Humanitarian OSM Team/Working groups/Technical/meeting 2014-07-07

From OpenStreetMap Wiki
Jump to: navigation, search

Meeting to discuss Tech topics on Monday 7th July 2014

IRC log:

19:02:27< dodobas> hello hello
19:02:56< pierregiraud> hello
19:03:02< dodobas> are we ready to start with the TWG meeting... we have a tentative agenda
19:03:14< harry-wood> hello
19:03:20< dodobas> is everyone ok with that, do you want to add something?
* clara can just say hi, and then is off 19:03
19:03:58< pierregiraud> agenda OK for me
19:04:17< harry-wood> hi clara
19:04:49< dodobas> pierregiraud: do you want to start... I've migrated the tmv1->tmv2
19:05:04< pierregiraud> is the code also up to date ?
19:05:14< dodobas> yes
19:05:19< pierregiraud> weird
19:05:44-!- larryone [] has joined #hot
19:05:53< pierregiraud> ah OK seems fine now
19:05:59< pierregiraud> cache issues
19:06:03< pierregiraud> OK, so
19:06:33< pierregiraud> I decided to write a script to migrate data from V1 to V2
19:07:03< AndrewBuck> Hello all.
19:07:16-!- alexbarth [] has quit [Remote host closed the connection]
19:07:48< pierregiraud> V2 is still in progress but has enough features to be put in production
19:08:09< pierregiraud> features from v1 are almost all there
19:08:26< pierregiraud> I asked dodobas to run the script on the dev instance
19:08:32< pierregiraud>
19:09:09< sev_hotosm> hello
19:09:09< harry-wood> Did it create a lot of "[HTML_REMOVED]" ?
19:09:10< pierregiraud> it looks like it run with no issues and all the jobs of the V1 were correctly migrated in this instance
19:09:10-!- alexbarth [] has joined #hot
19:09:28< harry-wood> hi sev_hotosm
19:09:29< pierregiraud> HTML_REMOVED means that we don’t allow HTML tags anymore
19:09:29< dodobas> pierregiraud: yes, a bit slow... but it finished
19:09:44< sev_hotosm> hi harry-wood
19:09:52< pierregiraud> project descriptions need to be cleaned up a bit
19:10:08< harry-wood> ah right. no HTML but we do allow markdown?
19:10:15< pierregiraud> yes we do
19:10:37< sev_hotosm> (I know I ask this question again, sorry) Do you have an idea when the TMv2 is launched?
19:10:42< pierregiraud> if the HTML_REMOVED statement is too agressive, we can easily remove it
19:11:00< pierregiraud> sev_hotosm: it’s now a question of days
19:11:03< AndrewBuck> it might be good to have the script convert a couple of the more common tags or just allow the simple ones like <b>, <em>, and <i>
19:11:07< sev_hotosm> tadam!
19:11:09< pierregiraud> I hope before the end of the week
19:11:44< dodobas> pierregiraud: which week ? :)
19:11:45< pierregiraud> AndrewBuck: you can still achieve the same goal with markdown
19:11:51< pierregiraud> the current one
19:12:04< pierregiraud> don’t you think it’s possible?
19:12:04< dodobas> :)
19:12:27< dodobas> sure it is, I'm just asking to be sure...
19:12:44< pierregiraud> of course the change will be a bit frightening at first for users
19:13:00< dodobas> pierregiraud: what about old links...
19:13:30< dodobas> we MUST support /job/:id
19:13:45< pierregiraud> if feasible, it would be nice to redirect job/:id to host/project/:id
19:14:00< dodobas> becuse all of the links in the wikis, emails, documents will stop working
19:14:10-!- jaakkoh_ [] has quit [Ping timeout: 480 seconds]
19:14:13< pierregiraud> can you do it with a nginx conf?
19:14:45< dodobas> pierregiraud: should be doable
19:14:59< pierregiraud> are there any other kind of links that we should support?
19:15:00< dodobas> i'll test it after the meeting...
19:15:02< pierregiraud> I don’t think so
19:15:40< AndrewBuck> pierregiraud: I know it can be done in markdown, my point was if the tags are not allowed to have the script do a search/replace on the job instructions to fix the easy to do tags, i.e. put in the markdown code in place of the old html tag
19:15:41< dodobas> i guess, /job/:id was the only entry point
19:15:42< pierregiraud> if not feasible, it’s also possible to add a route at the application level
19:16:18< sev_hotosm> @pierregiraud markdown is fine, html was a bit cumbersome anyway
19:16:20-!- neiljp_ [] has quit [Remote host closed the connection]
19:16:22< AndrewBuck> Unless you mean those tags are still allowed in markdown, in which case no change needs to be made.
19:18:09< pierregiraud> I have no clue about how much effort it would require to convert some of the tags to something valid in markdown
19:18:22< dodobas> AndrewBuck: it depends on the markdon parser... you can't be as expressive as with HTML... but it's much safer
19:18:52< pierregiraud> however, as proposed earlier, we can also remove the « HTML_REMOVED » statements
19:19:05< pierregiraud> to get something less agressive
19:19:09< harry-wood> Could leave the HTML unaltered
19:19:13< AndrewBuck> yeah, I am familiar with markdown, I just don't want to have to manually fix the ~575 job instructions.
19:19:36< harry-wood> unaltered HTML would look messy, but would allow people to manually cleanup without losing information ( link URLs)
19:19:37< pierregiraud> AndrewBuck: I don’t think we have to fix the 575 jobs
19:19:43< pierregiraud> only the latests
19:20:03< pierregiraud> HTML is kept as is in the databse
19:20:09< ptressel> Is there a danger in allowing those particular tags?
19:20:13< AndrewBuck> ok, as long as you have a plan then.
19:20:37-!- neiljp [] has joined #hot
19:21:01< pierregiraud> ptressel: messy DOM code with unclosed tags
19:21:08< dodobas> AndrewBuck: it's hard to convert HTML -> markdown, it depends on the context...
19:21:39< pierregiraud> but there are security issues as well
19:21:56< pierregiraud> though not a big threat
19:22:15< dodobas> there is ... panadoc (which I used on one of the projects), written in Haskell ->
19:22:24< AndrewBuck> well, I will leave it up to you guys, like I said, if you know how to make it work that is fine.
19:22:30< dodobas> so, pierregiraud, you could try that one
19:22:30< ptressel> If the links were dangerous, they'd have been dangerous in the original, no?
19:22:44< ptressel> These are not new tags, just migrated ones.
19:23:22< AndrewBuck> the dager is not links per say, it is things like the <script> tag which can be used to execute attacks and such.
19:23:40< pierregiraud> AndrewBuck: you get the point
19:23:45< ptressel> These are migrated descriptions, no?
19:23:59< ptressel> Do any contain <script> added by humans?
19:24:43< sev_hotosm> I can clean my jobs, just tell me where to do it
19:24:48< AndrewBuck> no, but the point is if you allow html tags nothing stops someone from adding one in the future.
19:24:54< dodobas> ptressel: the problem is... humans forget to close tags... and unclosed tags brake the site...
19:24:58< ptressel> Aren't we talking about /job/:id ?
19:25:23< pierregiraud> ptressel: nope, sorry
19:25:25< dodobas> ptressel: no, that will just be a redirect
19:25:37< pierregiraud> we’re talking about HTML no supported anymore in descriptions
19:25:45< AndrewBuck> dissallowing html in favor of markdown definitely makes sense, it just needs to be migrated over which it sounds like they will do.
19:25:59< ptressel> Sounds like talking about two different thing. User-added tags going forward, and /job/:id in the old descriptions.
19:26:24< pierregiraud> AndrewBuck: I’ll do what I can to automatically do the migration of HTML tags to markdown
19:26:40< pierregiraud> or at least provide adviced on how to do it easily
19:27:18< pierregiraud> manual migration will still be required though
19:27:30< pierregiraud> for example when description are mixing different languages
19:27:38< AndrewBuck> ok, yeah. Fixing a few odd ones manually is not a big deal, my idea was just to have the script do the "low hanging fruit" which should take care of most of the stuff. The only other thing commonly in use is the <div> tag which is used mostly on the more recent jobs.
19:27:47< dodobas> pierregiraud: try panadoc... it 'migth' just work
19:27:51< pierregiraud> the project managers will have to edit the projects
19:28:01< pierregiraud> dodobas: thanks for the link
19:29:25< pierregiraud> any suggestion for the HTML_REMOVED thing ?
19:29:31< pierregiraud> for new projects?
19:29:55< pierregiraud> do we want to inform the project manager that he/she cannot use HTML tags?
19:30:05< pierregiraud> or do we remove the tags silently?
19:30:16< mkl> just remove tags silently on new jobs
19:30:18< pierregiraud> both options are just config options
19:30:29< harry-wood> it will rapidly become obvious :-)
19:30:37< pierregiraud> I agree
19:30:38< mkl> as long as its explained in the help text that only markdown is allowed, we're fine
19:31:38< pierregiraud> mkl: OK
19:31:47< pierregiraud> I’ll add some help text
19:32:12< pierregiraud> are people here satisfied with the migration except for the HTML thing?
19:32:19< mkl> sorry if i missed ... how is it going with "phase 2" features?
19:32:26-!- neiljp [] has quit [Remote host closed the connection]
19:32:38< pierregiraud> mkl: I didn’t talk about the phase 2 features
19:32:45< pierregiraud> only about V1 migration
19:33:00< pierregiraud> in order to get V2 up and running as soon as possible
19:33:13< pierregiraud> replacing the current V1 forever
19:33:41< ptressel> If you use a markdown editor, users may be less likely to try entering tags. Here's the one used by stackoverflow:
19:33:48< mkl> good good. definitely higher priority. just curious about the outlook on those, when appropriate
19:33:58< ptressel> They have a markdown sanitizer that removes <script>.
19:34:38< pierregiraud> ptressel: I’ll have a look
19:34:40-!- neiljp [] has joined #hot
19:35:04< pierregiraud> ptressel: there’s already a preview in the current implementation
19:36:03< ptressel> I'm thinking more of wysiwyg controls, so the stackoverflow editor might not offer enough encouragement not to enter html.
19:36:03-!- neiljp [] has quit [Remote host closed the connection]
19:36:11< pierregiraud> so, if people agree, I’ll give the descriptions migration from HTML to markdown a try
19:36:52< pierregiraud> if I get something working I think we’ll be ready to shut V1 down
19:37:09< mkl> ptressel: we used this wysiwyg markdown editor on a recent project, liked it
19:37:14< pierregiraud> are there any objections?
19:37:47< pierregiraud> mkl: nice too
19:37:52< harry-wood> We've got (v2) and (v1) and will always be the actually live instance to use correct?
19:37:59< AndrewBuck> no objection from me.
19:38:08< mkl> pierregiraud: i would only suggest, not letting html->markdown conversion slow down production switch
19:38:14-!- neiljp [] has joined #hot
19:38:28< dodobas> harry-wood: yes
19:38:39< pierregiraud> mkl: I’m not sure I understand what you mean
19:38:55< harry-wood> he means… no objection :-)
19:39:04< mkl> i just mean, no objection. just suggesting, don't let it take too much time
19:39:15< pierregiraud> even if I don’t have something to propose for automatic conversion?
19:39:45< mkl> sounds like job admins were willing to go through and edit
19:40:35< harry-wood> we do have 578 jobs though. Automatic conversions are good
19:40:56< pierregiraud> +1
19:41:21< AndrewBuck> one alternative is to make a backup column of the original html text, then just do the conversion later.
19:41:38-!- neiljp [] has quit [Remote host closed the connection]
19:41:43< pierregiraud> AndrewBuck: the original text is still there
19:41:51-!- neiljp [] has joined #hot
19:41:53< AndrewBuck> ok, then that is no problem.
19:42:19< dodobas> pierregiraud: anything else ? :)
19:42:34< pierregiraud> nope thanks you
19:42:58< dodobas> so... next on the agenda...TWG wiki update...
19:43:28< dodobas> I've been updating... the wiki, a bit ...
19:43:54< dodobas> trying to make it a bit more readable...
19:44:22< dodobas> one of the issues, we should think about how to name stuff...
19:44:57< dodobas> do we have projects, services, resources, external services, or ?
19:45:35< mkl> what are active projects vs active services?
19:46:52< dodobas> well, service is something that's hosted, reachable
19:47:12< dodobas> a project is something that we can't host ?
19:47:27< dodobas> i guess, i'm not to happy with those
19:47:47< dodobas> basically, we need a taxonomy (?)
19:48:25< mkl> maybe just make those projects a child section of HOT Services Overview
19:48:29< mkl> called "Other Projects"
19:48:36< mkl> or "Additional Projects"
19:48:42< mkl> with a note that they aren't hosted services
19:48:50< ptressel> Additional is good.
19:49:12< mkl> The whole section could be HOT Projects Overview
19:49:24< dodobas> well this is something that we sohuld disscuss, maybe with the CommunicationTWG
19:49:29< dodobas> *WG :)
19:50:00< harry-wood> most of the things are "services". Things involving code, but also involving hosting of a production instance… but there's one or two which aren't such as HDM-JOSM-style
19:50:13< mkl> sure. otherwise than this point, i think it's pretty clear.
19:51:03< dodobas> ok, you can update the wiki, then... :)
19:51:33< mkl> oh boy
19:52:07< dodobas> so.. all of the HOT services are running normally
19:52:12-!- cypher_ [] has joined #hot
19:52:18< harry-wood> Hurray!
19:52:47< dodobas> i've been updating HIRC service, but.. it's not live yet... probably sometime during this week
19:53:16< dodobas> i've also added Myanmar region to the HOT export service...
19:54:05< harry-wood> So what is this "Project manager template report" you have on the agenda?
19:54:16< dodobas> and started preparing Project manager report for the exports service
19:54:34< dodobas> well, harry-wood , we introduced project managers for HOT projects...
19:54:41< mkl> (wiki edited...)
19:54:51< ptressel> \o/
19:55:03< dodobas> in short, a person responsible for a project (communication, managing repo, ...)
19:55:34< dodobas> so for example... i've pruned hot-export issues
19:55:43-!- larryone [] has quit [Ping timeout: 480 seconds]
19:55:51< dodobas>
19:56:17< dodobas> so.. there are three types of issues
19:56:29< harry-wood> Oh I see. Project Managers of tech projects. And so "Project manager template report" … a template of the report you would give to the group?
19:56:45< dodobas> harry-wood: yes, still working on it...
19:57:04< dodobas> bug: simple, doesn't require a lot of work
19:57:32< dodobas> enhancement: a bit more work, it probably should not change models/apis
19:58:03< dodobas> feature request: a lot of work, might break other stuff, changes to everything
19:58:44< dodobas> you can also add, a question tag, if an issue requires discussion
19:59:03< dodobas> for example:
19:59:12< dodobas> not sure what to do with that one
19:59:51< dodobas> for others, we could create an initial 'task list'
19:59:54< dodobas>
20:00:28< dodobas> to steer wandering developers ... :)
20:00:32< harry-wood> Yeah I saw you'd done that
20:00:46< harry-wood> Not seen those before. Do all the tasks get listed somewhere?
20:00:59< dodobas> nah, it's per issue
20:01:06< harry-wood> ah ok. only relate to that issue
20:01:24< dodobas> so organizing/pruning issues is one responsibility
20:02:13< dodobas> the other one... is preparing a report, or something like 'how-to-contribute' document... which issues are simple, what is urgent, and so on
20:02:37< dodobas> the problem... i don't know how to communicate it
20:03:20< dodobas> for example, we could use github wiki... but some of the projects might already use the wiki
20:03:34< dodobas> also the wiki is not something a user goes to...
20:03:48< ptressel> Don't know what github lets you control, but it trac, one can color the rows in a report, e.g. by severity or priority.
20:03:54< ptressel> Or sort by priority.
20:04:06< ptressel> s/it/in
20:04:37< dodobas> the second idea was to create a file in the repo... or and keep the information in the repo
20:05:02< ptressel> On another project, we ask people entering issues to mark them as "easy", so we something to point newcomers too.
20:05:06< dodobas> ptressel: well, you could organize issues a bit, but im not sure if you can reorder the issues
20:05:06< pierregiraud> projects often use this option
20:05:18< harry-wood> it's a standard thing to have a file per repo. but you're talking about across all projects hey?
20:05:59< dodobas> harry-wood: just somting that we could use a standard...
20:06:33< dodobas> also... if it's in the repo...then the information will never get lost... even if me migrate to another hosting option
20:06:47< harry-wood> has 'for developers' section which we can edit/improve
20:07:01< mkl> +
20:07:02< mkl> +1
20:07:35< dodobas> harry-wood: also, you could then 'gather' everything from with a simple script... and create a 'projects report'
20:07:36< ptressel> Do you want a way to emphasize the issues that are urgent visibly in the list, to get people to take those first?
20:07:58< dodobas> ptressel: we could abuse milestones for that purpose
20:08:17< ptressel> Also tags, by appropriate choice of colors.
20:08:39< harry-wood> One thing I've been wanting to try is to bring in some dynamic content on to via scraping e.g. content from the wiki page or from a file could be auto-scraped onto the 'for developers' page on the site.
20:08:42< dodobas> ptressel: yeah, but we need to be clear across all projects...
20:09:14< AndrewBuck> I am going to drop out for the rest of the meeting, don't think I have anything else to add.
20:09:19< ptressel> E.g. use one set of similar colors for the issue type (bug, feature request, etc.), so that another range of colors can be used for priority tags.
20:09:19< harry-wood> it's probably an building in drupal feature
20:09:30< harry-wood> built-in sorry
20:10:12< harry-wood> Yeah. I have to be going soon too
20:10:30< dodobas> ptressel: so... something along the lines of ... 'type' 'priority' 'flag'
20:10:44< ptressel> Right.
20:11:10< dodobas> flag could be 'question'|'wontfix'|...
20:11:37< dodobas> harry-wood: do you want to talk about OAM ?
20:11:55< harry-wood> Yeah I sneaked an item onto the agenda "OpenAerialMap".
20:12:00< ptressel> Use colors violet, red, red-orange, orange, yellow-orange, for priority (or something like that). Use bluish hues for type, etc.
20:12:00< harry-wood> Just a heads up that this will a big project involving lots of tech work.
20:12:17< harry-wood> The information so far… is all nicely summarised here:
20:12:31< dodobas> there is an IRC meeting this week
20:12:50< harry-wood> Yeah Wednesday 9th July 2014 at 12:00 UTC
20:13:52< harry-wood> Also it's on github here:
20:14:08< harry-wood> all we need now is some code :-)
20:14:14< ptressel> ;-)
20:14:34< dodobas> harry-wood: yeah... we'll talk more on 9th
20:14:44< dodobas> *the 9th
20:14:56< harry-wood> Speaking of github repos. I brought this one into the fold:
20:15:03-!- mkl [] has quit [Ping timeout: 480 seconds]
20:15:19< harry-wood> Because I kept mentioning Kuona, but without any proper description of it. So wrote something there
20:15:27< harry-wood> It's not running anywhere these days though
20:15:32< dodobas> harry-wood: yeah... there a similar service in the making...
20:15:46< ptressel> So somewhat like MicroMappers?
20:16:04< dodobas> I should start it up... on the server
20:16:18< harry-wood> Its a very simple idea, so it wouldn't surprise me if we end up with many similar implementations
20:16:29< ptressel> Or TomNod?
20:16:33< harry-wood> I want to see some clever integration with tasking manager
20:16:48< dodobas> ptressel: it's more like... prepare task for the TM
20:16:49< harry-wood> and that OKFN Pybossa thingy
20:16:58< ptressel> Ok
20:17:14< dodobas> harry-wood: yeah similar, but different use
20:17:28< ptressel> MicroMappers is more for crowdsourcing classification of images.
20:17:48< harry-wood> Kind of annoying that we don't have a screenshot of Kuona from when it was running
20:17:54< ptressel> And TomNod allows marking items in images.
20:18:02< harry-wood> likewise for that sandy mapmill instance. no screenshots anywhere as far as I know
20:18:16< dodobas> harry-wood: yeah, i saw that...
20:18:22< dodobas> it's not running anymore
20:18:33< dodobas> nor we have access to the source code
20:19:25< harry-wood> Right. I have to go guys. Good chatting to you
20:19:47< dodobas> see you on the 9th @ 12UTC
20:20:09< ptressel> Google finds info about it, e.g.
20:20:22< ptressel> Are we going to talk about SSL?
20:20:27< dodobas> anyone else, wants to report something, that's not on the agenda
20:20:38< dodobas> ptressel: well, i don't have anything prepared
20:20:51< ptressel> Is this for ?
20:21:16< dodobas> idea was to see which services would actually use it.. and then find out what do we nes
20:21:18< dodobas> *need
20:21:33< ptressel> Anything that transmits passwords...
20:21:58< dodobas> AFIAK.. and
20:22:18< dodobas> basically we should get a wildcard certificate for the domain
20:22:42< ptressel> Just need two subdomains? or more?
20:22:56< dodobas> but... there are a lot of 'deals' which look good, but then you find out that wildcard is not actually a wildcard
20:22:57< ptressel> Price goes up for wildcard.
20:23:19< ptressel> Right. Which is why I buy certs from
20:23:20< dodobas> ptressel: well, that's somthing we/I should invesigate...
20:23:53< ptressel> Though a friend recommended startssl -- we just used that for our localwiki site last week.
20:24:27< dodobas> which services would actually use the SSL certificate
20:24:42< ptressel> Auth.
20:24:58< ptressel> Mainly don't want to transmit passwords in plain text.
20:25:00< dodobas> i mean... if it's not available to general public... we could use self-signed certs
20:25:24< ptressel> But other sites I've worked on just enable it for everything.
20:25:52< dodobas> as most users are highly technical persons, so they understand what is a self-signed certificate
20:26:07< ptressel> Could do self-signed, with appropriate warnings.
20:26:14< dodobas> yeah, exactly
20:26:54< ptressel> I should mention why my friend recommended startssl -- it's free for single-domain certs.
20:27:22-!- alexbarth [] has quit [Remote host closed the connection]
20:27:28< dodobas> ptressel: there was one 'free for open-source' projects
20:27:34< ptressel> offers free certs for the first year with domain purchase.
20:27:38< dodobas> just a sec..
20:27:41< ptressel> Ah
20:29:18< dodobas>
20:29:45< dodobas> i haven't tried it.. nor I know if there are any issues...
20:30:03< ptressel> Huh.
20:31:03< ptressel> Looks nice.
20:31:07< dodobas> it's probably not a wildcard certificate...
20:31:14< ptressel> Unlikely.
20:31:37< dodobas> and ... 'Order must meet all vetting requirements', is what ? :)
20:31:46< ptressel> :D
20:32:12< ptressel> Means they reserve the right to decline orders.
20:32:44< ptressel> Probably a good idea, that their legal folks said they needed.
20:32:57< ptressel> Let's them avoid arguments.
20:33:15-!- pierregiraud [] has quit [Quit: pierregiraud]
20:33:22< dodobas> so, there are options, but... do we really need it, or we could just use a self-signed
20:33:59-!- neiljp [] has quit [Remote host closed the connection]
20:34:07< ptressel> Re wildcard, reason I was asking if there were only a few subdomains is, some CAs include a small # of subdomains in a normal single site cert.
20:34:56< ptressel> Using a cert recognized by browsers will avoid scaring people, but it's only once, and then they add the exception.
20:35:24< ptressel> So if there's a notice that they will get the scary warning when they register, it's probably ok.
20:36:12-!- neiljp [] has joined #hot
20:36:30< ptressel> Then use http if the session is not logged in, and https on the auth page and optionally while logged in.
20:36:48< ptressel> If you have a real cert, you can just use https for everything.
20:37:25< AndrewBuck> My recommendation would be to get a real cert (don't think they are that much) and use it everywhere..
20:37:50< ptressel> I second's what I've been doing elsewhere.
20:37:52< AndrewBuck> We need to fight censorship and spying in every way possible.
20:39:13< ptressel> https for everything is set and forget -- no need to only do https on some pages, so you can show the notice about the self-signed warning before people encounter it.
20:39:52< ptressel> Could ask the Globalsign folks if they will include a handful of subdomains.
20:40:13< dodobas> ptressel: it msut be for everything... it would require code update to use https selectivly
20:40:39< dodobas> ptressel: well, you could... ask for wildcard certificate :)
20:40:40< ptressel> Yep.
20:40:47< ptressel> Heh.
20:41:34< dodobas> anything else, or the TWG has ended ?
20:44:22< AndrewBuck> nothing from me
20:47:56< ptressel> Oh, should mention, if you get a cert for xxx.yyy, most CAs throw in for free.
20:48:56< ptressel> I usually get single site certs, and use path rather than subdomain.
20:50:25< ptressel> This looks like it has some Sandy MapMill images:
20:51:00-!- neiljp [] has quit [Remote host closed the connection]
20:52:23< ptressel> Looks like code is available:
20:53:26< ptressel> dodobas harry-wood ^^ if you want images / code for the sandy mapmill
20:54:02< dodobas> ptressel: can you add that information to the TWG wiki, repos at leas
20:54:05< dodobas> *least
20:54:33< ptressel> Ok.
20:54:44< ptressel> Found the Knight News challenge page about it.
20:59:15< ptressel> Or not.
20:59:23< ptressel> dodobas, I don't have permission to edit.
20:59:38< dodobas> what ?, you need to login
20:59:44< ptressel> Eeeuw, it logged me out.
20:59:50< ptressel> I just logged in.