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

IRC log for #mojo, 2016-07-12

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

All times shown according to UTC.

Time Nick Message
00:23 Grinnz damn
00:23 Grinnz i haven't gotten a new SSD in a year or so but mine are more around the 600 range
00:24 bpmedley Is that nearly double the SSD speed?
00:25 sri more than double for write speed
00:25 bpmedley Nice.  Unexpected.
00:26 sri was 600 the limit for sata?
00:26 sri these are both pcie
00:27 bpmedley Ahh, different buses?
00:28 sri ah, yes, sata-3 had a limit of 600 mb/s
00:30 Grinnz yeah, that's probably why mine are there
00:33 sri aha, this is pcie 4.0, which can go up to 2000 mb/s
00:33 Grinnz http://nickroshon.com/wp-content/uploads/2010/02/Go-Faster-WordPress-Racer.jpg
00:34 Grinnz http://pic.jpgdump.com/14347.jpg
00:34 sri oh wait, that's per lane
00:36 sri this is pcie 2.0 with 4 lanes
00:37 sri 500 mb/s each lane
00:38 sri guess pcie will work for a very very long time :)
00:38 sri considering devices can use 16 and more lanes
00:39 sri now i'm starting to get why thunderbolt 3 is so cool
00:39 sri it's basically a pcie connector
00:40 jberger when I was reading about these external gaming video card boxes they basically all beg for thunderbolt 3
00:41 sri even thunderbolt 1 had 4 pcie 2.0 lanes
00:42 sri but that's not enough bandwidth for a video card
00:43 sri thunderbolt 3 has 4 pcie 3.0 lanes, for 4000 mb/s
00:44 sri still a little limited actually
00:44 sri if state of the art video cards use 16 pcie 2.0 lanes
00:45 sri that's 8000 mb/s
00:45 Grinnz the new ones definitely make use of 3.0
00:47 jberger my 2014 MBP only has thunderbolt 2, there are homebrew mechanisms for getting an external video card working but it doesn't seem very legit
00:48 jberger and the built-in one is just good enough for the few games I play
00:48 sri actually, i've heard some success stories for thunderbolt 2 recently
00:48 jberger oh? if you have a link I'd be happy to read
00:48 sri with these https://www.amazon.de/Akitio-Thunderbolt2-PCIe-Expansion-Box/dp/B00NQ23TCU/
00:48 jberger yeah, that's basically what I've seen too
00:48 jberger they all seem to need to jerry-rig some part of it
00:49 sri you just need a video card with some special feature
00:49 sri ehich i don't remember
00:49 sri but people keep recommending evga cards
00:50 slobo joined #mojo
00:50 Grinnz my last couple have been evga, good stuff and good warranty
00:50 * sri would love to get his hands on one of those akitio boxes :)
00:51 sri maybe they get cheap now that thunderbolt 3 is all the rage
00:51 tchaves joined #mojo
01:02 sri 500 - 1000 euro for new components does seem a little expensive
01:04 jberger https://odd-one-out.serek.eu/thunderbolt-2-egpu-setup-using-akitio-thunder2/
01:07 sri neat
01:08 * sri puts it on his amazon wishlist
01:10 jberger using directly for a mac seems harder though
01:10 jberger there is a link at the end
01:10 sri wonder if a low end video card is fine too
01:11 sri like gtx 970 vs 960
01:14 jberger man it has been a long time since video card model numbers meant anything to me, guess I'm getting old
01:15 sri haha, same for me
01:15 sri it's kinda weird that now that i'm getting old games interest me more again
01:17 sri who knows, maybe i'll even build a gaming pc again :o
01:19 dvinciguerra joined #mojo
01:19 sri although, when mojolicious started i was still addicted to world of warcraft
01:23 jberger I think I might be too spoiled by laptops now, don't know if I could sit in a dark office gaming anymore
01:24 genio I rarely go into my office (unless I need the Windows desktop for some reason)
01:24 Grinnz 970 is basically enough for anything modern IMO, though the new 1060 is something to watch for and its actually going to be $50-$100 cheper
01:25 Grinnz with more vram and much better performance than the 970
01:25 Grinnz (supposedly)
01:25 sri getting a state of the art card to put it into an external pcie enclosure seems like a bit of a waste
01:26 Grinnz depends what it's for i guess :)
02:11 sri gta 5 runs surprisingly well on the macbook
02:14 sri has to be one of the best optimized games ever
02:30 zivester joined #mojo
02:45 Vitrifur joined #mojo
02:49 noganex_ joined #mojo
03:30 wingfold joined #mojo
04:47 dvinciguerra joined #mojo
04:54 stryx`_ joined #mojo
05:00 stryx` joined #mojo
05:10 stryx` joined #mojo
05:21 stryx` joined #mojo
05:34 stryx` joined #mojo
05:46 stryx` joined #mojo
05:52 stryx`_ joined #mojo
05:56 anton joined #mojo
06:00 stryx` joined #mojo
06:15 stryx` joined #mojo
06:18 waghso joined #mojo
06:22 waghso From url www.host.com?user=harry how to get 'user' variable?
06:22 waghso M new to mojolicious
06:23 Adura How far did you get in the docs?
06:23 Grinnz https://metacpan.org/pod/distribution/Mojolicious/lib/Mojolicious/Guides/Tutorial.pod#GET-POST-parameters
06:26 waghso I tried $c->req->query_params->param('user');
06:26 Grinnz that would work, but just $c->param('user') is enough
06:26 Grinnz what happened when you tried?
06:27 waghso It dosen't get value user in stash
06:28 Grinnz that's a GET parameter, unrelated to the stash
06:30 waghso I pushed that GET parameter to stashusing $c->stash(user => $user);
06:30 Grinnz ok, then what?
06:31 waghso Then I can't access $user in .ep
06:34 stryx` joined #mojo
06:40 dod joined #mojo
06:51 punter joined #mojo
06:54 stryx` joined #mojo
06:59 dod joined #mojo
07:00 dod joined #mojo
07:03 stryx` joined #mojo
07:04 AndrewIsh joined #mojo
07:24 trone joined #mojo
07:27 stryx` joined #mojo
07:28 Vandal joined #mojo
07:33 stryx` joined #mojo
07:40 jberger waghso: could you show us a minimal example of what you are working on? Use a paste site
07:40 bpmedley jberger: Did waghso leave?
07:53 nothing joined #mojo
07:54 stryx` joined #mojo
07:54 jberger Oh maybe
07:54 jberger Didn't check, should have
07:55 bpmedley Heh, no worries.. a fly by irc encounter.. ;)
08:15 punter joined #mojo
08:19 ningu joined #mojo
08:21 ashimema can someone sanity check my understanding of a piece of code for me: https://github.com/jhthorsen/json-validator/blob/master/lib/JSON/Validator.pm#L706  The 'sprintf' on that line will never act on anything right? or am I misunderstanding somthing
08:21 ashimema oh.. sorry.. that was meant got the #perl channel
08:22 ningu how is the body parsed in Mojo? is it available from the beginning of when a route handler is called?
08:23 ningu it seems not to me in the thing I am testing. just wondering what I am missing.
08:23 ribasushi ashimema: it seems like a thinko, yes: just returning '()' will DTRT
08:24 ningu not to be*
08:24 ribasushi ashimema: but really I think batman meant to do (@$e == 1) ? ... : ...
08:24 ashimema i reckon so yeah
08:25 ashimema thanks ribasushi, I'm not going mad then ;)
08:25 ningu hmm... maybe because it's multipart/form-data
08:25 * ashimema goes to submit a pull request with that.
08:25 ribasushi ashimema: light up the bat sign
08:25 ashimema hehe.. think he's on vacation ;)
08:27 ningu oh I see, it ends up in body_params
08:27 ningu what if I just want the raw body?
08:28 ningu ah looks like ->content
09:17 bpmedley ningu: Can you just use $c->req->body ?
09:25 ningu bpmedley: not for multipart
09:35 cuechan joined #mojo
09:36 kaare joined #mojo
10:39 jamesaxl joined #mojo
10:40 jamesaxl hi
10:47 jamesaxl_ joined #mojo
10:48 rshadow joined #mojo
10:48 rshadow hi
10:48 rshadow I have one question
10:49 rshadow I want to make custom render format
10:49 rshadow $self->render(xmlazy => \%hash );
10:50 rshadow it`s something like $self->render(json => \%hash );
10:50 rshadow can I do this?
10:52 jamesaxl__ joined #mojo
10:57 anparker rshadow: this looks like what you want http://mojolicious.org/perldoc/Mojolicious/Guides/Rendering#Adding-a-handler-to-generate-binary-data
10:59 tchaves joined #mojo
10:59 anparker except you probably don't need to delete encoding
11:00 jamesaxl_ joined #mojo
11:02 kid51 joined #mojo
11:09 jamesaxl__ joined #mojo
11:17 absolut_todd joined #mojo
11:19 punter joined #mojo
11:26 inokenty-w joined #mojo
11:33 rshadow anparker: thanks!
12:35 che-quest joined #mojo
12:35 hummeleBop joined #mojo
12:37 che-quest greets! have a question on controller has attribute
12:37 che-quest has foo => sub {state $foo = 156};
12:39 che-quest does $c->foo always static value?
12:40 sri no, and the "state $foo = " does nothing
12:41 hummeleBop joined #mojo
12:43 hummeleBop joined #mojo
12:44 ashimema_ joined #mojo
12:46 che-quest but: sub foo {state $foo ... does?
12:46 sri yes
12:46 che-quest ok, big thanks
12:47 dvinciguerra joined #mojo
12:48 nic che-quest: the sub you define for 'has' only gets called if something reads the attribute and it has never had a value (or has been deleted)
12:48 nic so you can see it isn't useful to make the value within the sub persist if the sub is only invoked once
12:52 che-quest i need static(constant) attr per each objects of the module in process
12:53 che-quest first object init attr nexts - get attr
12:54 zivester joined #mojo
12:55 che-quest earlier i use class vars
13:03 punter joined #mojo
13:06 kes joined #mojo
13:07 kes Hi. When you are planning to make release v7 onto cpan?
13:07 sri when i get around to writing the announcement :)
13:08 genio Well, stop toying with your new laptop and get to work!  *cracks whip*
13:08 kes ok. Will be waiting...
13:09 sri due to my laptop troubles i do have a work backlog that needs to be addressed first
13:09 sri might take some time
13:10 kes sri: take your "round TUIT" ;-) http://blog.liruoko.ru/images/tuit/tuit_002_small.jpg
13:11 sri still have one from mojoconf
13:17 sri must be defective though, because my open source time keeps shrinking since i got it -.-
13:20 gizmomathboy joined #mojo
13:20 cuechan joined #mojo
13:25 PryMar56 joined #mojo
13:26 anton hi. i would like to ask you what do you think about "async DBI"? i mean you can write "normal" DBI queries and they don't block your event loop
13:27 sri http://mojolicious.org/perldoc/Mojo/Pg
13:27 anton and?
13:27 anton i already looked
13:27 anton at this module
13:27 anton you can either block eventloop or use callback
13:28 anton i wouldn't call it "notrmal"
13:28 nic Is there a question?
13:28 anton i mean you can do it with coro thou with some care
13:28 sri sorry for trying to help
13:29 anton the question is you are interested or not?
13:29 anton sri: thanx.
13:30 anton uh. i was told that you can be interested in such a module
13:31 genio anton: We're a bit confused by the order of your questioning.  Can you try again to explain what it is that you're after?
13:32 genio Are you looking for how to write asyncronous DB requests? If so, Mojo::Pg is a good place to look.   Are you looking to try to make another module that works with Coro with the help of people in here?
13:32 anton i don't use mojo, but i am interested in some testing of my modules
13:32 nic I'm going to assume you have no question
13:32 anton i know how to write that thing. i just don't like "callback hell"
13:33 sri there is no callback hell in mojolicious
13:33 sri we already have control flow libraries
13:35 anton ok
13:35 anton as you wish
13:38 genio anton: What is your primary language?  Maybe we're having a language barrier problem.  Can you writeup what you mean to propose/accomplish in your first language and paste it to a paste site somewhere? Then maybe someone who also speaks that language can help us understand your goals. I only make this suggestion because so far your language usage seems terse and a bit confusing.
13:39 anton my primary language is perl
13:39 genio spoken language
13:39 anton i understand
13:40 anton i am just joking
13:43 anton $db->query -- Execute a blocking SQL statement ... You can also append a callback to perform operations non-blocking. -- am i missing something?
13:44 anton is there a way to perform nonblocking without callback?
13:49 nicomen sure, you just can't do anything with its results
13:49 nicomen or you use a job-system, with something else polling
13:50 anton ok, i'll try to explain. i have an ajax handler that have to make _many_ db calls to collect all needed data. the number of queries can vary.
13:50 anton this one handler when used in blocking mode can block entire process for some time
13:50 bobkare anton: have you looked at Mojo::IOLoop::Delay?
13:51 marty joined #mojo
13:52 anton no. that's interesing
13:52 marty joined #mojo
13:54 anton if there were some example with db queries that were executed in some particular order -- that would be excelent
13:56 nic all my db queries are executed randomly
13:56 nic written randomly too sometimes
13:59 anton i mean sometimes you can execute independent queries in no particular order but some queries can depend on results from another
14:00 nic Sounds like ::Delay is what you want
14:02 newbie joined #mojo
14:02 anton do you mean $delay->steps(sub {...}, sub {...}); ?
14:03 anton sequentional callback execution?
14:04 nic there are few different ways of writing it -- what's best depends on the context
14:04 nic not just sequential
14:05 anton anyway, as i see it, it's a kind of "callback linearization", am i right?
14:05 nic yes and no
14:07 anton it's interesting. probably the best variant without coro
14:09 sri around here we tend to like reliable code, so there's very few coro users
14:10 anparker is coro still alive?
14:10 Grinnz depends who you ask
14:11 sri who knows, maybe some day there will be a Cpanel::Coro and it comes back
14:12 * sri wrote this in 2013 https://gist.github.com/kraih/6082061
14:18 anton is there a mojo::pg:pool?
14:20 anton i mean do you queue queries to the only one db connection?
14:21 ningu joined #mojo
14:38 genio blocking vs non-blocking:  http://imgur.com/gallery/ZeSTCsS   :)
14:42 mcsnolte joined #mojo
14:54 zivester joined #mojo
14:59 disputin joined #mojo
15:07 batman joined #mojo
15:21 jnbek joined #mojo
15:22 noize joined #mojo
15:29 Akron joined #mojo
15:30 sri a wild Akron appears
15:30 sri o/
15:33 Akron Yeah!! o/
15:34 Akron Just saw that heise is using Mojolicious - didn't know. http://www.heise.de/newsticker/meldung/Die-Technik-hinter-heise-online-3262514.html?wt_mc=rss.ho.beitrag.atom
15:35 sri oh, cool
15:37 lluad joined #mojo
16:02 disputin joined #mojo
16:20 ningu joined #mojo
16:25 ptolemarch joined #mojo
16:34 dod joined #mojo
16:50 neilhwatson joined #mojo
16:57 trone joined #mojo
17:05 _dave_ curious...anyone here use webpack with morbo?
17:17 dod joined #mojo
17:20 anton is coro a bad taste here?
17:23 ribasushi sigh... ok let me do a bit of a translation, as the cultural/language clash won't go away on its own
17:24 asarch joined #mojo
17:26 ribasushi what anton is working on is a set of DBD subclasses (like direct DBD::Pg / ::mysql / etc subclasses ) to be used for "invisible async" with a connection pool etc. The idea (supposedly) is to get something like DBIC to become async-aware without extra work on the lib's side. His current work is very coro-centric (as this is his async-paltform of choice), but he is interested in validating his approach within a wider audience. Given many folks here would
17:26 ribasushi like to see an "async DBIC" (whatever this may mean) I sent anton here to get feedback. Although I warned him ahead of time that he *must* provide a non-coro-specific wrapper for his stuff before the current crowd will even look at anything
17:26 ribasushi this about sums it up
17:26 ribasushi ashimema_: ^^
17:26 ribasushi jberger: ^^
17:26 ribasushi mcsnolte: ^^
17:27 ribasushi I have not looked into any of his code as my interest in async in general is low, so i can't comment on the actual idea/implementation
17:28 anton ribasushi: thanx
17:28 mcsnolte nice, i look forward to seeing what you come up with anton
17:31 anton currently i can show only some test code as i just finished initial testing
17:32 anton https://github.com/jef-sure/dbd-pg-cae -- main module to get async pg in DBI
17:33 anton https://github.com/jef-sure/DBIx-Connector-Pool -- pool for async connectors
17:34 sri coro has become very popular in the perl community in general after the recent drama
17:34 anton you can see in t/* how this all works
17:34 anton sri: debian has it working
17:34 sri umm
17:35 sri *unpopular
17:35 Grinnz_ I was gonna say... lol
17:35 anton https://packages.debian.org/sid/perl/libcoro-perl
17:36 sri the only way i will ever use coro again would be if there was an independent alternative
17:36 anton i would love to have one
17:36 sri like how we use EV with a fallback if mlehmann goes rogue
17:37 anton and btw, DBIx::Connector::Pool is "pure" in terms of dependens on thread model
17:37 sri as long as there is only coro i'm not interested at all i'm afraid
17:37 Grinnz_ for context, p5p actually came up with a way to get Coro working on perl 5.22, and this was the author's response. http://www.nntp.perl.org/group/perl.perl5.porters/2016/06/msg237331.html
17:38 anton np. i am not interested in mojo also
17:38 sri and i'm no longer interested in talking to you anton
17:38 sri seriously, i'm trying to be nice and you keep making snippy remarks
17:40 sri i know after the last two remarks i should have just gome away...
17:41 ningu I am interested in mojo because I just used it to bring an ancient, huge cgi script into the modern world (without using the cgi wrapper)
17:42 sri anton: anyway, good luck with that attitude, you'll need it
17:43 anton thanx.
17:46 ribasushi <headdesk> sorry for the noise everyone I guess
18:31 ningu would there be a way to auto-reload my hypnotoad-deployed app, whenever a change is made to a *.pm file in its directory? I know that's kind of a strange thing to do in production but I am working with someone who is used to modifying the cgi script and having the change take effect immediately...
18:32 ningu I guess I'd just need a watcher for the directory
18:32 ningu I could hack something up
18:49 inokenty joined #mojo
19:03 marty joined #mojo
19:19 meshl joined #mojo
19:20 nic ningu: From time to time I think about integrating one of the inotify modules to do what you're describing
19:20 nic but in reality I just run morbo on a separate port in parallel
19:21 nic this is actually much better and safer
19:28 _dave_ and easier :)
19:29 dod joined #mojo
19:33 nic indeed
19:34 nic Once you (or your friend) are happy with the changes visible via morbo, you ask hypnotoad to do a hot reload.  Simplz
19:38 Grinnz_ having changes take effect immediately in production is a terrible idea. just hot-restart hypnotoad when you need to
19:44 ningu Grinnz_: can you hot-restart hypnotoad in -f mode?
19:44 ningu I tried and it didn't seem to work. said couldn't create socket, it was in use
19:44 Grinnz_ don't think so
19:44 ningu yeah, thing is I want to be sure hypnotoad runs on startup and restarts if it dies, so I am using -f
19:45 nic c'mon, be serious.  If you're editing on Live, then hot reloading the server is not your biggest problem
19:45 ningu right
19:45 nic If you were in a sensible enterprise situation you wouldn't be doing that
19:45 ningu we're not going to get to the point of having a development server anyway
19:45 ningu and we are not an enterprise
19:46 nic that was between the lines :)
19:46 Grinnz_ dev servers aren't just for enterprise :P VMs are easy
19:47 Grinnz_ also you can just run morbo and hypnotoad on the same box from different checkouts
19:47 nic and "hot-restart hypnotoad when you need to" has its own risks
19:48 nic remember that workers get restarted automatically too
19:48 _dave_ ugh VMs ... great for dev but then someone always wants to put that VM in prod
19:49 _dave_ in my experience, VMs are not cost effective at medium to high traffic loads
19:49 nic heh, I have come to the opposite point of view
19:50 Grinnz_ yeah, production deployment is a different matter
19:50 nic VMs are wonderful for production and terrible for dev
19:50 Grinnz_ nic: not sure what risk that adds
19:50 Grinnz_ beyond the risk you've already accepted of editing production without first testing on a dev server
19:50 _dave_ it's probably my personal experience speaking here, since I typically have to fix a VM that I never provisioned in the first place
19:51 _dave_ it's left a sour taste in my mouth for anything virtualized
19:51 Grinnz_ luckily, dealing with the VMs other than my own personal use ones is not in my job description :P
19:51 _dave_ "yeah we put the database on the same VM as the web server because it was cheaper" :/
19:51 Grinnz_ ahah
19:52 _dave_ I could so do stand up comedy for geeks. Wouldn't need any material.
19:53 ningu Grinnz_: we barely have the competence to use git
19:53 ningu different checkouts and push/pull for each change in development would be too much
19:54 Grinnz_ heh
19:54 ningu also, there is little "risk" in editing "production" live when there are less than 10 users and most are in the same room so when there is an issue someone can just bring it up
19:54 nic ningu: Once you 'get into it' git starts working for you and it saves _so_ much effort
19:54 Grinnz_ balancing workflow around "what I can expect the rest of the team to want to do" is another matter
19:54 ningu nic: I am not talking about myself
19:54 nic k
19:55 ningu yes, talking about others on the team
19:55 ningu who find git incomprehensible and an annoyance
19:55 ningu I still make them use ie
19:55 ningu it*
19:55 ningu but they don't know anything about branching, merges, etc.
19:55 _dave_ that sounds risky
19:56 nic don't make them do merges.  that's where beginners die
19:56 ningu true
19:56 ningu but having multiple deployments would be similarly complicated for them, I think
19:56 nic dictate a very simple branching regime
19:56 ningu yes, I have done that, with detailed instructions for how to make a new branch
19:56 ningu but that happens fairly rarely
19:58 ningu until 2 days ago this was a CGI script with no .pm files... so...
20:00 Grinnz_ the rest of my team is starting to get the hang of git branches, I think
20:00 Grinnz_ it's very helpful
20:00 ningu and to think, humans are descended from arboreal brachiators so branching should be in our nature...
20:01 * nic goes to find some leaves
20:02 Grinnz_ arboreal brachiator is my acid rock band name
20:05 _dave_ lol...mp3s or it didn't happen :)
20:05 ningu I am a little confused by the docs at https://metacpan.org/pod/Mojolicious::Controller#session
20:06 ningu difference between expiration and expires
20:06 ningu if I understand right, expiration makes the session last for an hour, no matter how many further requests there are
20:07 ningu then it gets deleted/reset
20:07 ningu whereas expires sets the absolute expiration time, on each request
20:07 Grinnz_ I believe 'expiration' sets the expiration time on every request
20:07 Grinnz_ automatically resetting
20:07 ningu yeah
20:07 ningu that's how it sounds
20:07 Grinnz_ 'expires' sets it directly for this one request
20:08 ningu ah, ok
20:08 ningu are there standard plugins for server-side sessions?
20:08 Grinnz_ not in core
20:09 Grinnz_ generally, "store something in whatever database you're using and store the ID in the session cookie"
20:09 ningu yeah, makes sense
20:09 ningu probably overkill in my case anyway
20:09 ningu what's the message about "your secret needs to be updated"... how do you do that?
20:09 Grinnz_ https://metacpan.org/pod/Mojolicious#secrets
20:10 ningu ah, thanks
20:11 Grinnz_ https://metacpan.org/pod/Mojolicious::Sessions#default_expiration # a bit more info on expiration
20:57 sri i'll never get used to the choppy scrolling in windows
21:11 meshl joined #mojo
21:32 cuechan_ joined #mojo
21:37 trone joined #mojo
21:49 * marty is binge watching Person of Interest
21:49 marty Sure did start out slow.  Season 2 is finally starting to get beyond "bad guy of the week".  Hope it keeps getting better.
21:50 sri it does
21:50 marty \o/
21:54 sri season 3 will shake things up completely and start the real story :)
21:57 sri the main storyline will take over more and more of the show, until it's a big ai war
21:58 sri not gonna say more though ;p
22:15 genio and then the evil queen kills them all!
22:16 sri /kickban genio no spoilers!
22:16 genio :)
22:20 disputin joined #mojo
22:23 meshl joined #mojo
22:25 meshl joined #mojo
22:27 ningu why is $c->session documented as being "for the next few requests"? presumably it's valid until its expiration, nothing more, nothing less?
22:38 marty Thanks sri.  I heard that POI was good but I was starting to wonder.  Actually, even season 2 is starting to pick up a bit.
23:12 dvinciguerra joined #mojo
23:25 ningu poi in Hawaii is good
23:49 ningu is there a simple way to clear a whole session?
23:51 ningu I guess $c->stash->{'mojo.session'} = {};
23:56 sri mojo.* values are private
23:56 ningu ok
23:56 ningu so is there a way to clear it safely?
23:56 ningu I could delete all of the keys, I guess
23:58 Vitrifur joined #mojo

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