#MAPSERVER IRC Log - 2010-01-27

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
02:14:01 kondos: He guys i just upgraded to ubuntu 9.10 karmic, and i want to install mapsever using the repository. Can anyone help me with the link to the guide to do so
02:25:08 dreamil: kondos: you can apt-get it using :
02:25:24 dreamil: sudo apt-get install mapserver-bin cgi-mapserver
02:25:32 dreamil: depending on what you want to install
02:25:40 dreamil: also: php5-mapscript
02:27:02 kondos: dreamil: on 9.04 there were some two lines we use to add in /etc/apt/sources.list and now that we using 9.10 i dont know how they should look like
02:27:22 dreamil: kondos: which version you want to install?
02:28:11 kondos: any version its just that am using ubuntu 9.10
02:28:32 dreamil: 9.10 seems to have 5.4.2
02:28:45 dreamil: sudo apt-get install mapserver-bin cgi-mapserver
02:28:50 dreamil: ^^^ try this
02:30:20 kondos: in 9.10 there is no need to edit the sources.list file
02:30:41 dreamil: yes, the packages are there in default repos
02:34:17 kondos: Okay thanks a lot i was strugling. i just installed it is working and nowi can carry on
02:34:35 dreamil: kondos: glad that i could be of help
02:52:12 aliceinwire: http://pastebin.com/m13c7a5d6
02:52:13 sigabrt: Title: pastebin - collaborative debugging tool ( at pastebin.com )
02:52:37 aliceinwire: how i can adjust this ?
03:00:02 chippy: hi, is it possible to selectively turn on and off rasters within a tileindex shapefile using mapscript?
03:01:38 : * chippy is assuming not
03:02:46 ue: chippy, still here ?
03:02:58 chippy: yes.
03:03:19 ue: 2 options
03:03:35 ue: 1 ) Regenerate the shapefile
03:03:45 ue: 2 ) As I recall index could be read from postgis
03:03:56 ue: so you can form a select that suits your needs
03:04:23 chippy: oh, thanks ue, yes using postgis that could work
03:05:36 ue: as I remember correctly it works in the following way: You create an "vector" layer that represents your "tileindex" grid
03:05:51 ue: and the in your raster you specify the layer name
03:06:02 ue: something like this, better check the doc first
03:07:30 chippy: many thanks ue
03:07:42 chippy: Not sure if it's something we want, but good to know it's possible : )
03:08:16 ue: np
03:10:49 aliceinwire: how i have to set this conditional for make it work ?
03:10:55 aliceinwire: http://pastebin.com/m13c7a5d6
03:10:56 sigabrt: Title: pastebin - collaborative debugging tool ( at pastebin.com )
06:34:01 jmckenna: trying to build 1.7.0 on MSVC 2008. fails on frmts\nitf. http://pastebin.ca/1768042 As the error states, the directory jpeg/libjpeg12/ does not contain the file jpeglib.h
06:34:03 sigabrt: Title: pastebin - MSVC compile error - post number 1768042 ( at pastebin.ca )
06:34:27 jmckenna: doh wrong window again
09:21:58 CIA-6: assefa * r9763 /trunk/mapserver/mapogcsld.c: Fix possible crash with regular expression parsing #2720
10:20:57 CIA-6: aboudreault * r9764 /trunk/mapserver/ ( HISTORY.TXT mapshape.c ): Added support to write a null shape in a shape file. ( #3277 )
10:58:26 Flewellyn: HI folks.
10:58:45 Flewellyn: I know Mapserver is not thread safe. Is it multi-process safe?
10:58:47 Flewellyn: As in, fork( )?
11:00:12 Flewellyn: I want to fork multiple mapscript-using processes, to do work in parallel.
11:10:55 winkey: Flewellyn every time it runs its fork( )ed thats how apps start on *nix
11:11:03 Flewellyn: winkey: Yes, I know.
11:11:26 Flewellyn: What I'm wondering is, can I fork a mapscript-using process and have it work without issues?
11:11:41 Flewellyn: Or will the non-threadsafe aspects of Mapserver cause problems there, as well?
11:16:14 winkey: when something is forked all memory is copyed to a new place and it continues on at that point in both proccesses, after that any changes to that memory are not shared between the two, you actualy have to code in some form of ipc to comunicate between the two. if it were to write to the same file you could have a conflict but then that would be an issue anyway since a webserver would be likely to start it more than once.
11:17:59 hobu: Flewellyn: MapForker ( tm ) works fine as an alternative to threading
11:18:25 Flewellyn: Mapforker? Heh.
11:18:39 Flewellyn: winkey: That's encouraging to hear.
11:19:11 winkey: that wasn't quite textbook, but close enough i think
11:19:16 : * Flewellyn googles for MapForker, finds only one link.
11:19:42 Flewellyn: hobu: I don't want threading because of the insanity it induces. I'm forking my own mapscript processes.
11:19:48 hobu: yes, me http://lists.osgeo.org/pipermail/mapserver-dev/2005-September/000959.html ; )
11:19:49 sigabrt: Title: Help: Memory Leaks ( at lists.osgeo.org )
11:20:28 Flewellyn: Indeed.
11:20:30 Flewellyn: I am there right now.
11:20:49 : * hobu jokes and notes that Flewellyn will have to build his own MapForker ( tm ) but license the ( tm ) from him ; )
11:20:59 : * Flewellyn girns.
11:21:07 Flewellyn: So basically, forking my own mapscript processes.
11:21:15 Flewellyn: Okay!
11:21:19 Flewellyn: That is really good to know.
11:21:27 : * Flewellyn is not a fan of threads anyway, so...
11:21:28 hobu: yep, fork right before you do the draw
11:21:38 hobu: and setup IPC to get back data from the children
11:22:04 Flewellyn: Makes sense. In this case, I'd be generating separate maps anyway, and writing them to temp files.
11:22:05 winkey: hey hobu know anything about dvidk's kml stuff in his sandbox?
11:22:13 winkey: davidk
11:22:37 Flewellyn: THanks for your advice, winkey and hobu.
11:22:38 hobu: all i know is it is most likely a failed SoC project, but nothing else. danmo was his mentor, and may have more information if he has time to dig it up
11:22:56 : * Flewellyn will get back to hacking.
11:24:17 winkey: hobu failded because it was never merged into trunk?
11:25:03 hobu: i'm sure there are plenty of reasons and properties to call it a failure
11:27:38 winkey: well that i know of its one of 2 pieces of code that could write out ogr featurestyle to kml, the other is mine, therefore i'm interested
11:55:06 nhv: danmo winkey had some questions about davidk's SOC sandbox
11:55:50 danmo: um... I can try to help, but I think Assefa or tbonfort would be better to help, once they are online
11:56:30 nhv: k thanks will pass that on
12:56:46 chippy: Hi, is there a way to exclude a layer in a mapfile from showing up in a getCapabilities request ( but still potentially to be able to be requested by a getMap ) ?
12:58:50 aboudreault: unfortunately no.
12:59:01 chippy: aboudreault, okay, thanks : )
13:04:12 aboudreault: It has had some discussions to add this support but it's not yet implemented.
13:04:38 aboudreault: and I don't find the wiki page that talk about it... :/
13:04:46 aboudreault: danmo: ping
13:05:15 danmo: aboudreault: ack
13:05:54 danmo: I believe there is a ticket about this ( and more layer control stuff )
13:06:11 danmo: but it quickly gets complicated, that's why nothing has been done yet
13:07:38 chippy: Fair enough, I can live without it : )
13:08:55 danmo: aboudreault: you were thinking of http://trac.osgeo.org/mapserver/wiki/HidingLayersInOGCWebServices
13:08:56 sigabrt: Title: HidingLayersInOGCWebServices - MapServer - Trac ( at trac.osgeo.org )
13:09:16 aboudreault: danmo: exactly, thanks
13:09:51 aboudreault: chippy: If by any chance you can/want to add useful information about this feature, feel free : )
13:11:51 chippy: okay. My use case is that a couple of raster layers were mis-registered, and so I wanted to hide these from the list until they were fixed, whilst still being able for some clients to be able to view them directly.
13:12:44 chippy: But I'm putting that information in the abstract.
13:13:23 danmo: I added a note about this issue to our TODO list for the code sprint: http://wiki.osgeo.org/wiki/New_York_Code_Sprint_2010_Agenda#MapServer
13:13:26 sigabrt: Title: New York Code Sprint 2010 Agenda - OSGeo Wiki ( at wiki.osgeo.org )
15:02:30 jlivni: q
15:11:30 nhv: chippy you could put a proxy in front of your mapserv application that substituted the mapfile parm if it was a getcap request
23:24:10 leo-unglaub_work: hi friends, i have a very funny problem. i have got 2000 .ecw files from our government for a project but the files are in a false projection. so we have downloadet the tool ecw header editor and chanes the projection manualy by hand. it works but i can't do this for 2000 images. du you have any idea how i can change ecw headers by a batch?
23:33:27 racicot: leo-unglaub_work: not sure if you have gdal compiled for ecw, but maybe gdal_translate with -a_srs set to the new projection you want?
23:36:32 leo-unglaub_work: i have gdal with ecw support compiled. converting geotiff to ecw works very well, but i tryed gdalwarp to change the ecw files and gdal says no ECW soppurt..
23:41:09 racicot: leo-unglaub_work: so my understanding is that you just need to assign the srs, not warp... is that correct? if so the -a_srs will just "assign" the srs using gdal_translate...
23:41:19 racicot: maybe I am not understanding your issue...
23:43:12 leo-unglaub_work: maybe you are right, i will try it !!!
23:43:16 leo-unglaub_work: thank you !