that brings it down to 290ms :) | 13:58 | |
| wonderful | |||
|
15:28
brrt left
|
|||
| robertle | nine: imho of course, but CGI is a really terrible way to build web applications, not only because of the whole process-per-request thing. together with the templating this is what led to PHP in the first place, all of that could easily have been perl's | 15:29 | |
| again imho, but CGI is and was terrible 20 years ago in perl 5, we certainly should not carry that idea forward. build apps that embed the web server where possible (e.g. with cro) and something like mod_ where you really want to embed into an existing web server (although I see no reason why you would ever want to do that) | 15:31 | ||
| whatever floats your boat I guess, but I think it would be bad to suggest CGI to newbies | 15:33 | ||
|
15:34
Ven`` left
|
|||
| robertle | not that lucasb is a newb of course, but others pick this up as well... | 15:37 | |
| lucasb | robertle++ I understand your opinion and I think it's good to have your voice here to balance things :) | 15:38 | |
| robertle | I mean people talk about the bad press perl has gotten for whatever reason. I am quite sure that CGI.pm was and is one of the main reasons for that... | 15:40 | |
| that said, I have of course done my own share of CGI.pm abuse :) | |||
| lucasb | I like CGI scripts because they are simple, quick and dirty :) I see them as self-contained, just like other command line utilities. I certainly don't call them "web applications", I call them "scripts that spews HTML" | 15:41 | |
| robertle | that's the worst thing: it's so easy and convenient to go down the path towards cgi darkness | 15:42 | |
| japhb | robertle: Do you have an objection to a Perl 6 Plack equivalent? If so, why? | 16:03 | |
| (Curiosity about the edges of your opinion.) | |||
| timotimo | is that kind of what crust is? | ||
| robertle | japhb: fort me plack is a better mod_perl, so yes, it is much better than CGI. I still don't understand why you want to embed your application into a webserver rather than the otehr way around. it's like making your programa plugin to a midnight commander because you are working with files ;) | 16:05 | |
| of course there is the thing with the ports, so I guess in some cases it can't be avoided | 16:06 | ||
| but even then it seems cleaner to me to embed the webserver, run it on some other port and have a routing webserver/reverse proxy in front | |||
| and yeah, my opinion is as fuzzy around the edges as it is strong ;) | 16:07 | ||
| bbiab | |||
|
16:07
robertle left
16:09
domidumont left
16:37
robertle joined
16:38
p6bannerbot sets mode: +v robertle
17:01
brrt joined
17:02
p6bannerbot sets mode: +v brrt
|
|||
| lucasb | with this midnight command plugin analogy, robertle made me think of moarvm in... Vim, haha :D github.com/vim/vim/blob/master/run...f_perl.txt | 17:13 | |
| s/command/Commander/ ! | |||
| Well, a more fanatic Perl6er and Vimmer will arrive in the future with this idea | 17:17 | ||
| Should this fictional character be discouraged and said that "the correct software-engineering way is to embed the text editor in the application, not the other way around!" :D | 17:20 | ||
| but it is a good analogy, well-put | 17:23 | ||
| oh, even irssi embeds Perl 5. but ok enough, sorry for the maybe off-topicness, developers :) | 17:28 | ||
| Kaiepi | should there be a asyncsocketcommon.c file to refactor out shared code between asyncsocket.c and asyncsocketudp.c so codefactor stops screaming whenever changes are made to either file? | 18:02 | |
| robertle | I am seeing some build failures on mips that are new to me: buildd.debian.org/status/fetch.php...&raw=0 | 18:04 | |
| note the "Bytecode stream version too low" | |||
| any pointers on where to start digging? | 18:05 | ||
| same on s390x, both are big-endian architecturs | 18:06 | ||
|
18:07
brrt left
|
|||
| Geth | MoarVM: Kaiepi++ created pull request #1007: Fix some compiler warnings |
18:25 | |
| Kaiepi | timotimo, are there any socket options missing in raw.githubusercontent.com/MoarVM/M...sockopts.c that'd be useful to have? | 18:30 | |
| (i'm aware SO_TIMEOUT doesn't actually exist, i was mixing up c and java) | |||
| i'm thinking SO_REUSEPORT, SO_SNDTIMEO, SO_RCVTIMEO, maybe SO_SNDLOWAT and SO_RCVLOWAT | 18:31 | ||
| timotimo | sorry, i haven't done enough advanced socket programming to know | 18:34 | |
| Kaiepi | welp, thanks anyway | 18:35 | |
| Geth | MoarVM: a095b06388 | (Ben Davies)++ | 2 files Fix some compiler warnings |
18:39 | |
| MoarVM: fdc73f300b | niner++ (committed using GitHub Web editor) | 2 files Merge pull request #1007 from Kaiepi/warnings Fix some compiler warnings |
|||
|
19:01
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build failed. niner 'Merge pull request #1007 from Kaiepi/warnings | 19:01 | |
| travis-ci.org/MoarVM/MoarVM/builds/463484161 github.com/MoarVM/MoarVM/compare/e...c73f300b78 | |||
|
19:01
travis-ci left
|
|||
| Kaiepi | build error seems unrelated to the commit | 19:02 | |
|
19:31
brrt joined
19:32
p6bannerbot sets mode: +v brrt
19:37
lizmat left
19:50
lizmat joined,
p6bannerbot sets mode: +v lizmat
20:23
brrt left
21:44
zakharyas left
|
|||
| Kaiepi | can someone review github.com/MoarVM/MoarVM/pull/993 ? | 22:26 | |
|
23:52
ZofBot left,
ZofBot joined,
p6bannerbot sets mode: +v ZofBot,
huggable joined,
buggable left
23:53
buggable joined,
p6bannerbot sets mode: +v huggable,
p6bannerbot sets mode: +v buggable
|
|||