00:20:12 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
00:20:21 *** mperry has quit (Ping timeout: 260 seconds)
00:20:21 *** mperry_ is now known as mperry
00:38:25 *** springmeyer (~springmey@c-76-28-156-154.hsd1.wa.comcast.net) has joined #mapnik
00:41:04 <StormTide> any of you guys setup renderd before
00:41:26 <StormTide> getting a socket bind failed for: /tmp/osm-renderd error..
00:51:28 <tcb_> StormTide: just got into renderd
00:51:58 <StormTide> im trying to figure out the config
00:52:15 <tcb_> what do you have socketname set to in /etc/renederd.conf?
00:52:47 <StormTide> its commented
00:52:47 <tcb_> default may be commented out -- try removing ';'
00:52:50 <StormTide> ah
00:55:00 <StormTide> should renderd once started show up in the ps tree?
00:55:05 <StormTide> cuz for me it just seems to run and quit
00:55:13 <tcb_> check /var/log/syslog
00:55:36 <StormTide> it looks ok, lemme paste
00:55:38 <tcb_> maybe datasource not available (postgres  not running)?
00:55:57 <StormTide> im not using a postgres datasource
00:56:02 <StormTide> using a custom mapnik xml
00:56:11 <StormTide> and shp files
00:56:42 <StormTide> http://www.phpriot.com/2655
00:57:45 <StormTide> the mapnik file seems to load and produce a map just fine... so... im pretty sure its just the tile server thats having an issue
00:58:28 <tcb_> yah
00:58:59 <StormTide> [Mon Feb 01 16:59:19 2010] [info] [client 192.168.1.102] tile_translate: uri(/tiles/)
00:59:00 <StormTide> [Mon Feb 01 16:59:19 2010] [info] [client 192.168.1.102] tile_translate: baseuri(/tiles/) name(default)
00:59:00 <StormTide> [Mon Feb 01 16:59:19 2010] [info] [client 192.168.1.102] tile_storage_hook: handler((null)), uri(/tiles/), filename(/var/www/tiles/), path_info((null))
00:59:00 <StormTide> [Mon Feb 01 16:59:19 2010] [error] [client 192.168.1.102] File does not exist: /var/www/tiles
00:59:03 <StormTide> shows up in apache
00:59:37 <tcb_> not sure if it matters, but you might mkdir /var/www/tiles
01:00:09 <StormTide> then it just shows an empty dir
01:01:07 <tcb_> is that font line the last thing you see in syslog?
01:01:15 <tcb_> (or just the last on the paste?)
01:01:21 <StormTide> yes last in the long
01:01:28 <tcb_> I get a "Starting stats thread"
01:01:56 <tcb_> check yer fonts directory -- I've had renderd quit b/c of issues with fonts
01:02:04 <StormTide> hrm
01:02:25 <StormTide> it looks ok...
01:02:29 <StormTide> what would i be looking for
01:03:52 <tcb_> dunno -- are there any files in there that aren't fonts?
01:04:01 <StormTide> couple text files
01:04:16 <StormTide> Sconsscript
01:04:30 <StormTide> lots really
01:04:41 <StormTide> i can try nuking the files in there but it looks only to be loading ..tfs
01:04:43 <StormTide> ttfs
01:05:21 <tcb_> oh yeah, I got plenty of txts too, I gues
01:05:21 <tcb_> s
01:06:21 <StormTide> nuked everything but the Deja ttf files
01:06:23 <StormTide> still no good
01:07:04 <tcb_> are you running renderd as root (or as a user that can write to /var/run/renderd ?)
01:07:20 <StormTide> as www-data ... and it has perms there
01:08:13 <StormTide> aha!
01:08:15 <StormTide> -f
01:08:17 <StormTide> foregrounds it
01:08:19 <StormTide> and now i see the error
01:08:23 <StormTide> which didnt go to syslog
01:08:28 <tcb_> oh!
01:08:32 <StormTide>   what():  Could not create datasource. No plugin found for type 'shape' (encountered during parsing of layer '43shpdepare' in map '/var/www/mapnikpointswithtext.xml')
01:08:37 <StormTide> looks like it cant find its plugins
01:08:52 <tcb_> gotcha (and thanks -- "-f" good to know!)
01:11:13 <StormTide> now renderd starts right, but the /tiles dir still doesnt hit the right hook from apache
01:12:59 <Ldp__> did you install and configure mod_tile?
01:13:35 <StormTide> yep
01:13:40 <StormTide> and its loading and i think conf'd right
01:14:05 <StormTide> i have a tile.load file
01:14:09 <StormTide> and then the vhost entires
01:15:02 <Ldp__> your apache errorlog should tell you if it intercepted the tiles url
01:15:25 <Ldp__> tile_translate: baseuri(/osm-maxspeed/) name(osm-maxspeed), referer:
01:15:28 <StormTide> [Mon Feb 01 17:13:11 2010] [info] [client 192.168.1.102] tile_storage_hook: handler((null)), uri(/tiles/), filename(/var/www/tiles/), path_info((null))
01:15:35 <StormTide> its doing something...
01:16:30 <tcb_> does renderd get a connection (should show up in syslog)?
01:16:36 <Ldp__> not until you point it to your mapnik plugins directory
01:16:47 <Ldp__> + fonts
01:17:11 <StormTide> doesnt appear to show anything in syslog about a connectiojn
01:17:16 <StormTide> lemme try running it in foreground again
01:17:47 <StormTide> ya, no, it gets nothing
01:18:06 <StormTide> i have it configured with the right plugins dir now
01:18:06 <Ldp__> No plugin found for type 'shape'
01:18:08 <Ldp__> solve that first
01:18:18 <StormTide> i did
01:18:22 <StormTide> renderd is loading now
01:18:56 <StormTide> (shows up in ps tree, doesn't error out with -f mode)
01:19:50 <StormTide> http://www.phpriot.com/2656 <-- in my /etc/apache2/sites-available/default
01:19:56 <StormTide> and i have the module loading just fine
01:23:11 <tcb_> I haven't tried the AddTileConfig directive myself
01:23:19 <StormTide> http://www.phpriot.com/2657
01:23:20 <tcb_> am instead using LoadTileConfigFile /etc/renderd.conf
01:23:33 <StormTide> lemme try changing to that then
01:23:34 <StormTide> sec
01:24:49 <StormTide> appears to behave identically if i do that...
01:24:55 <StormTide> care to share your config setup for apache
01:24:57 <StormTide> so i can compare
01:26:16 <tcb_> http://www.phpriot.com/2658
01:26:40 <tcb_> I think that's just the unmodified file from svn
01:27:10 <StormTide> i dont have any of those CacheDUration etc things
01:27:12 <StormTide> think that matters
01:28:01 *** cgs_bob_ has quit (Ping timeout: 264 seconds)
01:28:08 <StormTide> im just trying to invoke by going to /tiles/
01:28:12 <StormTide> that should work right?
01:28:31 <tcb_> try tiles/0/0/0.png (that's the whole world)
01:29:08 <StormTide> that did something
01:29:09 <StormTide> hrmm
01:29:36 <tcb_> handy for picking a particular tile:
01:29:37 <tcb_> http://www.maptiler.org/google-maps-coordinates-tile-bounds-projection/
01:31:02 <StormTide> very strange... it looks like its serving
01:31:05 <StormTide> but i see nothing for the images
01:31:18 <StormTide> blank tiles
01:31:23 <StormTide> hrmph
01:31:35 <tcb_> so, mod_tile will try its best for a few seconds, then return a 404
01:31:36 <StormTide> [Mon Feb 01 17:33:04 2010] [error] [client 192.168.1.102] Planet timestamp file (/var/lib/mod_tile//planet-import-complete) is missing
01:32:04 <tcb_> don't think that matters, but you can just touch /var/lib/mod_tile/planet-import-complete if you want
01:32:21 <tcb_> check top and see if renderd is busy doing something
01:32:37 <StormTide> doesnt appear so
01:32:40 <StormTide> hrmph
01:32:55 *** tcb_ is now known as tcarobruce
01:33:03 <StormTide> i only have a small map tho...
01:35:52 <StormTide> i think it might be working but i cant find my tile :(
01:36:15 <StormTide> tiles/6/10/42.png
01:36:24 <StormTide> is what that url is telling me
01:36:31 <StormTide> but thats blank.. hrm
01:36:52 <tcarobruce> you are looking at the Google coords (not TMS)?
01:37:17 <StormTide> google would be 10,21
01:37:29 <StormTide> or 10,22
01:37:47 <StormTide> tiles/6/10/22.png doesnt work either tho
01:38:03 <StormTide> (trying to find a tile for victoria bc)
01:38:34 <tcarobruce> what projection is your data / map in?
01:38:47 <StormTide> wgs84
01:39:32 <tcarobruce> mod_tile uses spherical mercator (900913)
01:39:38 <StormTide> -123.22801,48.41538
01:40:01 <StormTide> ah, crap... so i have to reproject all this eh
01:40:25 <StormTide> nothing i can throw in the mapnik xml for mod_tile that'll fix it?
01:40:40 <tcarobruce> you might just be able to set your map.xml projection and have mapnik reproject
01:41:09 <tcarobruce> try: <Map bgcolor="#8cb6d3" srs="+proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs +over">
01:41:44 <tcarobruce> (am assuming you have datasource srs's set correctly)
01:43:42 <StormTide> hazaa
01:43:45 <StormTide> that seems to have worked
01:43:55 <StormTide> what do you guys use as a slippy map viewer
01:44:02 <StormTide> for debugging with mod_tile
01:44:23 <tcarobruce> OpenLayers XYZ is your friend - paste coming up....
01:45:19 <tcarobruce> actually this is good:
01:45:21 <tcarobruce> http://dev.openlayers.org/releases/OpenLayers-2.8/examples/xyz-esri.html
01:45:58 <tcarobruce> (just remove layerExtent stuff and change ESRI url to your own)
01:47:02 <tcarobruce> (and change js src to http://openlayers.org/api/OpenLayers.js)
01:51:12 <StormTide> blah, i dont have a world map so its hard to find where the data is
01:51:42 <springmeyer> mperry: just ran into your ironpython+arcgis post, very cool
01:52:21 <springmeyer> any idea what arcgis version would be first compatible with ironpython/net stuff?
01:54:36 <StormTide> how do you translate a wgs84 latlong into this co-ord system? eg if i have a bounding box i want to use....
01:56:49 <tcarobruce> some python src at the bottom of that maptiler.org page (mentioned previously) will convert geographic lat/lon to spherical mercator
01:58:46 *** cgs_bob_ (~bob@122.135-78-65.ftth.swbr.surewest.net) has joined #mapnik
01:59:32 <springmeyer> also:
01:59:35 <springmeyer> >>> from mapnik import *
01:59:35 <springmeyer> >>> Coord(-122,49).forward(Projection('+init=epsg:900913'))
01:59:35 <springmeyer> Coord(-13580977.8768,6274861.39401)
02:00:31 <springmeyer> and:
02:00:33 <springmeyer> http://mapnik-utils.googlecode.com/svn/example_code/longlat_merc_conversion/longlat2sphericalmerc.py
02:02:10 <StormTide> is there any way to easily copy the map extent from qgis, seems when i enable extent display i disable copy from that box
02:02:48 <springmeyer> ya, there is a tiny qgis plugin called CopyPasta
02:03:25 <springmeyer> I think it is available via the 'contributed' repos that you can fetch using the QGIS plugin manager
02:04:07 <springmeyer> yep. Plugin manager > Repos > Add third party repos > 'Barry Rowlingson's Repository'
02:04:25 <StormTide> kew
02:04:25 <StormTide> thx
02:04:38 <StormTide> *sigh&
02:04:43 <StormTide> this still wont play nice
02:05:07 <springmeyer> what won't?
02:05:18 <StormTide> im trying to setup a mod_tile server for testing....
02:05:36 <StormTide> http://www.phpriot.com/2659 <-- trying that as the map
02:05:49 <StormTide> but all i get is white
02:06:47 <springmeyer> StormTide: http://bitbucket.org/springmeyer/tilelite/ <-- try that development server along with the sample OpenLayers file that zooms to your data extent automatically
02:07:08 <springmeyer> then if that does not work you may have a projection problem
02:07:15 <StormTide> i tried that first, but i couldnt figure out where 'easy install' comes from
02:07:31 <springmeyer> ah, you don't need easy install
02:07:42 <springmeyer> just download the sources
02:07:46 <springmeyer> cd source_dir
02:07:52 <springmeyer> sudo python setup.py install
02:07:53 <springmeyer> done
02:10:23 <StormTide> ok so tilelite runs...
02:10:53 <springmeyer> open up the utils/openlayers.html file
02:13:25 <StormTide> i can see the file, but for some reason i cant load it file the :8000 server
02:13:29 <StormTide> just 404's
02:13:46 <springmeyer> you are running the server?
02:13:52 <StormTide> ya..
02:13:56 <springmeyer> $ liteserv.py <mapfile.xml> ?
02:13:57 <StormTide> i can get to that demo page thing
02:13:59 <StormTide> yep
02:14:07 <StormTide> Welcome, ready to accept a tile request in the format of /zoom/x/y.png
02:14:10 <StormTide> that page works
02:14:25 <springmeyer> then the other should as well
02:14:56 <StormTide> localhost - - [01/Feb/2010 18:16:54] "GET /utils/openlayers.html HTTP/1.1" 404 25
02:15:11 <StormTide> from the foreground of the tilelite.py
02:15:20 *** tcarobruce has quit (Quit: tcarobruce)
02:15:27 <springmeyer> umm, just open the openlayers.html in a new browser window
02:15:37 <StormTide> oh, you mean from the filesystem
02:15:41 <springmeyer> yep
02:15:43 <StormTide> sorry, being dense
02:15:49 <springmeyer> no worries
02:16:38 <StormTide> yay
02:16:40 <StormTide> thats working
02:43:21 *** myselfhimself has parted #mapnik (None)
02:57:30 <springmeyer> StormTide is quite now :)
02:57:39 <springmeyer> er quiet I mean
02:59:24 <StormTide> heh
02:59:42 <StormTide> im trying to get route-me to load this osm source now
02:59:48 <StormTide> so onto another problem ;)
03:00:01 <springmeyer> hmm, via localhost?
03:00:10 <StormTide> via localnet...
03:00:20 <springmeyer> k
03:01:04 <StormTide> ive got a feeling theres a projection issue...
03:01:08 <StormTide> but need to confirm that
03:01:34 <springmeyer> well, did your tiles line up with the other layers in the OpenLayers demo html?
03:01:58 <springmeyer> you can turn on other layers with the blue + in the upper right
03:01:58 <StormTide> they do
03:02:42 <StormTide> i wish i could opacity = 50%
03:02:43 <StormTide> ;)
03:02:52 <Ldp__> you can
03:02:59 <springmeyer> yep, thats not hard
03:03:02 <StormTide> but the discrepencies might just be the differences in the mapping data
03:03:10 <springmeyer> Ldp__ will show you the way
03:03:16 <Ldp__> will I? oops
03:03:23 <springmeyer> (if Ldp__ were actually awake)
03:03:25 <StormTide> ;)
03:03:26 <springmeyer> but that is the Ldp__ bot
03:03:32 <Ldp__> the ill one
03:03:36 <springmeyer> oh no!
03:03:48 <springmeyer> flu or something?
03:03:51 <Ldp__> ya
03:03:59 <StormTide> springmeyer, you dont do any iphone dev do you haha
03:04:00 <springmeyer> ooooh, no fun
03:04:14 <springmeyer> ha ha
03:04:27 <Ldp__> I'm in one of the winding down stages: coughing your lungs out. No fun
03:04:34 <StormTide> Ldp__, oh i hate that
03:04:45 <springmeyer> damn. ya I was there a month ago
03:04:49 * Ldp__ loves codeine
03:04:50 <springmeyer> thoughts are with ya
03:04:55 <springmeyer> oh dear
03:05:18 <Ldp__> it's an antitussive
03:05:55 <Ldp__> where was I? oh yeah
03:06:01 <Ldp__> layername.setOpacity(0.5);
03:07:36 <StormTide> nono i meant between the google map that was showing in tilelite and my map
03:07:46 <StormTide> so i can see if they line up right
03:08:22 <Ldp__> doesn't matter where your tiles come from. If you're viewing them in an OL layer, you can do setOpacity
03:09:06 <StormTide> but when i click between em, one is hidden and the other show
03:09:07 <StormTide> n
03:09:17 <StormTide> (talkin about the blue + window in the top right here)
03:09:19 <Ldp__> make one of them an overlay
03:09:39 <Ldp__> isBaseLayer: false
03:10:37 <Ldp__> var yourlayer = new OpenLayers.Layer.XYZ("Name","http://server/tiles/${z}/${x}/${y}.png",{isBaseLayer:false,opacity:0.5});
03:23:42 *** Ldp__ has quit ()
03:39:49 *** jpablo (~chatzilla@190.159.196.162) has joined #mapnik
04:17:21 *** jpablo has quit (Quit: ChatZilla 0.9.85 [Iceweasel 3.0.14/2009091008])
04:28:06 <StormTide> springmeyer, is it normal for the tilelite thing to run very slowly at higher zoom levels while not taxing my computer?
04:29:21 <StormTide> trying to figure out if its this route-me app that is painfully slow, or if its some effect of this server
04:29:46 <StormTide> cuz, like, its working, but its 5-10 seconds per pannel
04:47:51 <springmeyer> hey StormTide
04:49:02 <springmeyer> yes, that is normal. the development server 'liteserv.py' uses Python's wsgiref impl, which is quite slow
04:49:39 <springmeyer> but if you actually deploy it, using mod_wsgi in multiprocess mode it will be *much* faster
04:50:02 <springmeyer> see the readme
04:58:26 <StormTide> gotcha
04:58:33 <StormTide> i plugged in my mod_tile server
04:58:37 <StormTide> its working flawlessly
04:59:39 <StormTide> ive got a proof of concept with a single mapfile up. so now, just a matter of writing it all up, and figuring out if there's a solution for making decent placemarks, etc
04:59:42 <springmeyer> awesome
05:19:08 <StormTide> know of any marker image packs with a permissive license?
05:29:06 <StormTide> http://www.facebook.com/photo.php?pid=5053801&l=ba54cc0797&id=568930020 ;)
05:41:57 <springmeyer> nice looks like progress
05:41:58 <springmeyer> http://code.google.com/p/google-maps-icons/
05:45:14 <StormTide> cool.
05:45:49 <StormTide> cc sa tho... i wonder whats considered a derived work on an icon set ;)
06:08:04 *** mperry has quit (Read error: Connection reset by peer)
06:08:08 *** mperry (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
06:39:27 *** HounD (~HounD@unics1.grfc.ru) has joined #mapnik
08:21:40 *** dodobas has quit (Ping timeout: 256 seconds)
08:22:10 *** dodobas (~dodobas@open.geof.hr) has joined #mapnik
08:32:49 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
08:32:50 *** mperry has quit (Read error: Connection reset by peer)
08:32:50 *** mperry_ is now known as mperry
09:16:42 *** mperry has quit (Read error: Connection reset by peer)
09:16:50 *** mperry (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
09:17:44 *** nikq (~nikq@li21-121.members.linode.com) has joined #mapnik
11:09:24 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
11:09:24 *** mperry has quit (Read error: Connection reset by peer)
11:09:25 *** mperry_ is now known as mperry
11:09:53 *** Ldp__ (~thid@osm.xs4all.nl) has joined #mapnik
11:25:17 <twain47> springmeyer: Always nice to see someone claiming credit for my work :(  'All icons made by Nicolas Mollet, unless specified.' grrrr.
11:59:56 *** D3f0 (~D3f0@www.transpa-sa.com.ar) has joined #mapnik
12:19:25 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
12:19:25 *** mperry has quit (Read error: Connection reset by peer)
12:19:26 *** mperry_ is now known as mperry
12:53:05 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
12:55:53 *** mperry has quit (Ping timeout: 260 seconds)
12:55:53 *** mperry_ is now known as mperry
13:43:53 *** hobu has quit (Excess Flood)
13:44:37 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
13:44:37 *** mperry has quit (Read error: Connection reset by peer)
13:44:38 *** mperry_ is now known as mperry
14:33:57 *** chad_burt (~chad_burt@mm-01.msi.ucsb.edu) has joined #mapnik
14:49:50 *** ajashton (~aj@c-69-136-229-112.hsd1.dc.comcast.net) has joined #mapnik
15:05:49 *** HounD has parted #mapnik (None)
15:09:08 *** ajturner_ (~ajturner@209.155.228.129) has joined #mapnik
15:13:04 *** ajturner has quit (Ping timeout: 265 seconds)
15:13:04 *** ajturner_ is now known as ajturner
15:28:57 *** xreal (~mweber@mailgw.FB12.Uni-Dortmund.DE) has joined #mapnik
15:29:48 <xreal> Sorry for asking here: What's the opposite of PostGIS-operator A && B (A completely contains B)? I want all the data of A, which isn't in B.
15:30:32 *** ajturner has quit (Remote host closed the connection)
15:30:50 *** ajturner (~ajturner@209.155.228.129) has joined #mapnik
15:31:46 <Ldp__> && is not 'A completely contains B'
15:31:58 <xreal> it is ;)
15:32:00 <Ldp__> && - Returns TRUE if A's bounding box overlaps B's.
15:32:06 <xreal> http://www.bostongis.com/postgis_quickguide.bqg
15:32:15 <xreal> "A && B (A and B bboxes intersect)"
15:32:21 <Ldp__> intersect is not contains
15:32:26 <Ldp__> http://postgis.refractions.net/documentation/manual-1.4/ST_Geometry_Overlap.html
15:33:00 <xreal> ah damn
15:33:04 <xreal> wrong line, sorry.
15:33:11 <xreal> You'
15:33:13 <xreal> 're right.
15:33:32 <Ldp__> try ST_Disjoint
15:34:05 <xreal> okay
15:34:11 <Ldp__> ST_Disjoint — Returns TRUE if the Geometries do not "spatially intersect" - if they do not share any space together.
15:34:32 * xreal wants something like "&& NOT(...)" ;-)
15:35:29 <Ldp__> WHERE not ST_Intersects(a, b)
15:36:23 <xreal> ST_Intersects uses indicies, doesn't it?
15:36:27 <xreal> indexes
15:36:32 <Ldp__> yes, and ST_Disjoint doesn't
15:36:42 <xreal> okay ;-)
15:37:10 <Ldp__> http://postgis.refractions.net/documentation/manual-1.4/ST_Intersects.html
15:37:13 <xreal> works fine, let me verify the output.
15:39:08 <xreal> 216077 rows, wow.
15:40:18 <xreal> YES, it worked. Thanks a lot!
15:40:36 <Ldp__> no problem
15:41:23 <Ldp__> if you have a large dataset, doing NOT ST_Intersects will return lotsa data
16:13:37 <springmeyer> twain47: grrr is right, I've not used that set just had an easy link.
16:13:49 * springmeyer looks back to see which he has stole
16:14:35 <twain47> springmeyer: There PD so technically he is not in breach of anything.  But claiming he 'made' it is maybe going a bit far :)
16:15:01 <twain47> It's only about 10 or 20 of them - lots are his own work
16:15:05 <springmeyer> are many yours just "wrapped" in those shields?
16:15:10 <twain47> yup
16:15:13 <springmeyer> ah, k
16:15:24 <springmeyer> yup, certainly see some simliarities
16:16:23 <twain47> see dentist (http://code.google.com/p/google-maps-icons/wiki/SportsIcons) for a really obvious example
16:17:47 <springmeyer> I see his credit: 'Many icons are derived from the project SJJBMapIcons by SJJB Management, licensed under Creative Commons Public Domain Dedication.'
16:18:12 <twain47> Added a few hours ago :) I left him a message.
16:18:26 <springmeyer> ha! nice
16:21:02 <Ldp__> twain47: the Ford icon I made for Haiti, would that be something you could easily add to your set?
16:21:34 <twain47> springmeyer: That's actually a lousy choice of icon - one of the ones that I took from someone else :)  I'm still building my own 'derived' list because I realised I was rather guilty of the same thing.
16:21:58 <Ldp__> twain47: I lifted the car model from your set
16:22:03 <springmeyer> :)
16:23:12 <twain47> Ldp__: Yup - Saw it on your map - been meaning to build a version to add.  Did you have an svg? No problem if not.
16:23:36 <Ldp__> I do have an svg, but it's just a wavy line under your car
16:24:17 <Ldp__> I actually wanted a car from the side, with a wavy line under it, but this had to do at short notice
16:24:25 <Ldp__> and it's probably better for low px sizes
16:25:17 <twain47> I'll do a quick drawing then and get it added.  I've got a bit of a back log of icon requests.
16:27:14 <Ldp__> here's my quick doodle: http://mijndev.openstreetmap.nl/~ldp/ford.svg
16:27:25 <Ldp__> initially had the top one, but that didn't work
16:27:57 <twain47> I liked it - was simple and obvious.  What more can you ask for!
16:28:05 <Ldp__> do you publish your request list somewhere?
16:28:46 <twain47> Not at the moment - looking at the google-maps-icons has made me aware how helpful that could be
16:30:31 <Ldp__> I'm about to add ford to the main map, can wait a bit for your version if needed
16:30:52 <twain47> Oh, in that case give me half an hour :)
16:31:10 <Ldp__> no rush :)
16:31:47 <Ldp__> although not rendered at such early zoom levels as is now done for Haiti
16:41:32 <StormTide> mornin
16:43:13 <StormTide> twain47, interesting... im trying to find a marine navigation set... like lighthouses, and cans/spars and such
16:45:43 <StormTide> http://wri-inter-dev.glf.dfo-mpo.ca/folios/00020/images/cansfr-eng.jpg and http://wri-inter-dev.glf.dfo-mpo.ca/folios/00020/images/cansbk-eng.jpg -> icon really
16:47:30 *** cgs_bob_ has quit (Ping timeout: 240 seconds)
16:51:35 <twain47> StormTide: link broken?
16:51:46 <StormTide> gr lovely
16:51:47 <StormTide> sec
16:52:35 <StormTide> http://www.ccg-gcc.gc.ca/eng/Ccg/atn_Buoyage_System <-- from the bottom of that page...
16:52:43 <StormTide> seem to work only with the refererer
16:55:03 <springmeyer> ajashton: tmcw has mentioned trouble installing PIL for the mapbox toolchain...
16:55:22 <springmeyer> is PIL used for anything in that workflow except Cascadenik?
16:56:06 <ajashton> does TileCache use it?
16:56:25 <twain47> StormTide: Probably a little specialist and country specific for the set I'm working on
16:56:30 <springmeyer> ah, ah thats right
16:56:35 <StormTide> twain47, its the international system
16:56:43 <StormTide> used everywhere ;)
16:57:00 <springmeyer> interestingly though TileCache's use of PIL is problematic
16:57:22 <springmeyer> an unneeded since Mapnik can do the metatile chopping itself
16:57:39 <springmeyer> which is the only case where PIL is used in TileCache AFAIK
16:57:55 <twain47> StormTide: oh. Sorry :) I'll add it to the list, but I'm not likely to get to something this specialist very soon
16:57:58 <StormTide> http://en.wikipedia.org/wiki/Lateral_mark <-- for example ;)
16:58:36 <StormTide> twain47, ya ;(... im trying to track down what open sea map is using
16:59:46 <ajashton> springmeyer: is PIL used for making JPG tiles, or does Mapnik do that as well?
16:59:59 <springmeyer> Mapnik can do JPG just fine
17:00:20 <springmeyer> although there is no ability to control JPEG quality in TileCache currently
17:00:37 <ajashton> right, I ran into that recently
17:01:01 <ajashton> found a couple lines of code on the mailing list to change it
17:01:10 <springmeyer> ah?
17:02:41 <ajashton> on the tilecache mailing list, looks like: http://n2.nabble.com/Tilecache-seed-and-jpeg-quality-td3500069.html
17:03:33 <springmeyer> ah, gocha, ya Guillaume is great
17:03:45 <springmeyer> thats slighty different though
17:04:11 <springmeyer> if you are using type=Mapnik then you can request Jpeg format natively
17:04:21 <springmeyer> and set the quality at the rendering stage
17:04:24 *** D3f0 has quit (Ping timeout: 252 seconds)
17:04:47 <springmeyer> which would be more optimal than requesting a full blown png from Mapnik, then reducing to a JPEG with PIL
17:14:02 <twain47> Ldp__: added transport_ford.  Pretty much as draw - I like not having to do much work :)
17:18:30 <Ldp__> heh excellent :)
17:19:04 <Ldp__> it's not really demanding to draw a wavy line, eh? :)
17:21:02 <twain47> Expecially since I copied yours :)  I tried the standard water (used in other icons) and a few variations but they just were not obvious enough, but it means this icon doesn't quite match.  I might go back and change some of the others like water tower to match this one instead
17:23:06 <Ldp__> I never liked poi_tower_water anyway
17:24:20 <Ldp__> twain47: do you have health_veterinarian on your todo list?
17:24:47 <Ldp__> oh wait, I just spotted health_veterinary. Strange, didn't notice it before
17:25:02 <twain47> It was added a couple of weeks back
17:25:49 <Ldp__> is there a chronological list of additions and modifications? I'm interested to get newer versions for osm.xml if these show up
17:26:16 <twain47> opticians, dentist and doctor were tweaked in the process to make them more balanced so you might want to update them too
17:26:40 <Ldp__> those aren't used yet
17:26:51 <twain47> doctors definitely is
17:27:52 <twain47> no it isn't - weard swear I rememberd that being there!
17:28:24 <Ldp__> only hospitals and pharmacies
17:29:19 <Ldp__> twain47: but a few shops were added recently
17:30:28 <twain47> really?  I've been waiting for supermarket and convenience for a while but they don't seem to have turned up
17:30:57 <Ldp__> they have, but Jon didn't expire tiles en masse due to possible Haiti load on yevaud
17:31:31 <twain47> I'm trying to sort out a SOURCES.txt and CHANGELOG.txt at the moment but keep putting if off :)
17:31:43 <Ldp__> but there's a problem on yevaud, causing only the shop labels to show up. It works fine on steve8's and my copies
17:33:09 <Ldp__> as on tile.openstreetmap.nl
17:33:14 <Ldp__> bbl
17:41:43 *** tcarobruce (~tcarobruc@adsl-75-10-247-30.dsl.pltn13.sbcglobal.net) has joined #mapnik
18:09:05 <tomhughes> does anybody have ubuntu/debian packages (or build script) for mapnik 0.7?
18:14:18 <springmeyer> hey tomhughes - I'm not aware of a package yet. I've pinged both of the maintainers for ubuntu and debian but both have been busy
18:14:39 <springmeyer> but source install notes for Karmic/0.7 are here: http://trac.mapnik.org/wiki/UbuntuInstallation
18:14:55 <tomhughes> a source install is the cause of my problem ;-)
18:15:03 <springmeyer> what is going wrong?
18:15:35 <tomhughes> I'm trying to build a mod_tile package, but the build scripts are finding mapnik 0.7 in /usr/local and complaing that they can't work out the dependencies as those files are from a package
18:15:52 <StormTide> tomhughes, i had to remove the packages and install from src
18:16:01 <StormTide> but managed to get it working on karmic
18:17:03 <springmeyer> hmm, can't work out the deps because the 0.7 files are *not* from a package you mean?
18:17:15 <tomhughes> yeah
18:17:20 <springmeyer> k
18:17:35 <springmeyer> and installing into /usr instead of /usr/local does not work?
18:18:26 * springmeyer would be interested to see the error that dpkg is throwing...
18:19:07 <tomhughes> dpkg-shlibdeps: failure: no dependency information found for /usr/local/lib/libmapnik.so.0.7 (used by debian/renderd/usr/bin/renderd).
18:19:45 <springmeyer> ah, okay. so it really wants a pkg, gocha
18:22:25 <StormTide> tomhughes, anyway, i never got the package approach to work... had to mapnik and mod_tile from source...
18:26:06 <Ldp__> StormTide: that's why I like gentoo :)
18:26:34 <StormTide> Ldp__, 99% of the time the package approach works nice
18:26:47 <StormTide> its just leading-edge software that runs into issues
18:36:01 <StormTide> springmeyer, got any tips about how to expand this server to multiple files?
18:36:31 <springmeyer> what server?
18:36:35 <StormTide> theres probably a better way than shoving all the shpfiles into one big xml
18:36:49 <StormTide> this mod_tile or whatever thing i got working last night
18:37:14 <StormTide> it looks at one... pretty manually prepared file... but i need it to look at potentially hundreds
18:39:29 <springmeyer> so you need to expose potentially hundreds of different looking base maps in your app?
18:39:41 <StormTide> springmeyer, different areas
18:39:53 <StormTide> eg same base map, different coverage area
18:40:08 <springmeyer> okay. well that can all be in one xml file then still
18:40:10 <StormTide> (coming from multiple s57's which are then split into shps)
18:40:17 <springmeyer> sure
18:40:48 <PhurlIpv4> "why boost::spirit sux? Try and debug" - http://tinyurl.com/ylcbrav
18:41:05 <PhurlIpv4> i find it almost impossible to debug mapnik
18:41:17 <StormTide> springmeyer, at what size will the xml file parsing become significant in the overhead
18:41:21 <PhurlIpv4> it is such a hard project to develop with, omg
18:41:52 <Ldp__> StormTide: mod_tile only parses the XML file(s) once
18:41:54 <springmeyer> StormTide: depends. ideally whatever rendering setup you use will only parse the XML once at server startup
18:42:05 <StormTide> oh, cool then
18:42:26 *** cgs_bob_ (~bob@7.sub-70-210-174.myvzw.com) has joined #mapnik
18:42:29 <StormTide> so now all i need to do is create a shellscript and figure out how to do this s57 layer to shp conversion on the cli
18:42:37 <StormTide> and i should be good to takeover the world ;)
18:42:41 <springmeyer> that will be easy
18:42:45 <springmeyer> ogr2ogr
18:43:28 <StormTide> does mapnik xml have any kind of preprocessor? eg for repeated elements?
18:43:38 <Ldp__> XML entities
18:44:03 <springmeyer> PhurlIpv4: don't like boost::spirit huh?
18:44:11 <StormTide> Ldp__, anything that could refer to a directory structure tho?
18:44:49 <springmeyer> PhurlIpv4: what are you trying to debug?
18:44:58 <PhurlIpv4> springmeyer, i am working on debugging my branch
18:45:02 <Ldp__> StormTide: do you need something like Path Expressions (see: http://www.mapnik.org/news/2009/dec/08/future_mapnik2/)
18:45:08 <PhurlIpv4> just stepping through
18:45:14 <PhurlIpv4> and find it very very difficutl
18:45:28 <springmeyer> have you read the boost::spirit docs?
18:45:50 <PhurlIpv4> n
18:45:51 <PhurlIpv4> o
18:46:09 <PhurlIpv4> but, noone can read 10 pages of object dumps
18:46:13 <StormTide> Ldp__, soooomething like that... essentially i have a dir full of folders that look like US4WA34M WS5OR15M ... etc
18:46:17 <PhurlIpv4> it is impossible
18:46:23 <StormTide> inside each i have the shpfiles i want to include in a common way
18:46:33 <StormTide> basically i want a directory foreach ...
18:46:43 <StormTide> i can script something in php to spit out the xml tho...
18:46:59 <StormTide> but it'll probably be processing slower than if there was a native processor macro
18:47:02 <Ldp__> StormTide: last time I had a bunch of shapefiles with the same features, I concatenated them together before giving them to mapnik
18:47:40 <Ldp__> ogr2ogr -append
18:47:55 <StormTide> so one mega mapfile is better than hundreds of smaller ones?
18:48:04 <StormTide> i would have thought the opposite
18:48:22 <Ldp__> remember to run shapeindex on the shapefile
18:48:43 <StormTide> shapeindex and me dont get along ;)
18:48:55 <springmeyer> PhurlIpv4: boost spirit just underwent a huge refactor and is now spirit2, mapnik only started using that new stuff in the last month so we have a lot to learn about spirit2
18:49:05 <StormTide> when i create an index, it crashes whenever i use the files
18:49:06 <PhurlIpv4> springmeyer, ok.
18:49:12 <PhurlIpv4> i hope to remove it all
18:49:18 <PhurlIpv4> from the main code base
18:49:22 <springmeyer> heh. well it was just added
18:49:23 <Ldp__> StormTide: you were using a POINT shapefile?
18:49:26 <PhurlIpv4> and separate it out
18:49:30 <PhurlIpv4> as a clean module
18:49:39 <springmeyer> PhurlIpv4: to reduce compile times?
18:49:42 <PhurlIpv4> there should be no need to have all the modules infected
18:49:50 <StormTide> Ldp__, springmeyer could explain the crash better than i could... but i think it was occuring on both the point and area ones
18:49:52 <PhurlIpv4> springmeyer, not only, but run time
18:50:06 <PhurlIpv4> you should be able to compile a style xml file into a c++ module
18:50:11 <PhurlIpv4> and link it in
18:50:21 <springmeyer> PhurlIpv4: not following...
18:50:23 <PhurlIpv4> without needed the interpreter
18:50:43 <PhurlIpv4> well imagine you separated out the language
18:50:44 <springmeyer> hmm, over my head I think
18:50:48 <springmeyer> k
18:50:49 <PhurlIpv4> into a simple program
18:50:53 <PhurlIpv4> that would generate c++ code
18:50:56 <springmeyer> boost spirit parsing?
18:51:00 <PhurlIpv4> yes
18:51:01 <PhurlIpv4> all parsing
18:51:01 <springmeyer> k
18:51:05 <springmeyer> okay
18:51:11 <springmeyer> so xml parsing too...
18:51:12 <PhurlIpv4> and then you had a set of callbacks
18:51:21 <PhurlIpv4> yes, well xml is already a plugin
18:51:28 <springmeyer> no its not
18:51:32 <PhurlIpv4> the osm is
18:51:38 <springmeyer> not the xml parsing for the styles
18:51:41 <PhurlIpv4> ok
18:51:47 <springmeyer> but yes, libxml2 is also used in the osm plugin
18:51:50 <PhurlIpv4> well the style system could be a plugin
18:51:54 <PhurlIpv4> basically
18:51:58 <springmeyer> k
18:52:17 <PhurlIpv4> that would be the best.
18:52:22 <springmeyer> by plugin do you mean just fully separate code as static or dynamic lib?
18:52:27 <PhurlIpv4> yeah
18:52:29 <springmeyer> k
18:52:34 <PhurlIpv4> full isolation
18:52:38 <springmeyer> sure
18:52:50 <PhurlIpv4> and you could replace the dynamic bit with a precompiled c++ module that does the same
18:52:57 <PhurlIpv4> same interface
18:53:00 <springmeyer> there has been talk of making the image_io all plugins too...
18:53:00 <PhurlIpv4> but faster
18:53:07 <PhurlIpv4> ok.
18:53:27 <PhurlIpv4> well mapnik uses only a small percent of the includes it has
18:53:46 <PhurlIpv4> i hope to be able to create an exact list of all dependencies
18:54:02 <PhurlIpv4> we can replace some of the classes with template wrappers
18:54:07 <springmeyer> per cpp file you mean?
18:54:21 <PhurlIpv4> in general, like alll the boost usages
18:54:30 <PhurlIpv4> if you have an exact list of the methods used
18:54:41 <PhurlIpv4> you can create a new c++ tempalte that has that signature
18:54:42 <springmeyer> okay
18:54:46 <PhurlIpv4> but hides all the stl stuff
18:54:51 <PhurlIpv4> that will compile in seconds
18:54:53 <PhurlIpv4> not minutes
18:55:17 <PhurlIpv4> in the implementation of the template, you can call boost
18:55:20 <Ldp__> even a mod_tile compile takes a long time with trunk :/
18:55:33 <PhurlIpv4> well it really makes development not fun
18:55:41 <PhurlIpv4> i cannot even use my computer while compiling
18:55:45 <PhurlIpv4> even if i use pch
18:55:50 <springmeyer> Ldp__: yes, it is all the new boost::spirit, compile times are really really slow
18:55:53 <PhurlIpv4> an have the compile time down to 20 seconds
18:56:19 <springmeyer> PhurlIpv4: yep, I agree that in trunk my computer becomes useless
18:56:33 <springmeyer> I've only just seen this because it, again, is new in trunk
18:56:51 <springmeyer> and previously (0.7 branch) we did not use spirit2
18:57:03 <PhurlIpv4> ok
18:57:22 <springmeyer> I think you are artem are likely the only people so far to hack on trunk
18:57:24 <PhurlIpv4> well, i finally have a running system
18:57:31 <springmeyer> other than my work this past week repairing things
18:57:31 <PhurlIpv4> i can compile and run
18:57:37 <PhurlIpv4> i am still on my branch...
18:57:39 <springmeyer> PhurlIpv4: congrats, does it pass the tests?
18:57:49 <PhurlIpv4> i had to break things to get it to compile...
18:57:59 <springmeyer> ruh roh :)
18:58:02 <PhurlIpv4> so now I am trying to bring it back to speed
18:58:10 <springmeyer> okay, right on
18:58:13 <PhurlIpv4> and have to find out where it is broken
18:58:26 <PhurlIpv4> moving code from header files to source files
18:58:31 <springmeyer> ah, I see more clearly now
18:58:40 <springmeyer> why you might be "debugging"
18:58:40 <PhurlIpv4> making template methods
18:58:53 <Ldp__> PhurlIpv4: there's one file during the compile which is particularly bad on compile times and memory usage: filter_factor. So if you're looking for somewhere to begin, that's a good one
18:59:00 <springmeyer> well, certainly take a look at the boost::spirit docs
18:59:07 <PhurlIpv4> ok
18:59:09 <springmeyer> and run the latest boost trunk
18:59:22 <springmeyer> because I just saw that they've made many changes
18:59:23 <PhurlIpv4> i have the latest boost
18:59:30 <PhurlIpv4> at least when i started
18:59:36 <springmeyer> Ldp__: exactly, filter_factor is the file that pulls in all the boost::spirit stuff
18:59:43 <PhurlIpv4> http://github.com/h4ck3rm1k3/MapNickAutotools/tree/master
18:59:48 <PhurlIpv4> you can see my stuff there
18:59:50 <springmeyer> k
18:59:57 <springmeyer> 404 Not Found
19:00:18 <PhurlIpv4> http://github.com/h4ck3rm1k3/MapNickAutotools
19:00:36 <springmeyer> github looks to be down
19:00:38 <PhurlIpv4> it is not perfect.
19:00:54 <PhurlIpv4> but i am making slow progress
19:03:00 <StormTide> springmeyer, this is probably impossible but, is it possible to use a shell command as a file source in a mapnik xml... eg to generate the shp files on-demand... that way i only have to maintain the .000 files at their newest version
19:03:37 <springmeyer> not in the xml, no
19:05:01 <StormTide> what if the file source was a shell script that emitted binary data? ... or would that just confuse it
19:05:25 *** mperry has quit (Read error: Connection reset by peer)
19:05:25 <StormTide> it would read it instead of executing wouldnt it
19:05:41 *** mperry (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
19:05:46 <StormTide> i bet there's a way with a fuse script
19:06:45 <springmeyer> sounds tricky
19:07:11 <springmeyer> StormTide: you can pretty easily create your XML mapfile with only <Styles>'s
19:07:23 <springmeyer> and then script in python the creation of the <Layers>
19:07:42 <springmeyer> then serialize all that back out to a compile XML mapfile with styles+layers
19:09:22 <StormTide> ya, i'll probably just php-emit the xml file... cuz thats really easy to do... but its more about having the shp files get out of sync with the .000 files.
19:09:27 <PhurlIpv4> http://forum.openstreetmap.org/viewtopic.php?id=5944 i get this error :Failed to parse filter expression:"name" in TextSymbolizer in style 'text'
19:09:33 <StormTide> will have to basically make a regeneration jobbie
19:09:39 <PhurlIpv4> does the easymapnik even run?
19:09:40 <StormTide> which will take a long time
19:09:51 <StormTide> whereas if i could do it on-demand...
19:09:55 <StormTide> whenever it wants new tiles..
19:09:57 <StormTide> might be better
19:11:10 <Ldp__> PhurlIpv4: I explained that on the forum, right?
19:11:17 <PhurlIpv4> i think so.
19:11:17 <springmeyer> on demand conversion of 000 files sounds really slow
19:11:34 <springmeyer> StormTide: how often do they change?
19:11:35 <PhurlIpv4> but it is not clear if it applies
19:11:42 <PhurlIpv4> and if the current code in trunk is broken
19:11:49 <PhurlIpv4> etc etc
19:12:10 <StormTide> weekly
19:12:23 <PhurlIpv4> is there a style that is known to work?
19:12:24 <StormTide> least thats the collection update frequency
19:12:31 <springmeyer> hmm, interesting
19:13:35 <springmeyer> PhurlIpv4: can't you just edit your style? it is a very minor change
19:13:49 <PhurlIpv4> springmeyer, of course! but this is the code i got.
19:13:57 <PhurlIpv4> i will scan for newer xml files
19:13:59 <springmeyer> what code?
19:14:23 <PhurlIpv4> mapnik/plugins/input/osm/test.xml
19:14:36 <PhurlIpv4> ./easymapnik -x ../test.xml -i ../test.osm -t -s osm -z 1 -Z 10
19:14:39 <StormTide> springmeyer, right now im exporting as a wgs84 shp file.... but since the eventual projection is 900913... is that dumb
19:15:05 <StormTide>     <Layer name="test3_label_overlay" srs="+init=epsg:4326" clear_label_cache="1">
19:15:06 <StormTide>  <-- is how i define the layers tho
19:15:10 <springmeyer> yep, transforming to 900913 at the shapefile level will lead to much faster rendering
19:15:13 <StormTide> <Map bgcolor="rgb(255,255,255)" srs="+proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs +over">
19:15:13 <StormTide>  
19:15:18 <StormTide> is the map
19:15:20 <springmeyer> ya, that'll be slow
19:15:25 <Ldp__> PhurlIpv4: I also wrote on the forum just now, but perhaps you didn't know yet that current trunk will read the old stylesheet name="name" construct again?
19:15:39 <StormTide> so i should go from .000 direct to the 900913...
19:15:40 <StormTide> k
19:15:47 <springmeyer> ya
19:15:56 <springmeyer> ogr2ogr -t_srs EPSG:900913 ...
19:16:12 <PhurlIpv4> Ldp__, i just want to know if this file is ok or not. mapnik/plugins/input/osm/test.xml does it need editing?
19:16:25 <PhurlIpv4> i mean is it my error or not.
19:17:04 <springmeyer> PhurlIpv4: yes, simplest thing is to edit that file
19:17:20 <PhurlIpv4> ok i will
19:17:21 <PhurlIpv4> !
19:17:24 <springmeyer> it is not your error, but rather and change in your version of trunk (that we have now modified to be backward compatible)
19:17:45 <PhurlIpv4> yes, so we need to keep these test files in synch with the branch
19:17:55 <PhurlIpv4> and have some type of automated test
19:17:59 <springmeyer> that will be the easy part :)
19:18:15 <springmeyer> PhurlIpv4: question for you...
19:18:21 <PhurlIpv4> that works /mapnik/tests/data/good_maps/bool_values.xml
19:18:26 <PhurlIpv4> springmeyer, ?
19:18:32 <springmeyer> PhurlIpv4: if you are working on Mapnik with speed in mind....
19:18:51 <springmeyer> then why are you looking at the OSM plugin?
19:19:09 <springmeyer> the OSM plugin is 1) slow, 2) rarely used, 3) experimental
19:19:15 <PhurlIpv4> springmeyer, well
19:19:21 <springmeyer> despite being quite cool and useful...
19:19:21 <PhurlIpv4> cause it can be made faster
19:19:33 <PhurlIpv4> i want to be able to use mapnik on small osm files
19:19:36 <springmeyer> okay, that is certainly the case
19:19:40 <PhurlIpv4> in a distributed manner
19:19:54 <PhurlIpv4> so, say, giving one time full of data at max zoon
19:19:56 <PhurlIpv4> zomm
19:20:00 <springmeyer> okay
19:20:03 <PhurlIpv4> so, say, giving one tile full of data at max zoom
19:20:13 <springmeyer> sounds good, just wanted to make sure I said that...
19:20:15 <springmeyer> plus...
19:20:16 <PhurlIpv4> we could render all parts of the different zoomlevels
19:20:23 <PhurlIpv4> that is my first goal
19:20:35 <PhurlIpv4> to create partial pngs
19:20:42 <PhurlIpv4> that can be merged in one pass
19:20:46 <springmeyer> the postgis plugin is what is fastest, and certainly it could use profiling too
19:20:49 <PhurlIpv4> yeah
19:20:54 <PhurlIpv4> but postgis is a big thing
19:21:00 <PhurlIpv4> i have some rtree code
19:21:04 <PhurlIpv4> and all in c++
19:21:17 <PhurlIpv4> better to have a monlith that can be optimized
19:21:21 <PhurlIpv4> monolith
19:21:34 <PhurlIpv4> i am thinking about clould computing
19:21:41 <PhurlIpv4> where the database storage is very slow
19:21:48 <PhurlIpv4> and better to have lots of small files
19:21:55 <springmeyer> PhurlIpv4: okay, sounds like you are having fun. keep it up!
19:22:00 <PhurlIpv4> vision of a git repository
19:22:09 <PhurlIpv4> with a directory struture
19:22:15 <PhurlIpv4> instead of postgis
19:22:29 <PhurlIpv4> could have a fully distribute rendering with only a filesystem
19:22:46 <PhurlIpv4> only updates would be expensive
19:23:29 <PhurlIpv4> if you sort the data properly ahead of time
19:23:39 <PhurlIpv4> you can get really good speed
19:23:49 <PhurlIpv4> at least that is my working idea
19:25:28 <PhurlIpv4>  /tests/data/good_maps/osm-styles.xml --- Failed to parse filter expression:"place_name" in TextSymbolizer in style 'places' is that broken as well?
19:25:56 <springmeyer> "place_name" --> "[place_name]"
19:26:04 <PhurlIpv4> ok, let me change ti
19:26:46 <PhurlIpv4>  <TextSymbolizer name="place_name" face_name="DejaVu Sans Book" size="10" fill="#444" halo_radius="1" wrap_width="0"/> ->  <TextSymbolizer name="[place_name]" face_name="DejaVu Sans Book" size="10" fill="#444" halo_radius="1" wrap_width="0"/> ?
19:28:01 <PhurlIpv4> ok that worked
19:28:02 <PhurlIpv4> great
19:28:11 <PhurlIpv4> thanks springmeyer that was not painful
19:30:17 <springmeyer> thank Ldp__ not me! :)
19:30:33 <PhurlIpv4> danke schön Ldp__
19:30:57 <StormTide> PhurlIpv4, the new expressions syntax is very powerful with a regexp replacement too btw
19:31:04 * StormTide found that out last week
19:31:08 <PhurlIpv4> StormTide, i believe you.
19:31:11 <Ldp__> I keep saying: I'm not German :)
19:31:17 <PhurlIpv4> ohhh!
19:31:27 <PhurlIpv4> well, i am not german either
19:31:59 <PhurlIpv4> Ldp__,  Dank U Wel!
19:32:09 <Ldp__> teehee :)
19:33:09 <PhurlIpv4> mapnik is very very powerful
19:33:17 <PhurlIpv4> it is like a programming language
19:33:45 <PhurlIpv4> that is why the style program needs to be factored out into a smaller module that can but used standalone and tested etc.
19:50:25 <springmeyer> mishok13: JSON styles, check
19:50:28 <springmeyer> http://www.boost.org/doc/libs/1_41_0/boost/property_tree/json_parser.hpp
19:50:48 <springmeyer> YAML styles, got a parser in mind?
19:51:18 <springmeyer> better, easier zoom syntax, lets do it, check
19:51:47 <springmeyer> styles and layers in different files, absolutely, can be done right now
19:52:14 <springmeyer> just requires two calls to load_map(), one to styles (enclosed in map), and the other to layers (enclosed in map)
19:52:47 <springmeyer> which is clumsy, I agree, so we should add more intuitive functions I think...
20:03:53 <Ldp__> springmeyer: keep in mind that I group style+layer in my include files
20:04:20 <springmeyer> yes, using XML entities anything is possible
20:04:44 * springmeyer was reading: http://cogniance.com/expertise/research_papers/designing-cartography-systems
20:05:31 <Ldp__> sigh, including a mapnik screenshot and then calling it osmarender :)
20:05:59 <springmeyer> I think mishok13 is getting at programs being able to load styles and layers separately, or at least that is something I'm interested in...
20:06:12 <springmeyer> ya, i thought that looked like the osm.xml too :)
20:06:26 <springmeyer> mishok13: get the GIMP out again!
20:08:16 <Ldp__> his zoom syntax con? entities again
20:09:33 <Ldp__> but XML entities are cumbersome, limited, and often hard to understand
20:09:53 <Ldp__> I'm all for a syntax that separates styling from heavy technical intricacies
20:15:24 <Ldp__> how old is mishok13's report?
20:21:12 <springmeyer> not sure
20:21:35 <Ldp__> "In short: each time you use XML for cartography, god kills a kitten."
20:21:38 <Ldp__> mew :-)
20:32:38 <springmeyer> meow
20:44:42 <springmeyer> .startrss
20:44:42 <nikq> Okay, I'll start rss fetching...
20:44:44 <nikq> Mapnik Trac: Ticket #508 (Image type uneeded in Mapnik 0.7.0 but still enforced in load_map) updated | http://trac.mapnik.org/ticket/508#comment:1
20:44:58 <springmeyer> Ldp__: #508 is bad news for osm.xml
20:44:58 <nikq> Ticket #508: Image type uneeded in Mapnik 0.7.0 but still enforced in load_map, http://trac.mapnik.org/ticket/508
20:45:34 <springmeyer> in other words it looks the the work making type,width,height optional is still buggy
20:45:36 <springmeyer> sigh
20:45:44 <Ldp__> springmeyer: too bad. I though artem would've addressed this while making the width/height fix
20:45:57 <springmeyer> well, he did
20:46:04 <Ldp__> not in load_map.cpp
20:46:07 <springmeyer> it just not working right
20:46:08 <springmeyer> right
20:46:31 <Ldp__> that's bad news for any distro that will only use tagged releases
20:46:33 <springmeyer> although hopefully I'm doing something stupid because the code looks fine
20:46:46 <springmeyer> Ldp__: yes
20:47:00 <Ldp__> will you consider updating release-0.7.0, as you did before?
20:47:02 <springmeyer> I think this may call for a 0.7.1
20:47:15 <springmeyer> No, I really don't want to mess with a tag
20:47:18 <Ldp__> just for this?
20:47:50 <springmeyer> but maybe...
20:48:00 <springmeyer> it would be wiser to release a 0.7.1 I think
20:48:07 <Ldp__> I presume you were prepared to issue fixes to 0.7.0 in a branch?
20:48:40 <springmeyer> I was but artem really wanted to delete the branch to avoid confusion and encourage the momentum to 0.8.0
20:49:03 <Ldp__> right, but that leaves us without recourse to have fixes to 0.7.0 in svn
20:49:14 <springmeyer> however, trunk is still a bit sketchy, so I may push back to quickly open up a new 0.7 branch
20:49:16 <springmeyer> yep
20:50:39 <Ldp__> any other pressing issues for a 0.7.1?
20:53:30 <springmeyer> I just added: http://trac.mapnik.org/milestone/0.7.1
20:54:35 <Ldp__> ok
21:03:15 <springmeyer> really not sure why its not working correctly, but I don't have time to dig into it now...
21:03:25 <springmeyer> will let you know if I do
21:04:14 <StormTide> springmeyer, did my post about the python expressions ever make it to the mailing list?
21:04:20 <StormTide> im not sure im properly subscribed
21:04:42 <StormTide> also for 0.7.1, wasnt there an xml bug
21:04:56 <springmeyer> StormTide: no, for some reason the dev list requires artem to accept
21:05:01 <springmeyer> which is must not have done yet
21:05:07 <springmeyer> so try sending to the mapnik-users list
21:05:12 <StormTide> eg why my qnik wont put out right xml anymore
21:05:15 <springmeyer> which I think allows auto-signup
21:05:45 <springmeyer> StormTide: the XML output problem *was* a trunk problem
21:05:49 <springmeyer> and I've now fixed it
21:05:57 <StormTide> ah
21:06:01 <springmeyer> so if you rebuild Mapnik trunk then Qnik should now
21:06:03 <StormTide> so it never affected the 0.7 rls
21:06:05 <StormTide> ;)
21:06:06 <springmeyer> output proper XML
21:06:08 <springmeyer> no
21:06:25 <springmeyer> if the XML is multi-line/pretty printed you should be good to go
21:07:16 <StormTide> its not but i'll rebuild here ;)
21:07:48 <StormTide> i think i managed to crash tilelite if you care about such things...
21:08:49 <springmeyer> ya, certainly
21:09:03 <springmeyer> since it would likely be a mapnik problem
21:09:58 <StormTide> http://www.phpriot.com/2663
21:22:41 <springmeyer> oh, that is harmless
21:23:04 <springmeyer> I've been meaning to catch that and muffle it
22:21:43 *** xreal has quit ()
22:36:18 <springmeyer> ajashton: around?
22:36:28 <ajashton> yep
22:37:04 <springmeyer> I've got a patch worked up for Cascadenik stuff
22:37:23 <springmeyer> interested in testing?
22:37:30 <ajashton> sure
22:37:50 <ajashton> what kind of stuff?
22:37:52 <springmeyer> cool, I think I'll commit then so you can pull down from svn.
22:38:31 <springmeyer> 1) maintaining relative paths, 2) saving url based resources without temp dir if using -d flag, 3) other bug fixes
22:39:06 <springmeyer> all tests are passing and the doc/*mml files run nicely
22:40:02 <ajashton> nice
22:40:50 <springmeyer> In the doc directory this test also works:
22:40:52 <springmeyer> for i in $(ls *.mml);
22:40:52 <springmeyer> do cascadenik-compile.py $i | nik2img.py $i.png; nik2img.py $i $i_.png;
22:40:52 <springmeyer> done
22:41:02 <springmeyer> I'll have you run that in a sec
22:43:03 <springmeyer> I think the the -d flag will still move files, need to look into that next
22:44:39 <springmeyer> k, ajashton svn up and give it a go
22:46:51 <CIA-31> mapnik-utils: dane.springmeyer * r900 /trunk/serverside/cascadenik/ (cascadenik/compile.py test.py):
22:46:51 <CIA-31> mapnik-utils: upgrade paths logic to support relative usage for more recent mapnik versions,
22:46:51 <CIA-31> mapnik-utils: fix usage of local files on windows, achange the behavior of the -d flag to
22:46:51 <CIA-31> mapnik-utils: avoid using tempfile for local files - closes issues #19,#26,#27,#28
22:46:52 <nikq> Ticket #19: 'Shaping' support, http://trac.mapnik.org/ticket/19
22:51:21 <ajashton> springmeyer: have to take care of some things right now.. will get back to you in 20-30 min
22:51:37 <springmeyer> ajashton: k, sounds good no worries
23:13:26 <StormTide> springmeyer, when i export the SLCONS layer as a shpfile... and then reload it, most of the lines disappear... is there a trick to this?
23:13:44 <StormTide> mapnik cant seem to render it either, but i suspect that might be because of a prob with the shp file
23:14:02 <springmeyer> I recommend using ogr2ogr
23:14:30 <springmeyer> then if there are problems with the input file (which there very likely may be) you'll know what they are and how to fix them
23:14:30 <StormTide> k lemme try that
23:19:19 <StormTide> -t_srs EPSG:WGS84
23:19:24 <StormTide> whats the right form of that
23:19:37 <Ldp__> EPSG:4326
23:21:21 <StormTide> ogr2ogr -t_srs EPSG:4326 -f "ESRI Shapefile" -select SLCONS ../43shpslcons.shp US5WA43M.000 ... results in Field 'SLCONS' not found in source layer. .. but there is that layer
23:21:45 <StormTide> (shows up in -summary too)
23:24:47 <StormTide> http://www.phpriot.com/2665
23:31:38 <springmeyer> the second command looks close
23:31:49 <springmeyer> I think you need to give ogr a hint on geometry type
23:31:59 <springmeyer> something like -nlt linestring
23:32:35 <springmeyer> http://www.gdal.org/ogr2ogr.html
23:33:02 <StormTide> k that gets rid of the error 1: items but the error 6 stuff remains
23:33:32 <StormTide> looks to be loadable this time tho
23:33:34 <StormTide> so that might do
23:34:26 <StormTide> except now mapnik is really confused at how to render that
23:34:41 <StormTide> anything special in style thats needed?
23:34:46 <StormTide> to indicate a linestring
23:35:12 *** drewby (~dfaubion@128-8-138-156.umd.edu) has joined #mapnik
23:35:19 <springmeyer> LineSymbolizer is all you need
23:35:25 <drewby> moment of truth!
23:35:28 <springmeyer> !
23:35:29 * drewby holds breath.
23:35:32 <springmeyer> !
23:36:07 <StormTide> haha
23:36:10 <StormTide> ;)
23:36:11 <StormTide> yay
23:36:17 <StormTide> you guys are awesome
23:36:51 <drewby> false alarm
23:37:59 * springmeyer has had a few of those when hacking in ctrans.hpp :)
23:38:42 * drewby can see why.
23:39:50 <springmeyer> ya, I find the whole item.vertext(), move_to graphics stuff tricky
23:40:24 <drewby> you know that w3c adopted that style of doing things for canvas
23:40:29 <drewby> could never figure out why
23:40:37 <springmeyer> but I sense its much faster to handle the coords in that workflow rather than what would be easier - just looping over indexed lists
23:40:48 <springmeyer> it seems to the way
23:41:09 <springmeyer> canvas is based on cairo, right? (... and cairo works that way)
23:41:58 <drewby> canvas is just a standard ( but it may be that mozilla/webkit implemented it in cairo) or did you mean based on In that they designed it with cairo in mind?
23:42:36 <springmeyer> ah, ya what you said :)
23:43:55 <StormTide> springmeyer, every so often, a tile at a specific zoom level generates really wrong
23:44:06 <StormTide> is there any way to detect/avoid that
23:44:12 <springmeyer> how are you serving the tiles right now?
23:44:16 <StormTide> mod_tile
23:44:30 <springmeyer> hmm, that should not happen
23:44:43 <springmeyer> apache pre-fork or worker?
23:44:47 <StormTide> prefork
23:44:54 <springmeyer> hmm
23:45:06 <StormTide> i can expose the test interface
23:45:07 <StormTide> sec
23:45:39 <drewby> StormTide: are you postprocessing the tiles in any way? I noticed that pngcrush would occasionally f-up a tile.
23:45:55 <StormTide> nope
23:46:04 <StormTide> lemme just put the test server up one min
23:47:50 <springmeyer> drewby: ah, good point
23:48:02 * drewby hates his test server.
23:48:07 <springmeyer> I was thinking about ways things could got really wrong
23:48:16 <springmeyer> but that is likely the problem
23:48:25 *** myselfhimself (~chatzilla@85-171-136-33.rev.numericable.fr) has joined #mapnik
23:48:26 <myselfhimself> hi
23:48:36 <drewby> Query run time on test server: 480 seconds, in productions: 8 seconds
23:48:45 <springmeyer> e.g. that Mapnik's png256 quant stuff is not retaining all needed colors
23:48:54 <myselfhimself> who could help with me support on render_expired.c code ?
23:49:02 <springmeyer> drewby: ouch :)
23:49:33 <springmeyer> StormTide: solution is  like to switch to 'png' format instead of 'png256' when in gen_tile.cpp
23:49:42 <StormTide> http://dev.stormtide.ca/test.htm can you see that?
23:50:31 <StormTide> the tiles for the first few zoom levels are all screwy
23:50:41 <StormTide> but they'll be right if i remove the bad tiles from cache and regenerate em
23:50:51 <StormTide> usually get one that craps out per full reset
23:52:49 <ajashton> springmeyer: been testing your changes. works well when using cascadenik-compile.py
23:52:53 <springmeyer> StormTide: I see some odd seams, is that what you mean?
23:52:59 <StormTide> ya
23:53:07 <StormTide> now watch
23:53:10 <springmeyer> not sure what is happening there
23:53:11 <StormTide> ill nuke the cache
23:53:15 <springmeyer> ajashton: great news
23:53:41 <springmeyer> ajashton: do you use cascadenik any other way that 'cascadenik-compile.py' ?
23:54:19 <StormTide> it'll probably render fine now
23:54:22 <StormTide> if you reload
23:54:25 <ajashton> springmeyer: yes, usually when testing I call liteserv.py on the mml. but now..
23:54:28 <StormTide> seems kinda random when the seems get fubar
23:54:53 <ajashton> liteserv seems to be working from /tmp, thus the relative paths do not work
23:55:49 <springmeyer> hmm, ya, okay, I'll take a look
23:56:52 <springmeyer> oh, ya TileLite needs to be updated
23:57:38 <springmeyer> ajashton: it will need to drop the compiled XML beside (in the same directory) as the .mml
23:57:55 <springmeyer> or thats the change I made to Quantumnik will testing this...
23:58:05 <springmeyer> do you think that's okay?
23:58:19 <ajashton> I don't have a problem with that
23:58:26 <springmeyer> okay
23:58:44 <springmeyer> just need to make sure the filename does not overwrite anything you might have locally in that directory
23:59:28 <springmeyer> I'm thinking the compiled file could be auto-named as the filename of the .mml but with a _compiled.xml at the end
23:59:33 <springmeyer> sound okay?
23:59:45 <ajashton> sure
23:59:47 <springmeyer> so, 'map.mml' --> 'map_compiled.xml' ?
23:59:57 <ajashton> makes sense