#MAPSERVER IRC Log - 2010-02-08

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
05:37:00 mose: hi here
05:37:07 mose: is the crowd alive ?
05:37:26 hobu: yes, still quite a live
05:37:30 hobu: alive
05:37:34 FrankW: The crowd lives.
05:38:03 mose: I'm preparing to pass over the founder status to hobu
05:38:11 mose: I just wanted to publicize the fact
05:38:57 FrankW: You mean irc channel founder status for #mapserver I guess?
05:39:26 hobu: yep
05:39:39 hobu: which I will in turn give rights to the osgeo group
05:41:17 mose: oh, my bad
05:41:25 mose: I am not founder here
05:41:28 mose: franck is
05:41:36 hobu: ah
05:41:53 mose: probably franck martin but not sure
05:42:57 mose: you could reach him at franck at sopac.org I think
05:43:36 mose: he's not much on irc anymore
05:44:13 mose: I confirm that's him
05:44:22 mose: and I confirm his email
05:44:42 hobu: thanks
05:44:51 mose: my pleasure
05:46:46 hobu: bad email address
05:46:55 mose: what's new in mapserver btw ?
05:47:08 hobu: tons
05:47:10 mose: well he used it still some month ago, let me check
05:47:38 mose: ah try franck at avonsys.com
05:48:21 mose: I had long hollidays so I lost the thread of the geolife
05:48:47 hobu: we're still here ... : )
05:48:57 mose: that rocks
05:49:14 mose: googlemap could have killed
05:49:16 hobu: we even got http://mapserver.org last year too
05:49:17 sigabrt: Title: Welcome to MapServer MapServer 5.6.1 documentation ( at mapserver.org )
05:49:38 mose: google is damn predator
05:50:24 nhv: an enabling predator :- )
05:50:58 mose: for sure that's not such an evil predator as former dinausors
05:51:05 mose: but still
05:51:54 mose: they are liable to USA law as sourceforge is
05:52:08 mose: so they may be forced to change policy with no notice
05:52:33 mose: not to blame them, but that sourceforge blocking made quite a shock already
05:56:35 Dalian: test
05:57:00 FrankW: Dalian: ack
06:16:29 strk: SLD anyone ? two words on it ?
06:16:52 strk: like... can it be used to define classifications for areal features ?
06:19:23 hobu: strk: did you see http://mapserver.org/ogc/sld.html
06:19:25 sigabrt: Title: SLD MapServer 5.6.1 documentation ( at mapserver.org )
06:19:41 hobu: nice listing of tables there describing what mapserver supports
06:19:43 hobu: and links to the spec
06:21:14 strk: reading that now, thanks
06:22:24 strk: uhm, I think styling would come later in the process though. first I'll need to get the data in.
06:22:47 strk: the need is to create a kind of web service to return thematized maps at country and region level
06:22:59 strk: w/out having the actual data in advance ( except the countries and regions boundaries )
06:23:22 strk: can PHPMapScript provide a custom FeatureDriver to mapserver ?
06:23:39 strk: maybe using a JOIN or something
06:34:23 strk: MS_INLINE ?
06:35:54 FrankW: strk: I do not think you can provide a custom feature driver, but you can do a custom query in php and push the results into an inline layer.
06:41:54 strk: best would be for it to NOT include the geometries
06:42:07 strk: but only the associated data ( per-country )
06:42:27 strk: is there a way for that ? ( a DATA object or something like that )
06:46:52 strk: postgis is fine as a provider for actual feature, but classification/thematization should be done based on values to be found somewhere elase
06:47:01 strk: what options do I have to provide such "external" values ?
07:23:05 hobu: assefa: do you have a mssql instance that you can fix bugs against?
07:24:52 assefa: hobu: yes I have sql express 2008
07:25:01 hobu: I'
07:25:15 hobu: I'll cc you on this. someone was nagging me about it though I'm not sure it's mssql's fault
07:25:16 hobu: http://trac.osgeo.org/mapserver/ticket/3058
07:25:17 sigabrt: Title: #3058 ( MSSQL2008 and nquery return only: "Content-type: text/html" ) - MapServer - Trac ( at trac.osgeo.org )
07:27:53 assefa: I will try and see if I can reproduce it locally ( with cgi calls ). Won't be able to test c#.
07:28:03 : * hobu nods
08:17:08 aboudreault: assefa: hmm, getting an error when I try to load your data in my oracle db
08:17:15 aboudreault: ERROR 1: ORA-01653: unable to extend table USER_TESTOS.BODEM50 by 1024 in tablespace TESTOS
08:18:24 aboudreault: I'm looking to find an quick fix for that
08:33:14 assefa: aboudreault: mmm never run into this issue my self.
08:37:19 assefa: there seems to be ways to extend the tablespace ( using resize or autoextend ): http://74.125.47.132/search?q=cache:WRJ5KkBM13cJ:dbmotive.com/oracle_error_codes.php%3Ftype%3DORA%26errcode%3D01653+oracle+ORA-01653&cd=1&hl=en&ct=clnk&gl=ca&client=firefox-a
08:37:20 sigabrt: Title: ORA-01653: Unable to extend table %s.%s by %s in tablespace %s - DbMotive - Quality Oracle Software and Services ( at 74.125.47.132 )
08:45:00 aboudreault: assefa: yeah, I'm trying to alter my tablespace. It should work.
08:45:15 assefa: ok
09:09:37 : * aboudreault rebuilds php with curl extension
09:21:27 assefa: aboudreault: sorry it took more of your time as I thought
09:22:13 aboudreault: np assefa
09:22:24 assefa: you should switch to windows: every thing is always pre-build : )
09:22:28 aboudreault: ok, I executed the script. Am I supposed to see something?
09:22:44 assefa: yes
09:22:53 aboudreault: god.... forget it!! : )
09:23:09 aboudreault: for the moment, at see a blank page
09:23:28 danmo: pramsey: for #3291 you need to download the raw patch. I think it's missing the multi-file separator and Trac doesn't display it properly
09:23:40 pramsey: ah, thanks danmo
09:24:12 assefa: I mean you should get soemthing like wfs:FeatureCollection ... or at least an error
09:24:32 aboudreault: assefa: I've uncommented the echo at the end of the script, but still getting no output.
09:26:37 aboudreault: assefa: ha, the port was missing in my url... wait got a projection error now
09:26:48 assefa: ah ok
09:27:51 aboudreault: assefa: here's the script's result: http://pastebin.ca/1790367
09:27:52 sigabrt: Title: pastebin - Unnamed - post number 1790367 ( at pastebin.ca )
09:31:01 assefa: aboudreault: would you mind doing a view source and e-mailing me the source. Difficult to see in pastebin
09:32:33 aboudreault: sent
09:33:48 assefa: perfect thanks a lot.
09:34:29 assefa: you got the same results I am getting here.
09:35:17 aboudreault: good.
10:06:27 pwillis: Hello,
10:06:29 pwillis: I continue to have a problem selecting
10:06:31 pwillis: a single band from a multi-band ( 563 bands )
10:06:32 pwillis: ENVI format BSQ file.
10:06:34 pwillis: Using the following URL mapserver gives me a FLOAT32 GTiff
10:06:35 pwillis: file but it is always filled with zeros.
10:06:37 pwillis: http://[something]/cgi-bin/wcs?REQUEST=GetCoverage&SERVICE=WCS&VERSION=1.0.0&COVERAGE=NDVI&CRS=EPSG:4326&BBOX=-135,55,-121,46&WIDTH=256&HEIGHT=256&FORMAT=GEOTIFF_FLOAT&Band=122
10:06:38 pwillis: Is there any documentation that clearly
10:06:39 pwillis: shows how WCS can be set up to do this?
10:06:41 pwillis: My current LAYER definition is as follows:
10:06:43 pwillis: LAYER
10:06:44 pwillis: NAME VegetationIndex
10:06:46 pwillis: STATUS OFF
10:06:47 pwillis: DEBUG ON
10:06:49 pwillis: TYPE RASTER
10:06:50 pwillis: METADATA
10:06:52 pwillis: "wcs_label" "Data/NDVI"
10:06:53 pwillis: "wcs_rangeset_name" 'bands'
10:06:54 pwillis: "wcs_rangeset_label" "NDVI"
10:06:56 pwillis: "ows_extent" '-135 55 -121 46'
10:06:57 aboudreault: pwillis: use pastebin !
10:06:57 pwillis: "wcs_resolution" '0.08333333333333332 -0.083333333333333332'
10:06:59 pwillis: "ows_srs" 'EPSG:4326'
10:07:00 pwillis: "wcs_srs" 'EPSG:4326'
10:07:02 pwillis: "wcs_formats" 'GEOTIFF_FLOAT,GEOTIFF_INT16'
10:07:04 pwillis: "wcs_nativeformat" 'ENVI'
10:07:05 pwillis: END
10:07:06 pwillis: DATA /NDVI/ndvi.img
10:07:08 pwillis: PROJECTION
10:07:10 pwillis: "init=epsg:4623"
10:07:11 pwillis: END
10:07:12 pwillis: DUMP TRUE
10:07:14 pwillis: END
10:07:19 pwillis: pastebin?
10:07:37 aboudreault: to paste, yes. pastebin.com
10:08:13 pwillis: O.K. Sorry.
10:08:25 aboudreault: np
10:23:04 pwillis: Hello,
10:23:06 pwillis: I continue to have a problem selecting
10:23:07 pwillis: a single band from a multi-band ( 563 bands )
10:23:08 pwillis: ENVI format BSQ file.
10:23:10 pwillis: Using the following URL mapserver gives me a FLOAT32 GTiff
10:23:12 pwillis: file but it is always filled with zeros.
10:23:13 pwillis: A sample *bogus* parameter URL may be found at:
10:23:14 pwillis: http://pastebin.ca/1790437
10:23:15 sigabrt: Title: pastebin - sample WCS URL - post number 1790437 ( at pastebin.ca )
10:23:16 pwillis: Is there any documentation that clearly
10:23:17 pwillis: shows how WCS can be set up to do this?
10:23:19 pwillis: My current LAYER definition is as follows:
10:23:21 pwillis: http://pastebin.ca/1790434
10:23:22 sigabrt: Title: pastebin - ENVI mapfile - post number 1790434 ( at pastebin.ca )
10:24:41 FrankW: pwillis: Let me take a look and see if I can make a suggestion.
10:27:24 FrankW: I see the following line in mapwcs.c / msWCSGetCoverageBands10( ):
10:27:27 FrankW: value = msOWSLookupMetadata( &( lp->metadata ), "COM", "rangeset_axes" ); /* this will get all the compound range sets */
10:27:38 FrankW: This seems to imply you need to have rangeset_axes metadata on the layer.
10:28:05 FrankW: I'm a bit vague on what it should hold, but perhaps you could review the docs in this light.
10:28:16 pwillis: FrankW: wcs_rangeset_axes ?
10:28:32 FrankW: sorry, yes, wcs_rangeset_axes.
10:29:04 FrankW: I think Steve did this portion so I'm a bit unfamiliar with the details.
10:29:15 pwillis: I have only seen that in the context of timeseries. Perhaps if I put all the spatial axes in as well..
10:33:08 FrankW: The code to extract the band list is at: http://trac.osgeo.org/mapserver/browser/trunk/mapserver/mapwcs.c#L1241
10:33:12 sigabrt: Title: /trunk/mapserver/mapwcs.c - MapServer - Trac ( at trac.osgeo.org )
10:33:14 FrankW: perhaps that will help a bit.
10:33:58 pwillis: FrankW: OK, I have wcs_rangeset_axes 'bands' set. I will look at the source code.
10:34:00 pwillis:
10:34:34 FrankW: If you are still unsuccessful, you can file a ticket and assign it to me ( warmerdam ) and I'll try and prepare a working example of band selection with WCS 1.0
10:35:01 pwillis: FrankW: I am running a GetCoverage request and a file is downloading. I will know once I have looked at the file.
10:39:08 pwillis: FrankW: No joy. The resulting FLOAT_32 GeoTiff is still all zeros. I'll snoop at the source code.
10:40:58 FrankW: OK, if you do file a ticket with a simple mapfile, sample data file, and query, I will try to fix things up.
10:41:07 FrankW: It would be good for the test suite to be extended with such a case.
11:04:11 pwillis: FrankW: I'm making up some smaller sample ENVI data files that you can download to test with. I'll post the URL( s ) for you once I have them set up.
11:04:33 FrankW: ok, cool
12:47:55 pwillis: FrankW: Regarding the ENVI Format multiband WCS test files; There are some sample ENVI format files, with latlon geography, located at:
12:47:56 pwillis: http://filebin.ca/cpaux/ENVI_test_files.ZIP
12:47:58 pwillis: These are vary simple files. I burned text into each of the six channels. If they are mapped correctly you should see 'B-001' on a zero ( black ) ground for band 1, 'B-002' for band2, up to 'B-006' for band 6. The data statistics for each 'IMG' file are provided in the 'readme.txt'. You should be able to get he min and max of the data from that.
12:47:59 pwillis: The int32 files have *negative* data in them. They are actually mislabeled and are signed 16 bit. ( oops! ) It wasnt meant to be a tricky test or anything...
12:48:01 pwillis: There is a list of possible ENVI/IDL data types in the readme.txt file as well.
12:48:02 pwillis: Do you need anything else for testing?
12:49:29 FrankW: pwillis: taking a look.
12:50:56 FrankW: I don't see the ticket.
12:51:34 pwillis: FrankW: OOPS!, forgot the ticket....I'll fill one out. The data is at the URL though.
12:52:04 FrankW: I don't see a map file
12:52:25 FrankW: I'd also like an url that demonstrates the problem.
12:52:38 FrankW: I don't see that in the README
12:53:07 FrankW: I think there is more than enough data - it would have been better to pare it down to a single data file, minimally demonstrating the issue.
12:53:16 FrankW: but I can work with what you provided.
12:58:39 pwillis: FrankW: I figured you'd want some sample files to test various data types and interleaves. Just use the double precision BSQ file
12:58:41 pwillis: as the base test. That should be more than bad enough for all purposes.
12:58:42 pwillis: I will set up a test sample WCS with that particular ( double64 ) IMG file and HDR if you like.
12:59:45 FrankW: A sample wcs map file for one of the files would be good. I'd rather start from your setup and then tweak rather than recreate from scratch.
12:59:53 pwillis: FrankW: NOTE: the tif files are not GeoTiff files. Only the IMG/HDR files have geography.
13:00:05 FrankW: And whether envi interleaving formats are all properly supported or not is a distinct issue from MapServer WCS handling of bands.
13:00:24 FrankW: and would go in a GDAL ticket if there was a problem.
13:11:33 pwillis: FrankW: A sample map file has been uploaded to:
13:11:35 pwillis: http://filebin.ca/upkmxs/mapfile.txt
13:11:35 CIA-7: assefa * r9780 /trunk/mapserver/mapwfs.c: Correct wfs describefeature mimetype #3294
13:11:36 pwillis: Nothing too complicated there. One raster layer declaration.
13:14:25 pwillis: FrankW: Sorry, the pixel size dimensions in the layer should be 2.8125 degrees not 0.08333. The file is only 128x64 spatially.
13:34:14 pwillis: FrankW: There is a WCS site set up with the double precision file: I just tested the smaller file with a GetCoverage request and it works....<sigh!>
13:34:16 pwillis: Here's the link for band 6 of that file:
13:34:17 pwillis: http://grip.aslenv.com/cgi-bin/wcs_1?REQUEST=GetCoverage&SERVICE=WCS&VERSION=1.0.0&COVERAGE=BOGUStestData&CRS=EPSG:4326&BBOX=-180,-90,180,90&WIDTH=128&HEIGHT=64&FORMAT=GEOTIFF_FLOAT&bands=6
13:34:19 pwillis: I will go back to the bigger file and check that again. Maybe it's a byte order thing or size of the input file.
13:58:17 pwillis: FrankW: I'm at a loss. I just tested the big IMG file and now it works! The only difference being, I removed a bunch of in-line comments from the map file. Everything appears to work now.
14:08:10 pwillis: FrankW: Where can I document this so that people can use this information?
14:14:27 FrankW: pwillis: i'm not sure what aspect you would be documenting.
14:14:45 FrankW: I am dubious that removing comments would really have an effect unless there is a subtle bug we could investigate.
14:14:56 FrankW: Certainly we don't want to write a document suggesting people don't document their map files!
14:19:42 FrankW: pwillis: If you would like to propose an additional example demonstrating bands for the WCS Server or WCS Use Cases documents, that could be helpful.
14:19:52 FrankW: http://mapserver.org/ogc/wcs_format.html or http://mapserver.org/ogc/wcs_server.html
14:19:53 sigabrt: Title: WCS Use Cases MapServer 5.6.1 documentation ( at mapserver.org )
14:20:36 FrankW: Hmm, it looks like the last layer might be an example of making base selectable.
14:25:34 pwillis: FrankW: Re:"Last Layer" there was only one raster layer in the map file. What do you mean by selectable base?
14:26:34 FrankW: I mean the last example in the use cases document seemed to provide for a selectable band. ( it mistyped band as base, sorry )
14:28:00 pwillis: Ah!, I will add some example use case docs to those links you provided. Would I have access to that wiki or do I need a secret funny walk, etc.?
14:33:25 pwillis: FrankW: I think I'm being slashdotted....<yikes!>
14:33:46 FrankW: The mapserver.org site isn't a wiki, so you would need to file a ticket with new materials for the document( s ).
14:33:57 FrankW: You can assign it to me for the WCS docs and I'll apply it.
14:34:30 pwillis: FrankW: O.K. I'll set up a new ticket and try to be as complete as possible.
15:30:12 rd_nielsen: How do you debug a GetFeatureInfo call that returns 200 OK but with no data? That is, when it actually should return data.
15:30:36 rd_nielsen: Does MapServer ( still ) support the RADIUS parameter? It doesn't complain about it, at least.
18:04:12 CIA-7: pramsey * r9781 /branches/branch-5-4/mapserver/ ( HISTORY.TXT configure.in mappostgis.c ): Determine PgSQL version in a more backwards compatible way ( #3291 )
18:04:55 CIA-7: pramsey * r9782 /branches/branch-5-6/mapserver/ ( HISTORY.TXT configure.in mappostgis.c ): Determine PgSQL version in a more backwards compatible way ( #3291 )
18:10:21 CIA-7: pramsey * r9783 /branches/branch-5-4/mapserver/mappostgis.c: Remove odd self-strdup from patch.
18:12:16 CIA-7: pramsey * r9784 /branches/branch-5-6/mapserver/mappostgis.c: Remove odd self-strdup construction.
18:25:32 CIA-7: pramsey * r9785 /trunk/mapserver/ ( configure.in mappostgis.c ): Determine PgSQL version in a more backwards compatible way ( #3291 )