Steve and I are in discussion to solve this issue. If you have other useful information or want to follow the bug processing, here's the trac ticket:
http://trac.osgeo.org/mapserver/ticket/3467 regards, Alan On June 3, 2010 08:57:27 am Worth Lutz wrote: > Yes, adding the tileindex argument to the resultsGetShape call in PHP > mapscript seems to have solved my problem with moving my application from > MapServer 5.6.1 to 5.6.3. > > Worth > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Worth Lutz > Sent: Wednesday, June 02, 2010 8:12 PM > To: 'Christian Jauvin'; 'Daniel Morissette' > Cc: [email protected] > Subject: RE: [mapserver-users] MapScript resultsGetShape confusion > [solved!] > > This may be the problem I had when I tried to update to 5.6.3. from 5.6.1. > > I'm in the middle of a project and did not have time to figure out what was > causing the problem. The only change was the change to the MapServer > version so I quickly reverted back. I'm using PHP Mapscript. I'll see if > I can find the time to check to see if this is the problem I was having. > > Worth > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Christian > Jauvin > Sent: Wednesday, June 02, 2010 11:18 AM > To: Daniel Morissette > Cc: [email protected] > Subject: Re: [mapserver-users] MapScript resultsGetShape confusion > [solved!] > > Ok.. so with a lot of fiddling around (and with inspiration from a > Frank Warmerdam's testing script) I've been able to find the cause of > the problem, as well as its solution: with a PostGIS connected layer, > a call to resultsGetShape *has* to include the third parameter, > "tileindex": > > resultsGetShape(shp, res.shapeindex) --> hangs MS 5.6.3 > resultsGetShape(shp, res.shapeindex, res.tileindex) --> works great! > > I don't know if it's really a bug, but it contradicts the > documentation, in which the tileindex argument is optional, and it > wasn't behaving like that (to my knowledge) with MS in [5.6.0, 5.6.2]. > > By the way, an initial observation I made was wrong: it also affects > the PHP version, and the workaround is the same. > > Best regards, > > Christian > > > On Tue, Jun 1, 2010 at 5:02 PM, Daniel Morissette > > <[email protected]> wrote: > > Christian Jauvin wrote: > >> I did what you suggested, and added: > >> > >> CONFIG "MS_ERRORFILE" "<error file>" > >> CONFIG "CPL_DEBUG" "ON" > >> DEBUG 5 > > > > Did you also include DEBUG 5 inside the POSTGIS layer itself? > > > >> to my mapfile, and performed my query again. It generated a very long > >> file (200K+ lines!) with only two types of line: > >> > >> [Tue Jun 1 14:36:01 2010].765230 getSymbol(): Symbol definition > >> error. Parsing error near (null):(line 208) > >> [...] > >> [Tue Jun 1 14:36:01 2010].777963 msPostGISLayerResultsGetShape called > >> for record = 338 > >> [...] > > > > 200k times the same message sounds like an infinite loop of some sort. > > > > Perhaps look at the first few lines of debug/error messages. There may be > > a > > > different message that would have been output *before* entering the > > infinite > > > loop and that could give a hint about what is happening. Just a guess... > > can't do much more from here. > > > > What is the last version that worked fine? 5.6.2? > > > > Daniel > > -- > > Daniel Morissette > > http://www.mapgears.com/ > > _______________________________________________ > > mapserver-users mailing list > > [email protected] > > http://lists.osgeo.org/mailman/listinfo/mapserver-users > > _______________________________________________ > mapserver-users mailing list > [email protected] > http://lists.osgeo.org/mailman/listinfo/mapserver-users > > _______________________________________________ > mapserver-users mailing list > [email protected] > http://lists.osgeo.org/mailman/listinfo/mapserver-users > -- Alan Boudreault Mapgears http://www.mapgears.com _______________________________________________ mapserver-users mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/mapserver-users
