#MAPSERVER IRC Log - 2010-01-14

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
00:23:43 beatpanic: hi, it is possible to use SLD inside a .map file?
00:24:15 beatpanic: or to make an sld with a desktop gis software and then style it with mapserver?
00:24:26 beatpanic: thanks : )
00:33:04 lucadelu: beatpanic: you can make a sld with wgis
00:33:07 lucadelu: qgis
00:33:24 lucadelu: there is a python plugin for export the sld
00:33:57 beatpanic: lucadelu, yep, and then how could run it inside mapserver? I'm looking at the SLD url parameter but I'm not sure
00:35:00 lucadelu: beatpanic: i think you can use sld with wms
00:35:09 lucadelu: http://mapserver.org/ogc/sld.html
00:35:10 sigabrt: Title: SLD MapServer 5.4.2 documentation ( at mapserver.org )
00:36:03 beatpanic: lucadelu, exactly, I'm wondering if it is sufficient to use the SLD=URL parameter to add a layer via wms ( which I'm using )
00:41:16 beatpanic: lucadelu, do you mind to pass me the link of the qgis python script to export in sld? thanks
00:41:38 beatpanic: lucadelu, or the python plugin name
00:49:45 lucadelu: beatpanic: the name si SLD export
05:14:01 aboudreault: FrankW: around? I did a few tests with your rfc 52 changes for OGR ( also tested the autostyling you fixed ), but I'm wondering if this really should to be backported in 5.6.
06:53:15 aboudreault: jmckenna: what about branching 5.6 doc? We'll make the "images" folder shared for all languages in the future.
06:55:10 jmckenna: i have a couple of questions about this. will respond right after #osgeo fundraising meeting now
06:55:26 aboudreault: ah, all right
06:56:48 gislars: aboudreault: if I look at http://mapserver.org/dev/en/ the link to the german version contains /trunk/ in the url
06:56:49 sigabrt: Title: Welcome to MapServer MapServer 5.4.2 documentation ( at mapserver.org )
06:57:08 gislars: is it because of the some redirects?
06:57:18 gislars: or are there two version now online?
06:59:51 aboudreault: It's only the link the in the sphinx doc. I'm going to move /dev to /trunk right now.
07:00:28 gislars: ok
08:08:29 CIA-9: jmckenna * r9689 /branches/branch-5-6/docs/: attempting to create a 5-6 docs branch
08:24:05 jmckenna: aboudreault: it seems some things have changed in /osgeo/mapserver/ on xblade14 since last night? meaning: i could have sworn that a 'mapserver-web-directory' existed there, containing trunk docs in html format, and it was what was shown live at http://mapserver.org/dev/ Maybe i am missing something here???
08:24:43 jmckenna: directory missing is 'mapserver-web-new' containing trunk html docs. ?
08:27:25 jmckenna: ping me so we can work together on making some sense of it all
08:50:42 aboudreault: jmckenna: i deleted -new
08:50:56 jmckenna: um ok.
08:50:57 aboudreault: use /trunk only
08:51:08 jmckenna: i'm looking at it now
08:51:25 aboudreault: the /dev was only to do my test when I began to look at the doc.
08:51:56 jmckenna: i checked out 'docs.54' and 'docs.56' there
08:52:26 jmckenna: i imagine 'docs.56' should be renamed to 'docs' ( as this is the live directory )
08:52:50 aboudreault: hobu is supposed to ask the sphinx folks on the google group about the link problem ( that you can see on /trunk/de/documentation.html )
08:53:42 jmckenna: and i believe we can use the 'docs.54' dir to point to the mapserver_54.pdf ( that we can generate )
08:54:14 jmckenna: err that we can rename ( since i exists in 5.4 ) yikes i hate versions, grrrrrrrrrr
08:54:27 jmckenna: this is like pulling teeth lol
08:54:48 aboudreault: docs is already the branch 5.4, so we can just rename it
08:55:01 jmckenna: yup ( said that ) : )
08:55:13 jmckenna: oh i see what u mean..true true
08:55:32 aboudreault: we have to copy/add the redirection folder of that branch in the 5.6 branch
08:56:57 aboudreault: I can do it if you want.
08:57:14 aboudreault: I can put in place the 5.6 doc online
08:57:21 jmckenna: ok
08:57:45 jmckenna: i will remove the unnecessary 'docs.54' folder
08:58:00 aboudreault: ok, will do the change right now.
09:08:39 CIA-9: aboudreault * r9690 /branches/branch-5-4/docs/redirection/ ( mapserver-redirect-trunk.conf mapserver-redirect.conf ): Added redirect-trunk apache config
09:09:40 CIA-9: aboudreault * r9691 /branches/branch-5-6/docs/redirection/ ( 4 files ): Added redirection files in 5.6
09:10:25 : * aboudreault starts the doc generation
09:14:48 jmckenna: i tried to generate 'docs.56' on xblade14 and got a fatal error. is that what u are doing?
09:16:01 aboudreault: yes, I tought you said me "ok" to do it
09:16:19 aboudreault: to put in place the 5.6. doc
09:16:37 jmckenna: oh i did not understand what that meant. ok sorry, carry on : )
10:06:58 aboudreault: the link of the de/usa flag is not good...... trying to find how this is generated
10:14:40 hobu: aboudreault: no, you were supposed to ask the sphinx folks because you have a better understanding of the problem than I do : )
10:19:02 aboudreault: <hobu> I would craft an email to [...], you are right, probably read a "will" rather than "would"
10:19:20 aboudreault: will do.
10:22:02 aboudreault: hobu: perhaps you know ( on www.mapserver.org ) why the link of the different languages ( the flags ) are not ok? but they are ok in www.mapserver.org/trunk/
10:23:10 hobu: the link generated appears to be http://www.mapserver.org/html/de/index.html
10:23:11 sigabrt: Title: Now Hosted on www.mapserver.org ( at www.mapserver.org )
10:23:16 hobu: but you want http://www.mapserver.org/de/index.html
10:23:17 sigabrt: Title: Willkommen zum MapServer MapServer 5.6.1 documentation ( at www.mapserver.org )
10:23:25 aboudreault: yes, I don't know where this is generated
10:28:54 : * hobu looks
10:29:08 hobu: are we branched now?
10:29:20 hobu: and the website is running off the branch?
10:30:28 aboudreault: 5.6 is branched yes. the website is the branch
10:35:42 hobu: aboudreault: check out layout.html in _templates for the problem with the flag
10:35:56 hobu: I don't understand the logic or the relative paths in there though. We should ask lars
10:41:36 aboudreault: ok, will take a look
11:03:59 CIA-9: hobu * r9692 /branches/branch-5-6/docs/_templates/layout.html: don't point to html directory for language flags
11:05:03 aboudreault: ha... you can revert this change hobu. I found where to set the target
11:05:17 hobu: oh, ok
11:05:22 aboudreault: I just uncommentted the line in the Makefile, rebuildind now
11:05:41 CIA-9: hobu * r9693 /branches/branch-5-6/docs/_templates/layout.html: revert r9692
11:09:11 hobu: make sure to tweak the crontab if necessary
11:11:00 hobu: oh, a locally modified makefile. I don't know that I like that, but best of the worse I suppose
12:12:54 aboudreault: hobu: Is it you who updated svn on osgeo server?
12:34:20 rbranson: anyone think there would be an interest in LZO compressed shapefiles?
12:35:26 FrankW: rbranson: do you mean the ability to automatically utilize them in mapserver?
12:35:40 FrankW: I'm not familiar with LZO compression.
12:35:41 rbranson: yes, and also include a shp compression tool
12:35:55 rbranson: LZO compression is designed for "real-time" decompression performance
12:36:12 rbranson: it's not as good as gzip or bzip2, but it's much faster for decompression
12:36:45 FrankW: you are aware that this would almost certainly defeat .qix spatial indexing, right?
12:36:54 rbranson: why so?
12:37:04 FrankW: it only works well when we can effectively seek to a particular part of the file quickly.
12:37:13 FrankW: I presume a compressed file will need to be decompressed from the start.
12:37:26 rbranson: i'm just now thinking about it honestly
12:37:41 rbranson: afaik, the qix file would have to be rebuilt to support the compressed shapefile
12:37:48 rbranson: as the offsets would change
12:38:23 FrankW: If LZO is anything like gzip or other common compression methods, it is very hard to start decompressing in the middle of the file.
12:38:34 rbranson: you wouldn't compress the entire file
12:38:43 rbranson: just the geometry data
12:38:59 rbranson: it wouldn't be effective on point data obviously
12:39:04 FrankW: Are you suggesting that each geometry would be independently compressed?
12:39:08 rbranson: yes
12:39:36 FrankW: Does the algorithm would well on relatively small chunks?
12:39:49 FrankW: I know many of the algorithms don't get much benefit over short segments.
12:40:06 rbranson: yeah, i am looking at that now
12:40:11 rbranson: it probably won't be good for very short pieces of data
12:40:22 rbranson: but it will work well for polygons with a few hundred or so points
12:41:07 FrankW: I think this would be a relatively complicated implementation.
12:41:09 rbranson: we constantly deal with >50MB shapefiles with large polygons
12:41:33 rbranson: once the shapefile drops out of RAM, performance obviously plummets
12:41:59 rbranson: our largest poly file is ~1GB
12:42:29 FrankW: With spatial indexing, it should not be necessary for the whole file to be in ram cache for good performance.
12:42:37 FrankW: But clearly compressing big geometries could be helpful.
12:43:21 FrankW: For what it's worth, others with a similar issue have developed formats that first change the geometries so each point is stored as a difference from the last point.
12:43:34 FrankW: And that used some reduction in precision to get good compression.
12:43:42 rbranson: hmm, that's interesting
12:44:00 rbranson: so you'd go from doubles to single-precision floats?
12:44:31 FrankW: Likely to integers based on some "compression defined precision" provided by the user.
12:44:48 rbranson: yeah, hmm
12:45:09 rbranson: best case you're probably only looking at half the size though
12:45:22 FrankW: Anyways, if you wanted to try an LZO "per geometry" compression effort with mapserver, I'd be interested in hearing how it goes.
12:45:37 rbranson: well then if I get going on this I'll keep you posted
12:45:56 rbranson: we haven't hit any walls yet, but it's looking like we're about to
12:49:33 rbranson: bah, LZO is GPL
12:49:55 FrankW: Is that a barrier to you? Are you linking mapserver into something proprietary?
12:50:12 rbranson: not really, but I was hoping for BSD
12:50:18 rbranson: so it'd be compatible with the MapServer license
12:50:40 FrankW: It's ok for MapServer to optionally depend on GPL libraries.