--- Log opened Wed Apr 15 00:00:23 2015
01:05 -!- KaZeR [~KaZeR@64.201.252.132] has quit [Remote host closed the connection]
01:06 -!- KaZeR [~KaZeR@64.201.252.132] has joined #navit
01:06 -!- mode/#navit [+o KaZeR] by ChanServ
01:10 -!- KaZeR [~KaZeR@64.201.252.132] has quit [Ping timeout: 250 seconds]
01:15 #navit: < Navit> See compile results history at http://download.navit-project.org/logs/navit/stats.html
01:22 -!- xenos19841 [~xenos1984@43-29-196-88.dyn.estpak.ee] has quit [Quit: Leaving.]
01:48 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has joined #navit
01:48 -!- mode/#navit [+o KaZeR] by ChanServ
02:10 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has quit [Remote host closed the connection]
02:11 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has joined #navit
02:11 -!- mode/#navit [+o KaZeR] by ChanServ
02:16 -!- KaZeR [~KaZeR@c-67-161-64-186.hsd1.ca.comcast.net] has quit [Ping timeout: 245 seconds]
02:36 -!- j_f-f [~quassel@orion2589.server4you.de] has quit [Remote host closed the connection]
02:39 -!- j_f-f [~quassel@orion2589.server4you.de] has joined #navit
05:43 -!- xenos1984 [~xenos1984@43-29-196-88.dyn.estpak.ee] has joined #navit
05:55 -!- circleci-bot [~circleci-@ec2-54-89-147-137.compute-1.amazonaws.com] has joined #navit
05:55 #navit: < circleci-bot> Failed: jandegr's build (#254; push) in navit-gps/navit (jandegr-maptool-europe) -- https://circleci.com/gh/navit-gps/navit/254
05:55 -!- circleci-bot [~circleci-@ec2-54-89-147-137.compute-1.amazonaws.com] has quit [Client Quit]
06:33 -!- Robotaxi [3ef5dbf5@gateway/web/freenode/ip.62.245.219.245] has joined #navit
06:33 -!- circleci-bot [~circleci-@ec2-54-157-215-221.compute-1.amazonaws.com] has joined #navit
06:33 #navit: < circleci-bot> Failed: jandegr's build (#256; push) in navit-gps/navit (jandegr-maptool-europe) -- https://circleci.com/gh/navit-gps/navit/256
06:33 -!- circleci-bot [~circleci-@ec2-54-157-215-221.compute-1.amazonaws.com] has quit [Client Quit]
07:12 -!- jjelen [jjelen@nat/redhat/x-spxdfxpdkdvdzgwv] has joined #navit
07:12 -!- woglinde [~henning@fb-n15-11.unbelievable-machine.net] has joined #navit
07:21 -!- woglinde [~henning@fb-n15-11.unbelievable-machine.net] has quit [Ping timeout: 250 seconds]
07:51 -!- circleci-bot [~circleci-@ec2-54-144-202-16.compute-1.amazonaws.com] has joined #navit
07:51 #navit: < circleci-bot> Failed: jandegr's build (#257; ssh by jandegr) in navit-gps/navit (jandegr-maptool-europe) -- https://circleci.com/gh/navit-gps/navit/257
07:51 -!- circleci-bot [~circleci-@ec2-54-144-202-16.compute-1.amazonaws.com] has quit [Client Quit]
08:41 -!- saschao [c1af31ea@gateway/web/freenode/ip.193.175.49.234] has joined #navit
11:39 -!- saschao [c1af31ea@gateway/web/freenode/ip.193.175.49.234] has quit [Ping timeout: 246 seconds]
11:45 -!- jjelen [jjelen@nat/redhat/x-spxdfxpdkdvdzgwv] has quit [Ping timeout: 264 seconds]
12:59 -!- woglinde [~henning@fb-n15-11.unbelievable-machine.net] has joined #navit
13:07 -!- woglinde [~henning@fb-n15-11.unbelievable-machine.net] has quit [Ping timeout: 255 seconds]
14:02 -!- jjelen [jjelen@nat/redhat/x-odtxufdhpglgufmn] has joined #navit
15:29 -!- jandegr [57416e0a@gateway/web/freenode/ip.87.65.110.10] has joined #navit
15:31 -!- circleci-bot [~circleci-@ec2-23-20-225-32.compute-1.amazonaws.com] has joined #navit
15:31 #navit: < circleci-bot> Timed out: jandegr's build (#259; push) in navit-gps/navit (jandegr-maptool-europe) -- https://circleci.com/gh/navit-gps/navit/259
15:31 -!- circleci-bot [~circleci-@ec2-23-20-225-32.compute-1.amazonaws.com] has quit [Client Quit]
16:08 -!- Robotaxi [3ef5dbf5@gateway/web/freenode/ip.62.245.219.245] has quit [Ping timeout: 246 seconds]
17:14 -!- jjelen [jjelen@nat/redhat/x-odtxufdhpglgufmn] has quit [Ping timeout: 264 seconds]
17:21 -!- KaZeR [~KaZeR@64.201.252.132] has joined #navit
17:21 -!- mode/#navit [+o KaZeR] by ChanServ
17:32 #navit: <@KaZeR> hi there
17:58 -!- tryagain [~quassel@178.216.76.73] has joined #navit
18:03 #navit: < Navit> kazer_ * r6083 /trunk/navit/po/lt.po.in: Update:Translations:Updated Lithuanian translation from launchpad http://sourceforge.net/p/navit/code/6083/
18:16 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has joined #navit
18:16 -!- circleci-bot [~circleci-@ec2-54-163-154-185.compute-1.amazonaws.com] has joined #navit
18:16 #navit: < circleci-bot> Success: pgrandin's build (#265; push) in navit-gps/navit (trunk) -- https://circleci.com/gh/navit-gps/navit/265
18:16 -!- circleci-bot [~circleci-@ec2-54-163-154-185.compute-1.amazonaws.com] has quit [Client Quit]
18:18 #navit: <@KaZeR> hey tryagain
18:19 #navit: <@KaZeR> i was able to fully update a map using the map updater POC. good news: it works. Bad news : on the client side, it took 1931 minutes...
18:20 #navit: <@KaZeR> i will need to do some profiling but i think that the tile hashing is currently what's using most of the time.
18:21 #navit: <@KaZeR> hashing the full planet in order to update the md5 hash table took only 80 minutes on the server side
18:22 #navit: < tryagain> i do not think we need hash computation at all
18:23 #navit: < tryagain> is there any permanent place where i could write my thoughts?
18:24 #navit: <@KaZeR> not yet but we should do it. Either a ticket or a wiki page, whatever you prefer
18:25 #navit: <@KaZeR> the point of using hash computation for me is to avoid downloading tiles that do not need to be updated, and saving bandwidth on the go
18:25 #navit: <@KaZeR> but i'd love to hear your thoughts on this :)
18:26 #navit: < tryagain> i think we could compare two binfiles tile-by-tile byte-by-byte
18:26 #navit: < tryagain> Then on matching tiles we set timestamp equal to the one of the older map
18:26 #navit: < tryagain> so user should download the tiles newer than he has
18:27 #navit: < tryagain> this of course works only for one map source
18:28 -!- circleci-bot [~circleci-@ec2-23-22-164-64.compute-1.amazonaws.com] has joined #navit
18:28 #navit: < circleci-bot> Success: pgrandin's build (#266; push) in navit-gps/navit (master) -- https://circleci.com/gh/navit-gps/navit/266
18:28 -!- circleci-bot [~circleci-@ec2-23-22-164-64.compute-1.amazonaws.com] has quit [Client Quit]
18:28 #navit: < tryagain> but different sources could be known by a global map item and attribute url
18:29 #navit: <@KaZeR> good point. i was actually considering using timestamps as a way to resume an update
18:29 #navit: <@KaZeR> but checking only timestamps should be faster indeed
18:29 #navit: <@KaZeR> but checking hash is less likely to break the map too
18:29 -!- circleci-bot [~circleci-@ec2-54-224-140-130.compute-1.amazonaws.com] has joined #navit
18:29 #navit: < circleci-bot> Success: pgrandin's build (#267; push) in navit-gps/navit (trunk) -- https://circleci.com/gh/navit-gps/navit/267
18:29 -!- circleci-bot [~circleci-@ec2-54-224-140-130.compute-1.amazonaws.com] has quit [Client Quit]
18:30 #navit: < tryagain> yes, but we can have different measures to ensure map consistency. TCP itself is unlikely to change data inside the stream.
18:31 #navit: < tryagain> then we have crc32 inside zip file
18:32 #navit: <@KaZeR> it's definitely worth a try, as the current way using hash takes way to much time anyway
18:39 #navit: < tryagain> md5 is much better for some applications, but in our case crc32 should be sufficient.
18:40 #navit: < tryagain> I also think we should have some companion file to be used instead of zip file directory.
18:41 #navit: < tryagain> zip file directory is very inefficient to store 700000 tile records, many of which do not match our area of interestand left empty
18:42 -!- circleci-bot [~circleci-@ec2-54-196-103-150.compute-1.amazonaws.com] has joined #navit
18:42 #navit: < circleci-bot> Success: pgrandin's build (#268; push) in navit-gps/navit (master) -- https://circleci.com/gh/navit-gps/navit/268
18:42 -!- circleci-bot [~circleci-@ec2-54-196-103-150.compute-1.amazonaws.com] has quit [Client Quit]
18:43 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has quit [Ping timeout: 256 seconds]
18:45 #navit: <@KaZeR> oh yeah, and actually the zipdir is currently an issue, as my code will only update what you already have in your map, but will not expand your map. so it does not yet solve the "first map download" problem
18:46 #navit: < tryagain> i guess it should be possible to fix
18:47 #navit: < tryagain> regarding size efficiency: each zipfile record has an alphabetic file name (19 for each tile), but actual information needed is only 41 bits (2 bits to encode tile name letter a-d for most deep level 18, and 5 bits to store its  length)
18:48 #navit: < tryagain> also we need 64-bit pointer to where the tile starts in actual zip file
18:48 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has joined #navit
18:50 #navit: < tryagain> then 32-bit zip file number (so we do not have to store skipped tile records in this file)
18:50 #navit: < tryagain> tile size is probably available from zip file local file header, so no need to store it
18:53 #navit: < tryagain> so it looks like we need only 12 bytes per tile record, resulting in 9 megabyte companion file for whole planet
18:56 #navit: < tryagain> wait, we also need timestamp data here
18:57 #navit: < tryagain> we could use 32bit unix time, but thats probably an overshoot
18:57 #navit: < tryagain> (un)fortunately, we are not able to build a map per second
18:59 #navit: <@KaZeR> haha
19:01 #navit: < tryagain> with 1-byte counter, we could have 256 versions of planet file.
19:01 #navit: < tryagain> should be enough
19:02 #navit: < tryagain> so 13 bytes per file, 10 megabytes for planet
19:02 #navit: < tryagain> currently we probably have over 70 megabytes
19:04 #navit: < tryagain> 10 megabytes would have to be downloaded during each update (unless we split this companion file into smaller parts geographically)
19:05 #navit: <@KaZeR> that's an interesting approach. we should really put that somewhere to get the ball rolling
19:07 #navit: < tryagain> after companion file is downloaded, client has information enough to decide which parts of flat zip file they actually need
19:08 #navit: < tryagain> and central zip file directory could be reconstructed using data we have in companion file
19:10 #navit: <@KaZeR> that sounds good indeed
19:11 #navit: < tryagain> actually, our companion file could also be used to improve performance of normal map operations. We cache it in memory and use to access tiles instead of the central directory.
19:31 #navit: <@KaZeR> you would generate the companion file via maptool during the map processing too ?
19:34 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has quit [Ping timeout: 245 seconds]
19:42 #navit: < tryagain> I do not think it's a big deal wheher to put this code inside maptool or in some external utility. I don't think we'll have a speed benefit from merging this function into maptool,  but it would be handy to have a single utility for map processing which could be run in one pass.
19:43 #navit: <@KaZeR> agreed
19:46 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has joined #navit
20:22 -!- jandegr [57416e0a@gateway/web/freenode/ip.87.65.110.10] has quit [Quit: Page closed]
21:08 #navit: <@KaZeR> tryagain: i'll put a draft of discussion in the wiki based on your thoughts later on today
21:08 #navit: <@KaZeR> s/discussion/blueprint/ actually
21:09 #navit: <@KaZeR> unless you feel that a trac ticket would be better suited. which it might
21:13 #navit: < tryagain> Thank you. But i think a ticket would be a better option.
21:15 #navit: <@KaZeR> yeah right.
21:22 -!- tryagain [~quassel@178.216.76.73] has quit [Remote host closed the connection]
21:31 -!- _rd [~rd@pD9E7D97D.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
22:40 -!- xenos1984 [~xenos1984@43-29-196-88.dyn.estpak.ee] has quit [Quit: Leaving.]
23:06 -!- noradtux [~noradtux@2a02:8108:28bf:fc94::1] has quit [Ping timeout: 252 seconds]
23:07 -!- noradtux [~noradtux@2a02:8108:28bf:fc94::1] has joined #navit
23:38 -!- He4dShOt_ [~He4dShOt@2001:41d0:1:7678::1] has joined #navit
23:39 -!- He4dShOt [~He4dShOt@unaffiliated/he4dshot] has quit [Ping timeout: 248 seconds]
23:39 -!- He4dShOt_ is now known as He4dShOt
23:39 -!- He4dShOt [~He4dShOt@2001:41d0:1:7678::1] has quit [Changing host]
23:39 -!- He4dShOt [~He4dShOt@unaffiliated/he4dshot] has joined #navit
--- Log closed Thu Apr 16 00:00:23 2015