--- Log opened Sat Jul 08 00:00:29 2017
00:37 -!- xenos1984 [~xenos1984@22-164-191-90.dyn.estpak.ee] has quit [Quit: Leaving.]
01:46 -!- noradtux [~noradtux@2a02:2028:61b:4f01::1] has quit [Ping timeout: 276 seconds]
01:50 -!- noradtux [~noradtux@port-7551.pppoe.wtnet.de] has joined #navit
04:18 -!- xenos1984 [~xenos1984@22-164-191-90.dyn.estpak.ee] has joined #navit
05:55 -!- ColdFyre [~lenny@c-73-221-248-99.hsd1.wa.comcast.net] has quit [Ping timeout: 246 seconds]
05:56 -!- ColdFyre [~lenny@c-73-221-248-99.hsd1.wa.comcast.net] has joined #navit
06:32 -!- xenos1984 [~xenos1984@22-164-191-90.dyn.estpak.ee] has quit [Ping timeout: 240 seconds]
07:04 #navit: < jkoan> The build on ci for the pull request ogre oh2fzo is not failing because of him. The ci script for tomtom builds trys to download a non existing file, and this is what it brakes
08:20 #navit: < oh2fzo> I also noticed it failed. Damn libpng :D
09:03 #navit: < jkoan> i fixed it in some time only waiting for kazer to have the ci free
09:09 #navit: < jkoan> oh2fzo: i will send you a pull request to your fork when i am finished, so you could only merge it and your branch is clean :)
09:55 -!- pini [~pini@bou-fi.pustule.org] has joined #navit
09:59 -!- pini [~pini@bou-fi.pustule.org] has quit [Ping timeout: 248 seconds]
13:04 #navit: < oh2fzo> jkoan: Do you mean fixing the map movement?
13:17 #navit: < jkoan> No, only the ci building of your pull request (it's a problem with the ci)
13:18 #navit: < oh2fzo> Oh, ok.
13:20 #navit: < jkoan> Sleske also closed you pull request, but I think this is like thinking to fast
13:20 #navit: < jkoan> Did you have seen his comment?
13:23 #navit: < oh2fzo> Yeah, I just checked and he is correct. But I think the drawing engine isn't done too well. The delay system is kind of bad because it limits even powerful systems.
14:07 #navit: < jkoan> I also think that we must rework the rendering. Maybe not with the way you tryed, but we definitely must do something. I hate those edges while moveing, it only looks old and crappie
14:11 #navit: < oh2fzo> And the low/capped frame rate even on high-end hardware.
14:49 -!- pini [~pini@bou-fi.pustule.org] has joined #navit
14:53 #navit: < Celelibi> What is the "car detailled" vehicle?
15:04 #navit: < jkoan> as i see the only difference is the value of route_depth
15:05 #navit: < jkoan> http://wiki.navit-project.org/index.php/Configuration/Full_list_of_options#vehicleprofile here you find something about that
15:20 #navit: < jkoan> KaZeR: probably we dont need to update the android thing because circleci does have android sdk files included: https://circleci.com/docs/1.0/build-image-trusty/#android
15:43 -!- pini [~pini@bou-fi.pustule.org] has quit [Ping timeout: 240 seconds]
16:45 #navit: <@KaZeR> hi there
16:45 #navit: < jkoan> hi KaZeR
16:47 #navit: <@KaZeR> Celelibi: the default car routing profile was designed for lower ends devices. Sometimes it will fail to find a route like in this example : http://trac.navit-project.org/ticket/1296
16:48 #navit: <@KaZeR> because there is a segment of a smaller route in the middle which isn't picked up. Enabling this profile loads more roads with lower priority, but it consumes more cpu and memory
16:49 #navit: <@KaZeR> jkoan: what do you mean with updating the android thing?
16:49 #navit: < jkoan> the "android" command become deprecated
16:55 #navit: <@KaZeR> ah! yeah in our case it's packed in the circleci container. We can update when we want though
16:57 #navit: < jkoan> but right now i try to update the ndk to the last version
16:57 #navit: < jkoan> but not really lucky
17:00 #navit: <@KaZeR> did you get a chance to test the qt5/qml2 build?
17:00 #navit: < jkoan> http://paste.ubuntu.com/25047635/
17:01 #navit: < jkoan> yes, which plattform?
17:06 #navit: < Celelibi> KaZeR: that's what I thought. The Car detailled shaved off a good fraction of the time.
17:06 #navit: < Celelibi> Is the algorithm still based on Dijkstra?
17:07 #navit: < Celelibi> (Or A* for that matter.)
17:07 #navit: < Celelibi> I mean, there have been a shit ton of research on the topic of routing in large graphs.
17:08 #navit: < Celelibi> Especially for the kind of graphs that make up the typical road systems.
17:09 #navit: <@KaZeR> Celelibi: yes still Dijsktra. I believe that Jandegr was working on some changes but haven't seen the code yet
17:10 #navit: <@KaZeR> jkoan: odd. looks like one of your include path is borked
17:12 #navit: < jkoan> KaZeR: ttps://pastebin.com/nEKWt70b did this helps you? probaly you could say me whats wrong
17:16 #navit: <@KaZeR> where's your stdlib.h files on your system?
17:16 #navit: < jkoan> which one, the one for android?
17:16 #navit: <@KaZeR> Celelibi: would you like to help writing a more modern routing algorithm?
17:16 #navit: <@KaZeR> jkoan: yep
17:17 #navit: < jkoan> one of them: https://pastebin.com/qC8yf3AQ :D
17:18 #navit: < jkoan> (found pastebinitenv
17:18 #navit: < jkoan> *pastebinit
17:19 #navit: < Celelibi> KaZeR: I'd love to help. But I already have many other open source projects I'd love to help too.
17:19 #navit: < Celelibi> My priority goes to those I use more often.
17:20 #navit: <@KaZeR> jkoan: looks like you're building for android 8 but your ndk provides the files starting for version 9
17:20 #navit: <@KaZeR> Celelibi: makes sense. Out of curiosity, what are the other projects?
17:20 #navit: < jkoan> yes, but it dos not work with android-9 also
17:21 #navit: <@KaZeR> bonus question : what do you need to use navit more often? :)
17:21 #navit: < jkoan> KaZeR: nice trick ^^
17:21 #navit: <@KaZeR> ;)
17:22 #navit: < jkoan> okay, now its working at least this one. but next problem :D
17:24 #navit: < Celelibi> KaZeR: syslinux (a boot loader), HT Editor (a disassembler), yafc (an FTP client), and many projects of my own, like an image-based maze solver, a FUSE-based FS for a website, and many more.
17:25 #navit: <@KaZeR> Celelibi: nice. I've used syslinux a lot in the past. Nifty piece of software
17:25 #navit: < Celelibi> I've fixed some nasty bugs in it.
17:26 #navit: < Celelibi> Tracking bugs in it was painfully fun. :D
17:26 #navit: < jkoan> https://pastebin.com/jMWxHnue i hate souch problems....
17:44 -!- pini [~pini@bou-fi.pustule.org] has joined #navit
17:59 #navit: < jkoan> Kazer you asked me to test qml2, when, where and where I got it? :)
18:00 #navit: <@KaZeR> jkoan: one way is the raspi image, another way is the qt5_qml_gui branch. I'm updating it to match trunk right now
18:00 #navit: < jkoan> so you send me a new image for the pi?
18:01 #navit: <@KaZeR> it's currently building : https://circleci.com/gh/navit-gps/buildroot/54 but i expect a few more issues. Once it's fixed we can easily build for all raspi versions
18:02 #navit: <@KaZeR> Celelibi: if you like to track odd bugs you can definitely have fun with us ;)
18:02 #navit: < Celelibi> I don't doubt about that. :)
18:04 #navit: < jkoan> KaZeR: failed
18:04 #navit: <@KaZeR> jkoan: yeah circleci is not as good as it used to be. Sometimes builds die with no reason
18:06 #navit: < jkoan> probably 2.0 of cirlcle helps us
18:07 #navit: <@KaZeR> we could try. it has some bad limitations though. The doc sucks for v2 for example. You can't build on tags without a dirty hack, and a few more issues
18:08 #navit: < jkoan> not nice :/
18:08 #navit: < jkoan> in the worth case we need a own jenkins
18:10 #navit: <@KaZeR> yeah we could do that too. i can probably sponsor a slave.
18:10 #navit: <@KaZeR> it's too bad cause circleci used to be really, really nice. They are probably victims of the success of their free plan
18:11 #navit: < jkoan> who hosts the navit servers generally?
18:17 #navit: <@KaZeR> most of the services are currently on a server owned by cp15. Trac used to be on my own server, but i moved it to another server of mine which also hosts wiki2 and forums2
18:18 #navit: <@KaZeR> i'll probably do the switch wiki and forum to the new server today
18:19 #navit: <@KaZeR> we still have the map processing, map serving and the legacy build service running on cp15's server
18:20 #navit: < jkoan> the legacy build service is broken over more then 50%
18:28 #navit: <@KaZeR> true, so we need to port the missing builds to the new build system.. like wince :)
18:29 #navit: < jkoan> what about "only run the ci build scripts?"
18:30 #navit: <@KaZeR> what do you mean?
18:31 #navit: < jkoan> why we let the legacy build service not only run the ci/build*files ?
18:35 #navit: <@KaZeR> because the ci/build* files will only work on the circleci v1 machines. Paths, packages and packages version are really different
18:36 #navit: <@KaZeR> like your NDK issue :)
18:36 #navit: < jkoan> yes, okay, but if i build on my own machine i always work with that files ;) probably we can reorganize them to work also with normal builds. Shell Scripting is my every day job on work ;)
18:37 #navit: < jkoan> my ndk issue was that i forgot to add: -DANDROID_NDK_API_VERSION=19
18:40 #navit: <@KaZeR> ah, good to know
18:40 #navit: <@KaZeR> (that you rock bash and the fix for your issue)
18:41 #navit: < jkoan> on work mainly ksh :/ but it okay anyway
18:43 #navit: < jkoan> i am so happy to have the way back machine for www.opentom.org
18:44 #navit: <@KaZeR> haha yeah that thing is useful
18:48 #navit: < jkoan> i hope that we manage it to migrate the tomtom build to buildroot
18:58 #navit: <@KaZeR> well we could do it if the goal is to replace the tomtom OS, but some people like to just run navit on their usual tomtom OS
18:58 #navit: <@KaZeR> buildroot won't solve the second use case
19:00 #navit: < jkoan> yes, this is right. But if its a full os we can say then just to grap another sd card, and it will be a full compatible navi, with map downloadf over bluetooth and so on
19:03 #navit: <@KaZeR> that would be cool. Can you boot on another sd card?
19:04 #navit: < jkoan> my device only boots from the sd card :D
19:04 #navit: < jkoan> look at this: http://docplayer.org/storage/28/12552262/12552262.pdf (its german, but its really interesting)
19:05 #navit: <@KaZeR> Oops! Forbidden.
19:05 #navit: < jkoan> what?!
19:05 #navit: < jkoan> its not mine, but its loading...
19:05 #navit: < jkoan> wait
19:05 #navit: < jkoan> http://docplayer.org/12552262-Hacking-into-tomtom-go-thomas-kleffel-christian-daniel-und-das-www-opentom-org-projekt.html better?
19:06 #navit: <@KaZeR> all right i found one nice thing with circleci v2. Their caching system is much better
19:06 #navit: < jkoan> nice :)
19:06 #navit: <@KaZeR> yep that link works
19:06 #navit: < jkoan> and if i have seen it right, thy support that you use own docker images :D
19:08 #navit: <@KaZeR> yeah they do. That's actually the biggest improvement
19:09 #navit: <@KaZeR> allows us to build on 16.04 for example
19:09 #navit: <@KaZeR> https://circleci.com/gh/navit-gps/buildroot/72
19:09 #navit: <@KaZeR> the v2 boxes seems to be less impacted by the noisy neighbors effect. could be a good win
19:09 #navit: < jkoan> does it build faster? :)
19:11 #navit: <@KaZeR> it should as the boxes don't seem to have a permanent load of 80 :)
19:12 #navit: < jkoan> really nice :)
19:12 #navit: <@KaZeR> building the toolchain on v1 takes 47 min. we'll know soon about v2
19:13 #navit: < jkoan> and because cache is better on v2 we dont need to build it every time?
19:14 #navit: <@KaZeR> using v1, the cache is updated only if all your dependencies steps are successful. Using v2, you can cache (and restore) between steps
19:14 #navit: <@KaZeR> so we can build the toolchain, and cache it directly even if the actual build fails. Which saves us ~50 min
19:18 #navit: <@KaZeR> it does seem faster so far. Building glibc on v1 took 7 min, 4 on v2
19:22 #navit: <@KaZeR> wow 18 min for the toolchain on v2, vs 47 on v1
19:22 #navit: < jkoan> nice!!!
19:23 #navit: < jkoan> UPDATE IT FOR NAVIT ALSO?
19:23 #navit: < jkoan> ups
19:23 #navit: < jkoan> i dont want caps
19:23 #navit: <@KaZeR> yeah probably we should do it. But i want to try maptool first :) we were hitting some limitations before
19:27 #navit: < jkoan> If you say maptool I also have a related question, why it is inside navit and not in its own repo?
19:27 #navit: <@KaZeR> because it is important that they share the binfile code for compatibility. Also gives anybody the possibility to convert their own maps
19:27 #navit: <@KaZeR> was useful when we did not have a map server :)
19:28 #navit: < jkoan> Okay, this seems logical
19:30 #navit: < jkoan> BTW I have seen that navit permit to download files bigger than 2 GB this not always makes sense, because devices which uses ext4 for example could store bigger files. The only question is how to handle this eg. Find out what the device is able to store
19:35 #navit: <@KaZeR> that's tricky. I think that instead of downloading huge files we should do a better job at providing sub-maps instead
19:35 #navit: <@KaZeR> i doubt that a lot of people need "Europe" for example. Usually 2-3 countries at best
19:39 #navit: < jkoan> Hm okay
19:39 #navit: < jkoan> But back to topic ;)
19:42 #navit: < jkoan> Why you have running 2 builds on the same time
19:43 #navit: <@KaZeR> different branches?
19:43 #navit: < jkoan> 72 and 73
19:44 #navit: <@KaZeR> ah yeah. i fixed something in 73, but i did not kill 72 because if it's going to fail it'll fail sooner
19:45 #navit: <@KaZeR> the new caching feature will enable us to cache the maptool binary between builds in https://circleci.com/gh/navit-gps/maptool/798
19:45 #navit: <@KaZeR> that's neat
19:45 #navit: <@KaZeR> oh well it takes 23s to build so not sure it's worth it :)
19:48 #navit: < jkoan> 23 seconds to build full navit?
19:48 #navit: < jkoan> Currently we need around 10minutes?!
19:49 #navit: <@KaZeR> without graphics drivers :)
19:49 #navit: < jkoan> This makes more sense
19:49 #navit: <@KaZeR> we can shave off a couple more seconds, it did the png 2 svg conversion
19:51 #navit: < jkoan> Question about svg, does android (ci builds) use svg?!
19:52 #navit: < jkoan> Because the icons not seems to be
19:53 #navit: <@KaZeR> i'm not sure i'd need to check. i think that it wasn't by default but i don't remember if we changed it
19:55 #navit: < jkoan> I will try it when I have time to. Because svg looks very much nicer
19:59 #navit: < jkoan> Did you cache the osmosis file? It would make sense, then you can use the diff files to be faster.
19:59 #navit: < jkoan> *osm
20:00 -!- pini [~pini@bou-fi.pustule.org] has quit [Ping timeout: 276 seconds]
20:01 #navit: <@KaZeR> yeah the goal is to cache the pbf from geofabrik. first i want to see if the other issues (especially on time) are gone in v2
20:05 #navit: < jkoan> Too long with no output :D
20:12 -!- xenos1984 [~xenos1984@22-164-191-90.dyn.estpak.ee] has joined #navit
20:14 #navit: < jkoan> KaZeR: dot:mega
20:23 #navit: <@KaZeR> jkoan: thanks :)
20:35 -!- pini [~pini@bou-fi.pustule.org] has joined #navit
20:45 -!- pini [~pini@bou-fi.pustule.org] has quit [Ping timeout: 260 seconds]
21:00 #navit: < jkoan> Kazer same thing with the maptool, timed out
21:02 #navit: < jkoan> Try add a wrapper script that prints dots from the background and then run maptool
21:05 #navit: <@KaZeR> actually we can leverage the no_output_timeout config instead :) https://circleci.com/docs/2.0/configuration-reference/#run
21:17 #navit: < jkoan> This is even better
21:20 #navit: < jkoan> If you set this also to the wget you will have even smaller logs ;)
21:35 #navit: <@KaZeR> yeah true. But having a sense of how much data is downloaded might be kind of useful maybe
22:24 #navit: <@KaZeR> https://circleci.com/gh/navit-gps/maptool/802 phase 1 succeeded \o/
23:33 -!- Horwitz [~mich1@p200300800E101300022268FFFE64E7C4.dip0.t-ipconnect.de] has quit [Ping timeout: 240 seconds]
23:41 -!- xenos1984 [~xenos1984@22-164-191-90.dyn.estpak.ee] has quit [Quit: Leaving.]
23:46 -!- Horwitz [~mich1@p200300800E106900022268FFFE64E7C4.dip0.t-ipconnect.de] has joined #navit
23:46 -!- mode/#navit [+o Horwitz] by ChanServ
23:56 -!- ilovekiruna [~ilovekiru@ip2505a0ef.dynamic.kabel-deutschland.de] has joined #navit
--- Log closed Sun Jul 09 00:00:31 2017