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

IRC log for #mojo, 2016-09-20

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

All times shown according to UTC.

Time Nick Message
00:37 Grinnz 20:36:23 — GumbyPAN CPAN Upload: Minion-Backend-SQLite-0.007 by DBOOK https://v1.metacpan.org/release/DBOOK/Minion-Backend-SQLite-0.007
01:24 laidback_01 joined #mojo
01:37 itaipu joined #mojo
01:47 cpan_mojo Mojo-IOLoop-Subprocess-Sereal-0.001 by DBOOK https://metacpan.org/release/DBOOK/Mojo-IOLoop-Subprocess-Sereal-0.001
02:03 mcsnolte joined #mojo
02:17 noganex joined #mojo
02:24 itaipu joined #mojo
03:36 zivester joined #mojo
03:53 trobotham joined #mojo
04:22 polettix joined #mojo
04:22 jberger Bah, didn't get to the release tonight. I will in the morning. Plus that will let me run the tarball past someone first
05:22 trobotham joined #mojo
05:58 dod joined #mojo
06:07 dod joined #mojo
06:24 polettix joined #mojo
06:30 inokenty-w joined #mojo
06:41 AndrewIsh joined #mojo
06:53 dod joined #mojo
07:12 mbudde joined #mojo
07:13 Dandre joined #mojo
07:20 Vandal joined #mojo
07:26 trone joined #mojo
07:44 trobotham joined #mojo
07:44 ladnaV joined #mojo
07:58 jnbek joined #mojo
08:09 osfabibisi joined #mojo
08:17 trobotham joined #mojo
08:38 polettix joined #mojo
09:10 Vandal joined #mojo
09:36 irqq joined #mojo
09:50 Vandal joined #mojo
09:56 * sri yawns
10:26 dotan_convos joined #mojo
10:30 sri at this point i guess our best hope for http2 support would be an nghttp2 binding for perl
10:32 tyldis joined #mojo
10:32 sri which has a pretty big api https://nghttp2.org/documentation/apiref.html
10:51 nicomen sibbolethish language, po ruski ;)
11:06 jamesaxl joined #mojo
11:45 kes joined #mojo
11:48 polettix joined #mojo
11:56 sri it's official now, i'm joining SUSE :) https://twitter.com/kraih/status/778187976547397632
11:57 dotan_convos joined #mojo
11:57 bd still on mojolicious?
11:57 coolo sri: after the mandatory medicine check of course :)
11:58 sri :o
12:05 jberger sri: congratulations officially!
12:06 Sebbe sri: Congratulations!
12:06 pink_mist congrats!
12:11 anton joined #mojo
12:19 stephan48 sri: congratulations!
12:19 zivester joined #mojo
12:20 stephan48 sri: how effective are the hyperx headphones at blocking outside noise?
12:21 good_news_everyon joined #mojo
12:21 good_news_everyon [mojo] jberger pushed 1 new commit to master: https://git.io/viydX
12:21 good_news_everyon mojo/master 2c9d037 Joel Berger: update Changes
12:21 good_news_everyon left #mojo
12:24 sri thanks!
12:24 sri stephan48: just get the already
12:25 sri *them
12:25 stephan48 tomorrow :D
12:25 jberger sri: ok I updated the date
12:25 stephan48 made the mistake of ordering them a few minutes to late on amazon
12:26 jberger make manifest, make test, make distcheck, make disttest
12:26 sri very good at noise cancelling btw. since you basically have closed aluminium cans over your ears
12:26 jberger looked over the MANIFEST
12:26 jberger what am I missing?
12:26 sri release and bump the version
12:28 sri i think for good headsets there are only two choices really http://amzn.to/2cOejEa and http://amzn.to/2d1zDth
12:31 good_news_everyon joined #mojo
12:31 good_news_everyon [mojo] jberger tagged show at master: https://git.io/viybf
12:31 good_news_everyon left #mojo
12:31 jberger ok I found my screwup :-P
12:31 sri the hyperx seems to be so good because they've shamelessly copied from beyerdynamic
12:32 sri jberger: :O
12:32 jberger I hate git tag
12:32 jberger luckily that's an easy fix
12:32 sri well, you can always tag later
12:32 stephan48 thats quite a small price difference
12:33 good_news_everyon joined #mojo
12:33 good_news_everyon [mojo] jberger deleted show at 2c9d037: https://git.io/viybC
12:33 good_news_everyon left #mojo
12:34 jberger funny, I did make v7.07 tag, it just didn't announce it here
12:37 good_news_everyon joined #mojo
12:37 good_news_everyon [mojo] jberger pushed 1 new commit to master: https://git.io/viyNf
12:37 good_news_everyon mojo/master 11cc5ce Joel Berger: bump version
12:37 good_news_everyon left #mojo
12:38 jberger and I got the index successful email
12:39 sri jberger++
12:39 jberger I mostly get nervous about EUMM now
12:40 jberger the rest of it isn't that bad
12:41 sri what is your release workflow normally these days?
12:44 itaipu joined #mojo
12:46 jberger App::ModuleBuildTiny for authoring, Module::Build::Tiny for install
12:48 jberger so its basically: prove; mbtiny regenerate; mbtiny dist; mbtiny shell (opens a new shell into an unzip of the dist, prove, look at META* and MANIFEST); mojo cpanify
12:48 jberger not that different really, just different :-P
12:49 sri lol
12:49 sri that's quite a few extra steps ;p
12:50 sri git tag; make test; make manifest; make dist; mojo cpanify;
12:50 sri basically all i ever do
12:51 sri maybe make test before git tag ;p
12:51 sri using this oneliner instead of make test though
12:52 sri make clean; perl Makefile.PL; TEST_POD=1 TEST_HYPNOTOAD=1 TEST_MORBO=1 TEST_PREFORK=1 TEST_EV=1 TEST_ONLINE=1 TEST_IPV6=1 TEST_TLS=1 TEST_SOCKS=1 TEST_SUBPROCESS=1 HARNESS_OPTIONS=j9 make test
12:52 jberger well, some of those are mine, checking the META* and MANIFEST are things I do with any build system
12:52 sri i only create the manifest file on demand, so i never have to check it
12:52 jberger and I like to do a "dist test" whether its via eumm's make disttest or mbtiny shell && prove
12:53 jberger I only make it on demand too, but that doesn't mean that something nasty didn't slip in, like a local file that got forgotten about
12:53 batman sri: i use App::git::ship: "prove -l && git build && git ship"
12:54 jberger worst are files that are in .gitignore but not in MANIFEST.SKIP
12:54 jberger so you do git status and you don't see any untracked files, but they can still show up in the MANIFEST
12:54 sri hmm, that just doesn't happen to me ;p
12:55 * sri keeps his git repos clean
12:55 batman me neither... "git build" won't let me run that command if the repo is dirty
13:05 gizmomathboy joined #mojo
13:13 ramortegui joined #mojo
13:21 * jberger also just learned about osx color calibration expert mode
13:21 jberger zomg it looks so much better!
13:23 * jberger turns down terminal opacity even more now
13:37 kaare_ joined #mojo
14:02 mcsnolte joined #mojo
14:11 zivester joined #mojo
14:15 mib_zodsx6 joined #mojo
14:15 ptolemarch joined #mojo
14:15 mib_zodsx6 Hi everyone. I
14:15 sri o/
14:17 mib_zodsx6 I'm using Mojo::DOM->all_text quite a lot in 6.66, but in new versions smart trimming was removed. Please help me to get same results of all_text in 7.07 as it was in 6.66
14:18 sri you can't i'm afraid
14:18 sri we want to bring back trimming, but have so far been unable to figure out how to do it right
14:18 PryMar56 joined #mojo
14:19 mib_zodsx6 thanks, will search for workaround :)
14:22 ramortegui joined #mojo
14:33 lluad joined #mojo
14:42 itaipu joined #mojo
15:07 * Grinnz 's release process: dzil release
15:08 Grinnz also, Git::GatherDir ignores gitignored files for me :P
15:10 jberger what about gitignored files that you DO want in MANIFEST?
15:11 Grinnz such as?
15:11 jberger META*
15:11 jberger MANIFEST
15:11 Grinnz those are generated at build
15:11 jberger ah ok
15:11 jberger and it knows how to tell the difference I assume
15:12 jberger anyway, I've already dzil-ed and undzil-ed all my dists, with mbtiny available I don't think its likely I'll re-dzil again
15:12 jberger but we've had this discussion
15:12 pink_mist jberger: difference between what?
15:12 Grinnz there's a very distinct"root" (physical files) and "build" (what ends up being put in the tarball)
15:13 Grinnz root also called source in testing
15:13 jberger pink_mist: files that were generated and should be in the build but shouldn't be in git vs other files that aren't generated during build and thus shouldn't be in git or the manifest
15:14 Grinnz the build happens in a temp dir
15:15 dod joined #mojo
15:15 haarg Distar uses a whitelist for manifest, and forces a disttest on release.  pretty much always catches packaging issues.
15:15 Grinnz META.json is copied back to root in my bundle, but that's ancillary to the build process, so the repo is installable without dzil
15:16 Grinnz same with generated Makefile.PL/Build.PL
15:17 asarch joined #mojo
15:22 Grinnz which reminds me, I need to check out that regenerate phase if kentnl released it yet
15:23 haarg i think it's still RFC
15:43 cpan_mojo Mojolicious-Plugin-DbicSchemaViewer-0.0200 by CSSON https://metacpan.org/release/CSSON/Mojolicious-Plugin-DbicSchemaViewer-0.0200
15:49 jberger sri: can I get a reminder what the security implication is that causes Mojo::Path to need to normalize %2F?
15:49 jberger just got a question on another chat medium
15:50 jberger I've given the questioner another option, but he is curious anyway
15:50 jberger (he's trying to use a string value in a url_for which might have a / so he thought he could just urlencode the value before using it as a placeholder value, it doesn't work)
15:51 sri path https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2011-1589
15:51 jberger ah
15:51 jberger wow, I don't think I've ever seen that before
15:51 jberger (it is well before my time hanging around mojoland)
15:55 disputin joined #mojo
15:58 disputin joined #mojo
15:58 sri it was a very hectic time, since the one who found it was malicious
15:58 sri and made it public before contacting us
16:00 Grinnz :/
16:00 genio That's a crappy thing to do
16:00 Grinnz jberger: yeah, if you have a value with / generally the only good place to put it is the query string
16:01 Grinnz we have to do that with one value which would normally be better suited as a placeholder
16:01 jberger wow, that sucks!
16:02 jberger in the questioner's real use case, the answer is simple, they are using base64 encoding for something and they should use base64url
16:02 jberger but it did beg the question, I'm glad I asked, I learned something
16:03 Grinnz yeah there are diffferent base64 encodings for a reason
16:03 jberger but you have to know they exist
16:03 jberger lots of the sha modules provide a base64 function alongside
16:03 jberger but those are the non-urlsafe forms
16:04 jberger to the layperson they don't know that a base64url is an option
16:05 sri doesn't look like this nghttp2 binding is getting anywhere :/ https://github.com/gonzus/http-nghttp2
16:06 jberger are bindings to the c library really the best way?
16:06 jberger I guess it is how most others have done it
16:06 sri i think so
16:07 sri yea, everybody does it with nghttp2
16:08 sri and it shouldn't be too hard to integrate into mojolicious if the binding is good
16:08 dod joined #mojo
16:10 sri it can basically be a blackbox we feed bytes and that gives us http1.1-ish looking data (and the opposite)
16:13 sri doing the actual http2 protocol in perl seems a bit silly, since it's mostly compression
16:13 sri which is clearly better done in C
16:19 sri adding http2 support might even turn into a weekend project with a good nghttp2 binding
16:23 sri that's the good thing about alpn negotiation, you know right away you're only dealing with http2 and can let the blackbox deal with everything
16:24 sri heh, alpn is still waiting in a branch https://github.com/kraih/mojo/compare/alpn
16:26 good_news_everyon joined #mojo
16:26 good_news_everyon [mojo] kraih deleted zero_length_header_value at 1ef8a57: https://git.io/viS23
16:26 good_news_everyon left #mojo
16:27 good_news_everyon joined #mojo
16:27 good_news_everyon [mojo] kraih deleted data-line-in-error at 75b2b9a: https://git.io/viS2Z
16:27 good_news_everyon left #mojo
16:27 good_news_everyon joined #mojo
16:27 good_news_everyon [mojo] kraih deleted c_log at 8282e31: https://git.io/viS2u
16:27 good_news_everyon left #mojo
16:29 good_news_everyon joined #mojo
16:29 good_news_everyon [mojo] kraih deleted twemoji at 3e8bb80: https://git.io/viS2M
16:29 good_news_everyon left #mojo
16:29 good_news_everyon joined #mojo
16:29 good_news_everyon [mojo] kraih deleted json_large_integer at 9af9f10: https://git.io/viS2y
16:29 good_news_everyon left #mojo
16:29 sri don't mind me cleaning the closet
16:30 Grinnz https://www.youtube.com/watch?v=RQ9_TKayu9s
16:33 good_news_everyon joined #mojo
16:33 good_news_everyon [mojo] kraih deleted log-format-rfc-3339 at 938556a: https://git.io/viSaW
16:33 good_news_everyon left #mojo
16:35 sri jberger: you should remove some of the channel branches
16:37 dod joined #mojo
16:38 jberger man, there are no good youtubes of the R. Kelly "why won't Tom Cruise come out the closet?" song
16:38 jberger sri sure
16:38 jberger (forgot to say **from southpark**)
16:41 jberger honestly, I think we should just nuke the channel branches from orbit
16:41 jberger I'd have to audit them to know which ones were worth keeping
16:42 Grinnz i don't watch much south park but i think i remember that episode :P
16:42 disputin joined #mojo
16:44 jberger ok I looked through the commit graph
16:44 jberger channels/master is the most advanced, I'm nuking the rest
16:45 good_news_everyon joined #mojo
16:45 good_news_everyon [mojo] jberger deleted channel_jberger_3 at ee08396: https://git.io/viSwn
16:45 good_news_everyon left #mojo
16:45 good_news_everyon joined #mojo
16:45 good_news_everyon [mojo] jberger deleted channel_jberger_4 at 04f4a60: https://git.io/viSwc
16:45 good_news_everyon left #mojo
16:47 jberger awwwww, I forgot about :scope selectors .... https://github.com/kraih/mojo/compare/css_scope
16:47 jberger :'(
16:56 sri yea, those were interesting
16:56 sri most pressing issue with Mojo::DOM is text trimming though
16:57 sri with there was a simple browser api method we could mimic
16:57 sri s/with/wish/
16:59 jberger all of that happened essentially while I was traveling to vacation, which was the absolute worst time for me
17:00 jberger since I missed most of the arguments, and then it was "settled" by the time I landed
17:00 jberger and since it was my vacation I didn't dig deeply back into it
17:00 jberger if someone (and this doesn't have to be sri) could make a bullet-point summary of the problems and required solutions I could take a look
17:00 jberger but I've never completely understood the problems
17:01 sri short summary, we only handled <pre>, and people thought we also should handle <br>, citing certain browser/jquery api methods, and then nobody knew what to do
17:02 sri at least that's how i remember it
17:02 sri think there were other whitespace trimming inconsistencies too
17:02 Grinnz there was some inconsistency when trimming adjacent tags IIRC
17:03 sri now that you mention it, i remember that too
17:05 sri https://github.com/kraih/mojo/issues/974
17:05 jberger so my first thought is, how wrong would it be to extract all the text elements from an outer element, trim them each individually, then join them with spaces
17:05 jberger ?
17:05 sri lol, the issue was joining them with spaces
17:05 sri which was wrong
17:06 sri if it looks like an easy problem you're either a genius or don't get it yet ;p
17:07 go|dfish aa
17:07 go|dfish sigh, sorry
17:08 jberger so surely there must be some actual way to do this, browsers can after all. I'm assuming it would just be prohibitively hard to implement correctly?
17:09 jberger then the question becomes, how correctly can we do easily; which is where I'm assuming things broke down
17:12 sri yes
17:12 jberger is it known which tags preserve whitespace?
17:12 Grinnz i don't think anyone has tried that hard, because there isn't really a standard for it so it would be reverse engineering
17:13 jberger because generally html isn't whitespace preserving IIUC
17:13 Grinnz between what browsers do and what javascript libraries do
17:13 sri since nobody has found it yet, i guess we can assume there is no actual DOM/jQuery api for getting formatted text
17:14 jberger generally I think what people want is long blocks of consecutive whitespace normalized so that it can be easily compared with 'is'
17:14 jberger from a use-case perspective
17:15 jamesaxl joined #mojo
17:22 bombuzal joined #mojo
17:26 jberger so I'm coming up with a few ideas
17:26 jberger and it starts with a better "squish" function
17:47 irqq joined #mojo
17:52 jberger something like this maybe: https://gist.github.com/jberger/78abb3e9cc0bc6c98b12181494d0cffd
17:52 jberger but then again, that's just off the top of my head
17:52 jberger anyway, lunchtime
18:17 sri jberger: you'll never reach consensus on behavior
18:18 itaipu joined #mojo
18:41 jberger yeah, I guess that's why finding a spec or published algo is so necessary here :s
18:49 * sri nods
19:34 genio sri: Anything that would make you suggest that I don't upgrade to MacOS Sierra?
19:35 sri nope
19:35 genio \o/ seemless upgrades are nice
19:35 genio *seamless
19:35 sri it's a rather unspectacular release
19:36 sri siri is nice, but the novelty wears off quickly
19:36 sri only real improvement i've seen is monotonic clock support in Time::HiRes
19:46 sri bit of a bummer that linux desktops are not better yet at handling HiDPI displays
19:47 sri it's fun trying to decide between gnome 3 and kde when you've only connected to linux boxes via ssh for so many years :)
19:48 lluad Which one has the better terminal app? That's the only thing that matters. :)
19:51 vicash joined #mojo
19:52 sri gnome wins there imo
19:52 juikuen joined #mojo
19:53 sri sadly it looks like kde is better at dealing with multiple HiDPI displays with different scaling settings
19:54 sri not sure why gnome 3 gets so much hate, i like the style
19:54 sri btw. i hope you don't mind me talking more about linux in the future :D
19:55 Grinnz i dislike that gnome 3 removes so much from the desktop interface
19:55 Grinnz so i use cinnamon, which uses the same backend
19:57 sri coming from os x, gnome 3 feels very familiar
19:58 sri gotta say one thing, font rendering is great for both, qt and gtk with a scaling factor of 2
20:00 jberger when I moved from ubuntu to osx it felt very natural so I'm not surprised to hear it works the same going the other way
20:00 sri ubuntu unity is just a tiny layer on top of gnome 3, right?
20:02 jberger I don't know if it actually uses gnome3, I guess it probably has to
20:02 jberger but it came out of disagreements with gnome during the 2->3 thing, so its going to be similar either way
20:03 sri with the gnome terminal, atom and chrome i already have my usual work environment covered i suppose
20:03 sri all feel very similar to the os x version
20:03 sri and look good with 2x scaling
20:09 sri i have not even edited a config file to set things up
20:16 sri kde plasma is really slow
20:18 Grinnz my specific problems with gnome 3 are 1. no taskbar ??? like it's trying even harder than windows to be a tablet OS, and 2. when i used it, it had no concept of persistent notifications, especially since there was no taskbar or application tray
20:19 Grinnz thus a significant part of my workflow was negated
20:19 Grinnz you can do "anything" to it with extensions, except these extensions are maintained by the community and thus break every release
20:20 Grinnz whereas cinnamon has the expected taskbar, the expected notifications support, and the expected features so i don't need extensions
20:20 Grinnz /rant off :P
20:21 Grinnz i don't think unity uses gnome 3, they're two separate initiatives
20:21 Grinnz or at least, the gnome-shell backend
20:23 sri not sure what you mean, there is a task bar
20:23 Grinnz with all of the open applications?
20:23 sri yes
20:23 Grinnz huh, did they give up on that?
20:24 sri when you hit Activities
20:24 Grinnz err, not the overlay
20:24 sri why would i want to waste screen realestate with a permanent task bar?
20:25 Grinnz i can't work without it
20:25 sri all active apps have a spot in the top bar, and all minimized ones are in the overlay sidebar
20:25 Grinnz wait, they added minimize too?
20:26 sri no wait, i'm wrong, the top bar only shows the currently active app
20:26 Grinnz last i remember, they were still saying how minimizing apps is unnecessary
20:27 sri but yes, you can minimize to the overlay
20:27 Grinnz (which I agree with, but wouldn't implement in a desktop that people are supposed to like)
20:27 sri and alt-tab to bring back minimized apps
20:28 sri how is cinnamon with HiDPI displays?
20:28 Grinnz there's an overarching policy to it that really pushed me away from gnome, where they know the one way you should use the desktop and if you want somethign else you better code it yourself
20:29 Grinnz same as gnome, it's all GTK3 apps
20:29 Grinnz i use hidpi on one of my computers but not actively
20:34 Grinnz speaking of wasting screen real estate... http://www.tuxmachines.org/images/fedora15_gnome3-installer.jpg there's a lot of empty taskbar there :P
20:34 Grinnz thats about the last time i used gnome 3 too, so maybe it's different now
20:36 Grinnz actually this is what it looked like 1920x1080 http://i.imgur.com/BEDzs.jpg
20:36 Grinnz or 1200 in that case
20:37 sri that's true, lots of wasted space in the task bar
20:38 sri that's a horrible theme :o
21:18 cpan_mojo Mojolicious-Plugin-BootstrapHelpers-0.0203 by CSSON https://metacpan.org/release/CSSON/Mojolicious-Plugin-BootstrapHelpers-0.0203
21:32 sri actually tried cinnamon, not impressed
21:32 sri feels way less polished than gnome 3
21:32 sri like the chinese knock off of kde plasma
21:32 sri ;p
21:34 itaipu joined #mojo
21:41 Grinnz did you try the settings dialog :P
21:46 sri yes i did try settings, but cinnamon is just not for me
21:47 sri i kinda want to like kde plasma, but it just doesn't have any character
21:47 sri it's almost nice looking, but so many parts feel like someone just went through a checklist and didn't put real thought into it
21:48 * sri decides he's a gome person
21:48 sri *gnome
22:02 Grinnz i only tried KDE once briefly and it was when KDE4 was new, so it's probably not a fair judgment, but it didn't seem that good
22:02 Grinnz i hear plasma 5 is much better though
22:06 sri looks like a more polished cinnamon
22:08 sri plasma was also the only desktop that actually felt slow in vmware so far
23:04 jnbek joined #mojo
23:35 sri wow, looks like couchdb just got rid of the one thing i liked about it
23:36 sri which was the cool logo
23:46 Grinnz heh

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