#mapnik log: Monday 09, March 2009

2009 | 03

previous | next
00:05:31 *** scruggs (n=chris@72-161-105-25.dyn.centurytel.net) has joined #mapnik
00:10:07 <nikq> Mapnik Trac: Ticket #256 (Syntax error in ogcserver.cgiserver from [986]) created | http://trac.mapnik.org/ticket/256
00:10:28 <nikq> Mapnik Trac: cgi_err.patch attached to Ticket #256 | http://trac.mapnik.org/attachment/ticket/256/cgi_err.patch
00:11:16 <rcoup> maybe mapniks needs python syntax checking in a pre-commit hook... ;)
00:15:04 <nikq> Mapnik Trac: Changeset [1003]: ogcserver: Fix syntax errors accidentally introduced in r986 - patch from  ... | http://trac.mapnik.org/changeset/1003
00:15:45 <nikq> Mapnik Trac: Ticket #256 (Syntax error in ogcserver.cgiserver from [986]) closed | http://trac.mapnik.org/ticket/256#comment:2
00:24:14 *** springmeyer (n=springme@c-24-19-50-92.hsd1.wa.comcast.net) has joined #mapnik
00:25:17 <springmeyer> thanks rcoup :)
00:27:14 <nikq> Mapnik Trac: Changeset [1004]: scons: fix brain fade - need to return gdal lib name to populate  ... | http://trac.mapnik.org/changeset/1004
00:29:01 <rcoup> springmeyer: next job for ya, fix --clean :P
00:30:12 <springmeyer> ya, that's odd.
00:31:18 <springmeyer> rcoup: a keyerror like that if the 'configure' dictionary that is pickled gets messed up
00:31:29 <springmeyer> (will happen)
00:31:39 <springmeyer> try $ rm config.cache
00:31:39 <rcoup> springmeyer: you can't do --clean without configure
00:31:46 <rcoup> not sure if thats a bug or not
00:33:38 <rcoup> i have some debian scripts working for trunk
00:33:46 <rcoup> updated with your scons changes, cairo, etc
00:34:44 * springmeyer sorry distracted...
00:35:01 <springmeyer> you should be able to to -c without configure
00:35:10 <springmeyer> if not yes its a bug
00:35:24 <rcoup> ok, you try it on a clean checkout ;0
00:35:40 <springmeyer> yes I will
00:35:44 <springmeyer> of svn head right?
00:35:47 <rcoup> yep
00:36:04 <rcoup> oh, and if you do something like scons --clean SOME_CONFIG=foo etc
00:36:11 <rcoup> it warns you about ignoring stuff
00:36:16 <rcoup> but then exits with code 0
00:36:23 <rcoup> (--clean could be build or install too)
00:36:33 <springmeyer> yes. we can alter than behavior if needed
00:36:33 <rcoup> and doesn't appear to do anything
00:37:12 <rcoup> it should probly either: exit with !0 or actually do what it was asked to (ignoring the config) ;)
00:37:27 <rcoup> not a biggie, just something i came across
00:38:48 <springmeyer> thanks.
00:39:35 * springmeyer has a friend leaving town, be back in a bit after getting a chance at a clean checkout...
00:40:30 *** scruggs_ (n=chris@72-161-105-25.dyn.centurytel.net) has joined #mapnik
01:40:26 <springmeyer> hey rcoup - back finally
01:40:59 <springmeyer> so you must be doing --clean on a fresh checkout, before actually building
01:42:31 <springmeyer> I do see the keyerror if I try that, but not if I --clean after doing a configure/build
01:43:22 <springmeyer> --clean is only able/meant to clean locally build targets so I don't know if there is a use case for cleaning before configuring
01:44:07 <rcoup> that's fine, i changed the debian script to do configure before clean
01:54:17 <springmeyer> well, the alternate is to not allow the bypass of the configure step when --clean is specified
01:55:04 <springmeyer> which is just a convenience anyway...
01:56:49 <springmeyer> regarding specifying SOME_CONFIG=foo...
01:57:55 <springmeyer> I've thought more about that, and now figure that if someone is specifying SOME_CONFIG=foo then they obviously want to 'configure' their build/clean/install
01:58:28 <springmeyer> so it makes sense to me to have the addition of any arguments kick off the configure stage, just to make sure that they are honored
01:59:06 <springmeyer> warning/exiting that they will be ignored seems kinda mean :)
01:59:37 *** rcoup has quit ()
02:45:53 *** Foriskak (n=default@c-67-164-57-179.hsd1.ca.comcast.net) has joined #mapnik
02:48:19 <Foriskak> I'm getting this: /usr/include/boost/python/detail/wrap_python.hpp:50:23: error: pyconfig.h: No such file or directory
02:48:49 <Foriskak> when I try to: install mapnik
02:49:00 <Foriskak> sudo python scons/scons.py install
02:49:18 <Foriskak> any ideas how pyconfig.h is supposed to come into being as it does not exist
02:49:21 <Foriskak> what does it do?
02:49:59 <springmeyer> Foriskak: pyconfig.h sounds like the development header for Python itself, that boost python needs to build against
02:50:11 <springmeyer> Foriskak: what os/platform?
02:50:24 <springmeyer> and what mapnik version?
02:50:25 <Foriskak> fedora core 9/64
02:50:49 <Foriskak> springmeyer: whatever is trunk in the SVN repo right now
02:50:56 <Foriskak> I checked out last night
02:50:57 <springmeyer> okay
02:52:14 <springmeyer> so make sure you've got python development headers
02:54:33 <springmeyer> Foriskak: would be helpful to see your full compile output as well
02:54:47 <springmeyer> http://dpaste.com works well for pasting something that large
02:54:51 <Foriskak> yum install python-devel
02:54:53 <Foriskak> sure
02:55:01 <springmeyer> yes, that sounds right
02:58:04 <Foriskak> http://dpaste.com/10291/
02:59:37 <springmeyer> Foriskak: okay, I bet you have python-devel
03:00:18 <springmeyer> but Mapnik's scons scripts are failing to properly attach it to the compile args
03:00:27 <springmeyer> can you do
03:00:38 <springmeyer> $ python scons/scons.py configure
03:00:48 <springmeyer> and dpaste.com the output of that for me?
03:04:29 <Foriskak> ok
03:05:13 <Foriskak> http://dpaste.com/10296/
03:05:19 <springmeyer> thx
03:09:32 <springmeyer> hmm, okay no clues there.
03:09:40 <springmeyer> what about the build output ?
03:09:47 <springmeyer> $ python scons/scons.py
03:11:02 <Foriskak> hmm
03:11:06 <Foriskak> ok
03:11:18 <springmeyer> also, would be good to know what this gives (in a python interpreter)
03:11:21 <springmeyer> >>> import distutils.sysconfig
03:11:31 <springmeyer> >>> distutils.sysconfig.get_python_inc()
03:12:25 <springmeyer> the python include directory should hold 'pyconfig.h'
03:12:34 <Foriskak> http://dpaste.com/10300/
03:12:45 <springmeyer> that command is used internally in the scons scripts, so if it is failing that may be the problem
03:12:48 * springmeyer takes a look
03:12:49 <springmeyer> thx
03:12:59 <Foriskak> that file is nowhere to be found, let me do the python checks
03:13:32 <springmeyer> $ locate pyconfig.h ?
03:13:38 <Foriskak> yeah
03:14:12 <springmeyer> k
03:14:32 <Foriskak> The include dir is /usr/include/python2.5 definitely does not hold that file
03:15:05 <Foriskak> brb coffee
03:15:06 <springmeyer> interesting, okay
03:15:13 <springmeyer> it *should*
03:15:25 <springmeyer> so I guess you could be lacking python-devel
03:15:45 <springmeyer> not sure how boost and other stuff could have been installed without it...
03:16:37 <springmeyer> ya, I get it there on my ubuntu machine
03:16:38 <springmeyer>  ls /usr/include/python2.5 | grep pyconf
03:16:38 <springmeyer> pyconfig.h
03:16:52 <springmeyer> so install python-devel from yum and that should do it
03:16:54 <Foriskak> I think I recall somethign now
03:16:59 <Foriskak> that failed
03:17:09 <Foriskak> ah...now I recall
03:17:13 <springmeyer> but mapnik should be checking for pyconfig itself I guess
03:17:15 <springmeyer> ah, okay
03:17:47 <Foriskak> ok...now I recall, a bunch of docs conflict with some kde-utils crap
03:18:37 <springmeyer> huh. sounds fun
03:18:54 <Foriskak> this looks like it's getting further
03:20:18 <Foriskak> so about the DB choice of postgres over mysql (I notice osmosis can import into mysql), that was made due to the GIS package?
03:21:15 <springmeyer> well, postgres/postgis has excellent spatial indexing and functions so it is a great option for storing geodata
03:21:37 <springmeyer> but as far as mapnik, it can read from any datasource via a plugin architecture
03:22:01 <springmeyer> so a mysql plugin could be written but nobody has found the need yet
03:22:23 <springmeyer> postgres/postgis, sqlite, and oracle are the DB plugins written so far
03:24:01 <Foriskak> I see. I gather however that DB queries can be avoided when serving up slippy maps?
03:30:11 <Foriskak> It appears from the fragmented documentation that once you render the tiles, one can calc the tilename (or apply some hash if directories have too many files) with a simple function and then just get the tile that way without any DB interaction
03:35:50 <springmeyer> yes, tiles can be cached
03:36:27 <springmeyer> there are a variety of tools to handle tile serving and caching logic when using mapnik as the renderer
03:36:41 <Foriskak> I'm considering mod_tile
03:36:45 <Foriskak> anythign else of note?
03:37:22 <springmeyer> yes, that is the optimal choice
03:37:45 <springmeyer> as long as you want to serve tiles in the OSM scheme
03:53:44 <Foriskak> nearly there...how do I set python's include dirs? so it finds libmapnik.so
03:59:33 <springmeyer> you mean so that >>> import mapnik
03:59:39 <springmeyer> ... finds libmapnik.so?
04:00:59 <springmeyer> put '/usr/local/lib64/' in one of the files in your /etc/ld.so.conf
04:01:06 <springmeyer> and then run $ldconfig
04:02:25 <springmeyer> if you are running into more compile problems paste the errors and I'll take a look in a bit
04:02:29 * springmeyer heads off to dinner
04:13:19 <Foriskak> springmeyer: did that, hmm, let me double check
04:36:07 *** weizhuo (n=chatzill@nat/yahoo/x-51512f5ba93e1d51) has joined #mapnik
04:45:10 <nikq> Mapnik Trac: Changeset [1005]: scons: force configure stage if the user supplies commandline arguments,  ... | http://trac.mapnik.org/changeset/1005
04:54:08 <springmeyer> Foriskak: any luck?
04:54:26 <springmeyer> and so installing python-devel got your build working?
04:59:33 <nikq> Mapnik Trac: Changeset [1006]: scons: avoid skipping configure if --clean is requested since it will bork  ... | http://trac.mapnik.org/changeset/1006
05:00:54 <nikq> Mapnik Trac: Changeset [1007]: scons: exit with proper return code indicating an error if we are missing  ... | http://trac.mapnik.org/changeset/1007
05:09:51 <nikq> Mapnik Trac: MapnikInstallation edited | http://trac.mapnik.org/wiki/MapnikInstallation?version=3
05:25:14 <nikq> Mapnik Trac: UsingScons edited | http://trac.mapnik.org/wiki/UsingScons?version=12
05:26:25 <Foriskak> springmeyer: the build yes, but python can't find the libmapnik somehow
05:26:56 <Foriskak> despite stuff in /etc/ld.so.conf.d/
05:27:18 <Foriskak> springmeyer: so, the way that directory works it seems is that dirs in all files in there are searched
05:27:46 <nikq> Mapnik Trac: Ticket #257 (Improve Python checking/configure) created | http://trac.mapnik.org/ticket/257
05:27:48 <Foriskak> is the name of the file containing the dirs of any importance maybe?
05:28:51 <springmeyer> hey Foriskak
05:28:56 <Foriskak> hey
05:29:04 <springmeyer> glad the build worked, sorry the lib is a hassle!
05:29:08 <springmeyer> can you do this...?
05:29:09 <springmeyer> >>> from mapnik.paths import mapniklibpath
05:29:09 <springmeyer> >>> print mapniklibpath
05:29:44 <Foriskak> ImportError: libmapnik.so.0.6: cannot open shared object file: No such file or directory
05:30:04 <springmeyer> okay, worth a try :)
05:31:35 <Foriskak> ah...there seems to be an enviro var called PYTHONPATH
05:32:08 <Foriskak> heh... INSTALL file in checkout has necessary steps not mentioned in the online docs
05:32:10 <Foriskak> hah
05:32:18 <Foriskak> well that woudl have saved some time
05:33:19 <Foriskak> I'm sure a seasoned python person would now how to fix this fast, too bad the code's not in Perl
05:33:35 <springmeyer> Foriskak: did you fix it?
05:33:52 <springmeyer> PYTHONPATH won't fix you ability to find the lib
05:34:11 <Foriskak> really
05:34:28 <springmeyer> nope
05:34:33 <springmeyer> if you are getting 'ImportError: libmapnik.so.0.6'
05:34:50 <springmeyer> that means that the mapnik python bindings are 'on the PYTHONPATH'
05:35:00 <springmeyer> so the bindings themselves are found
05:35:04 <Foriskak> ok
05:35:15 <springmeyer> if they were not then you would get 'ImportError: No module named mapnik'
05:35:26 <springmeyer> as per the INSTALL doc
05:35:48 <springmeyer> so Foriskak: where did libmapnik.so get installed?
05:36:32 <Foriskak> In /usr/local/lib64
05:36:37 <springmeyer> can you confirm that  you have '/usr/local/lib64/libmapnik.so'?
05:36:41 <springmeyer> okay
05:37:01 * springmeyer boots up an ubuntu64 instance to see what his ld.so.conf looks like...
05:37:26 <Foriskak> include ld.so.conf.d/*.conf
05:37:44 <Foriskak> ahaha
05:37:45 <Foriskak> moment
05:37:50 <springmeyer> right, which reads in one of the *.conf files
05:37:54 <springmeyer> http://trac.mapnik.org/wiki/InstallationTroubleshooting#LinkingErrors
05:40:28 <Foriskak> excellent
05:41:26 <springmeyer> works?
05:46:05 <Foriskak> yes
05:46:20 <Foriskak> now to get postgres to accept my user
05:47:01 <springmeyer> cool
05:47:17 <springmeyer> so what were we missing in the docs to help make that linking easier?
05:48:18 <springmeyer> re: postres user I used edit the pg_hba.conf to 'trust' all local connections
05:49:22 <springmeyer> and then turn off listening for anything but local connections in postgresql.conf
05:49:38 <Foriskak> springmeyer: is the build step if scons.py necessary?
05:50:11 <springmeyer> Foriskak: I'm not sure what you mean
05:50:18 <springmeyer> explain more?
05:50:48 <Foriskak> $ python scons/scons.py
05:51:01 <Foriskak> this step before sudo python scons/scons.py install
05:51:10 <springmeyer> oh, okay
05:51:18 <springmeyer> no, its optional
05:51:21 <springmeyer> but a good idea
05:51:49 <springmeyer> in other works you could do a clean checkout and the simply do...
05:51:50 <springmeyer> sudo python scons/scons.py install
05:52:19 <springmeyer> but then you risk installing a partially build mapnik if something goes wrong
05:52:50 <springmeyer> and your local compiled targets will be owned by root which can be odd
05:53:21 <Foriskak> maybe mention that, this was mostly due to my ignorance of ld.so.conf.d's functioning
05:53:51 <Foriskak> I've been copying and pasting all the lines I've had to type since beginning this
05:54:08 <springmeyer> well ld.so.conf can be a bit obscure :)
05:54:31 <springmeyer> copying and pasting?
05:54:34 <Foriskak> getting intarray working, calling osm2pgsql so that it doesn't fail,etc.
05:54:41 <springmeyer> oh, because there is no guide for fedora huh?
05:54:59 <Foriskak> or so I don't forget what I had to do if I have to do it again!
05:55:07 <Foriskak> there is one by passback
05:55:28 <springmeyer> a new page on fedora could be your next paste :)
05:55:29 <springmeyer> http://trac.mapnik.org/wiki/LinuxInstallation
05:55:36 <Foriskak> heh
05:55:45 <Foriskak> sure
06:08:03 *** rcoup (n=rcoup@ip-118-90-3-165.xdsl.xnet.co.nz) has joined #mapnik
06:13:33 <springmeyer> hey rcoup: I updated scons based on your comments. let me know if you have any further thoughts
06:13:40 <springmeyer> r1005 and r1006
06:13:42 <nikq> http://trac.mapnik.org/changeset/1006, at , by dane: scons: avoid skipping configure if --clean is requested since it will bork scons given the need for the configured/pickled dictionary to populate env variables in Sconscript files - thanks rcoup for n
06:14:08 <rcoup> coolio
06:14:13 <rcoup> will have a look shortly
06:14:14 <rcoup> thanks :)
06:14:25 <springmeyer> naw thank you, for the review
06:14:35 * springmeyer is headed off now
06:14:44 <springmeyer> catch you all later
06:16:23 *** springmeyer has quit ()
06:42:11 *** weizhuo has quit ("ChatZilla 0.9.84 [Firefox 2.0.0.20/2008121709]")
07:04:55 <dodobas> yello
07:47:01 *** xcacou (n=aga@AToulouse-157-1-70-225.w81-50.abo.wanadoo.fr) has joined #mapnik
07:51:00 *** __d3f0__ (n=defo@190.176.196.131) has joined #mapnik
07:58:58 *** D3f0 has quit (Read error: 145 (Connection timed out))
08:12:28 *** d3f0 (n=defo@190.176.195.185) has joined #mapnik
08:13:30 *** d3f0 has quit (Read error: 104 (Connection reset by peer))
08:16:49 *** d3f0 (n=defo@190.176.194.219) has joined #mapnik
08:19:01 *** __d3f0__ has quit (Read error: 104 (Connection reset by peer))
08:22:53 *** d3f0 has quit (Remote closed the connection)
08:23:08 *** d3f0 (n=defo@190.176.194.219) has joined #mapnik
09:10:41 *** d3f0 has quit (Read error: 110 (Connection timed out))
10:06:40 *** xcacou has quit (Read error: 113 (No route to host))
10:21:44 *** xcacou (n=aga@AToulouse-157-1-70-225.w81-50.abo.wanadoo.fr) has joined #mapnik
10:22:22 *** rcoup has quit ()
10:27:32 *** rcoup (n=rcoup@ip-118-90-3-165.xdsl.xnet.co.nz) has joined #mapnik
10:31:28 *** rcoup has quit (Client Quit)
12:13:08 *** aub (n=aubrey@cpe-72-227-134-148.nyc.res.rr.com) has joined #mapnik
12:25:10 *** reallove (n=dan@unaffiliated/reallove) has joined #mapnik
12:26:12 <reallove> hello,anyone around ? I got a problem rendering with mapnik . Following the instructions from http://wiki.openstreetmap.org/wiki/Mapnik , when I run any of the generate python scripts I got errors
12:26:45 <reallove> RuntimeError: failed opening file: Bad file descriptor ; at a line that in the .py file is : load_map(m,mapfile)
12:27:06 <reallove> if I echo $MAPNIK_MAP_FILE , I got the correct osm.xml file
12:28:19 *** aub has quit ()
13:34:48 *** aub (n=aubrey@216.156.96.6) has joined #mapnik
13:35:25 *** reallove has parted #mapnik ()
14:07:44 *** aub_ (n=aubrey@216.156.96.6) has joined #mapnik
14:07:47 *** aub_ has quit (Client Quit)
14:10:38 *** aub has quit (Read error: 60 (Operation timed out))
14:14:20 *** aub (n=aubrey@216.156.96.6) has joined #mapnik
14:44:24 *** xcacou has quit (Remote closed the connection)
15:40:55 *** sanjiv (n=chatzill@59.180.145.180) has joined #mapnik
15:44:04 *** sanjiv has quit (Remote closed the connection)
15:44:51 *** sanjiv (n=chatzill@59.180.145.180) has joined #mapnik
15:47:24 *** sanjiv has quit (Read error: 104 (Connection reset by peer))
15:48:15 *** sanjiv (n=chatzill@59.180.145.180) has joined #mapnik
15:51:28 *** sanjiv has quit (Remote closed the connection)
16:26:54 *** tomhughe` (n=tom@gate.compton.nu) has joined #mapnik
16:27:59 *** tomhughe` is now known as tomhughes1
16:28:24 *** tomhughes has quit (Read error: 101 (Network is unreachable))
18:18:37 *** rcoup (n=rcoup@ip-118-90-33-69.xdsl.xnet.co.nz) has joined #mapnik
20:53:18 *** audifahrer (n=andreas@p57AF4DC6.dip.t-dialin.net) has joined #mapnik
21:05:15 *** rcoup_ (n=rcoup@ip-118-90-79-181.xdsl.xnet.co.nz) has joined #mapnik
21:14:51 *** rcoup has quit (Read error: 110 (Connection timed out))
21:52:11 <audifahrer> hello
21:53:10 <audifahrer> is someone here to help me with a XML style question?
22:18:08 *** weizhuo (n=chatzill@nat/yahoo/x-0236ea412061fcda) has joined #mapnik
22:35:21 *** audifahrer has quit ("Verlassend")
23:26:02 *** aub has quit ()