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

IRC log for #mojo, 2016-10-16

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

All times shown according to UTC.

Time Nick Message
00:11 zivester joined #mojo
00:37 bpmedley Any know if sjdy521 hangs out here?  I may have a fix for their issue (#1008)..
00:52 eseyman joined #mojo
01:15 aborazmeh joined #mojo
02:02 troglodita joined #mojo
02:04 noganex joined #mojo
02:42 zivester joined #mojo
03:18 asarch joined #mojo
05:30 juikuen joined #mojo
05:34 parv joined #mojo
06:59 dod joined #mojo
07:04 dod joined #mojo
07:52 dod joined #mojo
08:28 polettix joined #mojo
08:30 sri so, amazon finally has their first hit series
08:30 sri goliath is really good
08:41 zivester joined #mojo
08:47 polettix re Minion::Backend::SQLite memory usage: https://polettix.it/public/minion-backend-sqlite.png
08:47 polettix over about 16 hours
08:47 polettix well, probably more 15 hours
08:48 polettix the same code running at the same time with ::Pg is always still at the initial memory usage
08:50 polettix doh, someone noticed the same thing and already opened an issue :) https://github.com/Grinnz/Minion-Backend-SQLite/issues/8
09:21 zamanf joined #mojo
09:30 good_news_everyon joined #mojo
09:30 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vPiYF
09:30 good_news_everyon mojo/master 70b728c Sebastian Riedel: fix a few proxy bugs in Mojo::UserAgent (closes #1004)
09:30 good_news_everyon left #mojo
09:30 Janos joined #mojo
09:34 Vandal joined #mojo
10:12 polettix joined #mojo
10:22 rshadow joined #mojo
10:54 lluad joined #mojo
11:02 ivi_ joined #mojo
11:05 ivi joined #mojo
11:17 lluad joined #mojo
11:28 hrupp joined #mojo
12:19 lluad joined #mojo
12:30 sri jberger: no sticker yet? :o
12:30 sri *+s
13:06 dod joined #mojo
13:35 juikuen joined #mojo
13:56 pink_mist polettix++ # well written addition to the bugreport
14:01 tyldis I suspect Mojo::SQLite or DBD::SQLite. Even without Minion I see what appears to be leaks with sqlite
14:02 zivester joined #mojo
14:02 tyldis One of the reasons I forked off my inserts in my current project (but I'l be replacing that part of the code with StatsD instead of SQLite anyways)
14:19 pink_mist I used to have an irssi plugin interfacing with an SQLite database, and it also exhibited leakage, but there were just too many moving parts between irssi, the plugin, and SQLite to be able to narrow down which bit was causing it
14:19 pink_mist I always blamed the plugin, but maybe I should have blamed SQLite
14:47 jberger sri: not yet
14:47 jberger When were they expected?
15:09 sri actually i do
15:09 sri https://wwwapps.ups.com/WebTracking/track?track=yes&trackNums=1ZA576230297627061&utm_swu=2840
15:10 sri oops
15:21 asarch joined #mojo
15:55 Janos joined #mojo
16:29 jberger If it is in palatine then it's close
16:30 jberger In other news, someone is finally asking this question
16:30 jberger https://tierion.com/blog/who-pays-for-the-decentralized-web/
16:31 jberger I'm in a scientific channel that loves to talk about putting all their data on a decentralised system
16:31 jberger Of course they talk about data longevity but they also talk about reduced cost of hosting
16:31 jberger Ann's that just makes no sense to me
16:32 jberger Errr *and
16:33 jberger Of course you have reduced cost if lots of other people will host your giant days sets
16:33 jberger Data
16:33 jberger But who are those people? Who pays for their servers?
16:33 jberger And why are they going to let you use them for cheap/free like you think they will?
16:46 disputin joined #mojo
16:47 tyldis jberger: Costs are divided among many users? :)
16:55 bpmedley jberger : What would a href in an anchor look like for a decentralized web, or am I mis-understanding?
17:05 tyldis Oh, I should probably have read the link...
17:09 abra joined #mojo
17:44 marty_ joined #mojo
17:52 abra joined #mojo
17:58 khfeng joined #mojo
18:01 abra joined #mojo
18:05 jberger bpmedley: i don't know exactly what is being proposed
18:06 jberger I assume it is basically decentralized name resolution in a way like decentralized dns
18:06 bpmedley jberger : I see, you're just saying, in general, issues as basic as payment have been thought through and that's a little odd?
18:06 bpmedley have *not
18:07 jberger Not payment so much as ownership
18:07 jberger If you replicate all the content to multiple places
18:07 bpmedley I see, and that's a fundamental piece of any data sharing, eh?
18:08 jberger And have decentralized (read redundant) resource resolution then you've more than doubled the hardware required
18:08 jberger And doubled is just for single redundancy
18:08 jberger Who pays for all that hardware
18:08 jberger Operations
18:08 purl hmmm... Operations is like.. duhh.. these applications are slow and they suck.. buy more boxes
18:08 jberger Bandwidth
18:08 purl somebody said Bandwidth was definitely not free - TANSTAAFL
18:09 jberger purl you're really ruining my rant here ;p
18:09 purl jberger: huh?
18:09 abra joined #mojo
18:09 bpmedley Without purl we are nothing.. :-O
18:10 jberger True, true
18:11 bpmedley Odd that such basics have not been considered in any talk or proposal..
18:11 * jberger hugs purl
18:11 * purl hugs jberger back
18:11 jberger bpmedley: decentralized web is a nice talking point
18:11 jberger Who needs specifics when those are the tricky point
18:12 jberger All I'm saying is that this is just feel good crap
18:12 bpmedley If ownership nor payment have been considered, then what is there for discussion?
18:12 jberger Or worse
18:12 jberger All of the web will be hosted by the few companies with the resources to do this already
18:14 jberger bpmedley: "the feel good parts" ;)
18:14 jberger "We're gonna stick it to the man", "we are going to save the internet"
18:14 jberger It's easy to say, much harder to do
18:20 vicash eventually the decentralized hosts will not be able to support themselves and will eventually merge to create an oligarchic system
18:21 PryMar56 joined #mojo
18:22 vicash the whole thing has a basis in game theory.. eventually the payoffs will not be worth it and some users will prefer to take the higher payoff and screw others rather than take the lower payoff and be community friendly
18:23 vicash you see that with open source software. some software requires several thousand man hours of work and cannot be supported by donations or similar and needs Enterprise sales to be able to fund development. eventually some of the enterprise sales reduces feature development in the open sourced portion...
18:23 bpmedley vicash : Are you saying we'd be right back to where we started?  In some animal farm sense?  :)
18:25 vicash bpmedley: yes that's my theory. the web was sort-of decentralized from the start and still is if you run your own webservers and all that but it is not cheap to run because infrastructure costs increase as your site becomes popular
18:26 vicash bpmedley: eventually more folks want in and companies spring up to serve those users and you have the same situation as today
18:27 vicash bpmedley: looking at Bitcoin as an example. i ran a few miners until june this year but had to be part of a pool and even that became prohibitively expensive.. now  you have professional datacenters mining bitcoins or some lone rich folks but otherwise it isn't feasible anymore. the incentives have disappeared for the small miners like me
18:28 vicash bittorrent works because movies/files are ephemeral... you host a movie for maybe 3 months or so.. then the usage dies down and so do the peers.. unless the movie is popular.. same with OS torrents ..
18:29 bpmedley All interesting points..
18:30 jberger actually, vicash is right, the web is already decentralized
18:30 jberger what people want now is some kinda of automagically redundancy
18:31 jberger with trackers
18:31 jberger and I just don't see how that is viable
18:32 dave actually, I think people want free speech...until it's inconvenient ;)
18:34 dave many normal users are tired of having a web site they use a lot go away
18:45 jberger sure, but why does it go away
18:45 jberger usually because the company that runs it stops wanting to run it
18:45 vicash jberger redundancy can be achieved for say informative sites that are openly crawl-able as archive.org already does that. but the restricted stuff stuck in databases for user comments is not archivable. and maybe rightly so.
18:45 jberger decentralized web doesn't prevent companies from taking down their sites
18:46 jberger and who is going to want to host content that even the original owners don't want to pay for anymore
18:46 vicash exactly because the only incentive is monetary
18:46 jberger (except for archive.org, which reminds me that this only works for static pages anyway)
18:46 vicash and it makes sense because the users were looking for a free tool for hosting their content..
18:47 vicash but yes archiving dynamic content can be a big business if it can pay for itself. it doesn't unfortunately.
18:47 jberger pick maybe one: free, self-controlled, dynamic, hosted indefinitely
18:48 vicash there is a secret archive called google Cache too ;)
18:48 jberger yeah
18:51 vicash i think VintCerf and TBLee don't have it right.. they want companies to let go of the data which isnt easy... esp since several companies sell this data to data brokers as a source of revenue
18:52 vicash internet isnt free and electricity isnt free
18:52 jberger I can promise you that :-P
18:53 vicash one would think the air above us is free but that isn't either, you cannot transmit anything without licenses and fees and purchasing spectrum etc.
18:53 vicash except for diseases.. that's free
18:53 jberger nor can you build above it at a certain height
18:54 jberger or especially sell the airspace above you
18:58 Gedge joined #mojo
19:02 alilles joined #mojo
19:02 alilles joined #mojo
19:19 dod joined #mojo
19:32 marty_ joined #mojo
20:22 mib_faw0wr joined #mojo
20:22 mib_faw0wr Hi Guys,
20:23 mib_faw0wr I have a comment form, I;d like to allow a limited set of HTML and strip the rest, e.g. I'd allow:
20:23 mib_faw0wr tags and attributes: <a href="" title="">, <abbr title="">, <acronym title="">, <b>, <blockquote cite="">, <cite>, <code>, <del datetime="">, <em>, <i>, <q cite="">, <strike>, <strong>
20:24 mib_faw0wr maybe "<a href="" title="">, <abbr title="">, <acronym title="">, <b>, <blockquote cite="">, <cite>, <code>, <del datetime="">, <em>, <i>, <q cite="">, <strike>, <strong>"
20:24 mib_faw0wr Can I use Mojo::Dom for this ?
20:24 vicash mib_faw0wr: i wouldn't recommend this. instead use a pre-designed markdown like editor and translate the text on the server side.
20:25 mib_faw0wr can you give me an example
20:25 mib_faw0wr what do you mean "pre-designed markdown"
20:27 vicash mib_faw0wr:  for one, sanitizing HTML is not a perfect system, it can still lead to complex XSS attacks. https://en.wikipedia.org/wiki/Cross-site_scripting#Safely_validating_untrusted_HTML_input
20:28 mib_faw0wr do you mean something like Text-Markdown
20:29 vicash mib_faw0wr: well what i mean is that instead of filtering out HTML, rather support a markup language like Mardown and render it in your controller and then inject it back. maybe that isn't optimal either as that will also allow <script> tags which you may not want injected.
20:30 vicash mib_faw0wr: if you just want to build an engine that filters out HTML you can use HTML::Scrubber https://metacpan.org/pod/HTML::Scrubber
20:32 mib_faw0wr Well, I was thinking allow a few tags and scrub the rest; i.e. Russian security model :)
20:34 mib_faw0wr OK, after a quick look https://metacpan.org/pod/HTML::Scrubber seems to do what I want, thanks
20:46 lluad joined #mojo
20:56 lluad joined #mojo
20:56 lluad joined #mojo
21:08 mib_faw0wr Thanks vicash, tha seems to work nicely, I suppose I should also include a profanity santisier, any suggestions ?
21:11 vicash mib_faw0wr: i don't know. it is quite difficult to implement that. how will you remove words that are spelt like 'I D I O T' for the word idiot or id10t etc. to get around your sanitizer. it isn't a simple problem. but doable i would say if you can parse the Urban dictionary and filter out words.. https://metacpan.org/pod/WebService::UrbanDictionary .. my suggestion would be to download the data from there, and create a regex automa
21:12 vicash mib_faw0wr: another thing you can do is use Aspell or Hunspell to find the closest match to a non-offensive word and replace it with that.
21:13 vicash so *uck becomes duck.
21:13 mib_faw0wr In this case, the messages would be sent to internal staff, not republished online so the user would not get any feedback as to if their words where filtered
21:13 jberger I remember a site that tried to prevent kiddie porn by censoring all words related to children, like ... children
21:14 jberger "I took my ******* to the park today"
21:14 mib_faw0wr That's just ****** *** ***
21:15 jberger :P
21:16 vicash mib_faw0wr: it is an interesting problem but i would relegate that to a research problem. You can use Lingua::EN::Levenshtein and Urbandictionary to basically see similar forms and parse them out. but it isn't trivial. Twitter hasn't been able to do it for whatever reason
21:17 jberger well, in the end, I think masking the common ones in common form at least shows that you are trying (if you are supposed to be trying)
21:17 jberger if you actually have to succeed, yeah, maybe just give up and move to someplace without internet access :-P
21:18 vicash jberger: or wait for Trump to get elected, then you won't need to do it because free speech
21:18 jberger I don't even want to get into that ...
21:18 vicash </ sarcasm>
21:19 jberger no, I mean, I could rant for longer than my usual rants
21:19 jberger as anyone still following me on twitter knows by now (sorry about that everyone :s)
21:19 vicash mib_faw0wr: i like some modules on CPAN .. Text::Fuzzy for one is an interesting module.. it can find similar words near to a specific word. so you can filter out slangs
21:20 vicash but nothing is perfect, however, something is always better than nothing
21:22 mib_faw0wr Just read this, extreme example: http://www.bash.org/?178890
21:23 vicash mib_faw0wr: see this is where it is subjective.. i cant see any word that is offensive. "ass" means donkey
21:25 vicash mib_faw0wr: if this is an internal corporate tool, it is best to generate a regex of slangs from UrbanDictionary, update it every few weeks or so and get it approved by HR. then whenever something shows up that isnt appropriate, retroactively censor it
21:27 mib_faw0wr ok, so "Corporate" would be over stating the size of the entity :)
21:27 mib_faw0wr As to approving :) , I'll be doing that :)
21:30 mib_faw0wr I'm thinking a quick regex with top 100 swear words - if I can find that many
21:31 mib_faw0wr looks like this is the ticket http://www.noswearing.com/dictionary
21:31 jberger please please please start with the Carlin 7 words at the top, I love it when swear files do that
21:35 mib_faw0wr :)
21:35 marty joined #mojo
22:03 vicash mib_faw0wr: you can make it funny too.. replace all the F* words with love
22:05 pink_mist "you motherf****er" -> "you love and honour thy mother" ... I approve
22:06 pink_mist *thine?
22:07 suede this problem reminds of the issue the lego universe people were having, trying to censor lego genitalia
22:14 mib_faw0wr ok, that worked great, anyone looking for a job as an ****yst
22:16 jberger mib_faw0wr: awesome
22:17 jberger to be fair ****ysts deserve it :-P
22:19 vicash Google ****ytics
22:19 jberger another fine target
22:20 jberger ok ok, we are straying a little far afield, lets rein it in folks :D
23:00 zivester joined #mojo
23:01 mib_faw0wr Off to the day job (fudgepacker), Thanks Guys.
23:24 jberger I'm going to interpret that very literally and move on

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