#MAPSERVER IRC Log - 2009-12-22

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
06:28:08 CIA-6: dmorissette * r9636 /branches/branch-5-4/docs/download.txt: Update for 5.6.0 release
06:38:47 mdev: hi. I'm having an issue with extent rounding
06:39:36 mdev: between any two given loads of a map using zoomscale to set a zoomlevel, when all other parameters are the same, the resulting extent gets rounded slightly differently on each load but I need it to be identical
06:45:30 FrankW: I am not aware of rounding done on extents in mapserver.
06:45:47 FrankW: How are you determining there is rounding? Is this viewing raster or vector data?
06:46:33 mdev: sometimes $map->extent->maxx yields -4.346991818024 and other times -4.3469918180251
06:46:33 mdev: etc
06:47:09 FrankW: Ah, way way down near the limit of the signifiicant decimals.
06:47:12 mdev: the values seem to drift from load to load... I am trying to determine if it's an issue with the php in the mapscript but it is highly doubtful cause I know I did not add any rounding
06:47:16 mdev: yes
06:47:36 FrankW: That could easily be MapServer and there isn't likely anything that can be done about it.
06:47:38 mdev: the problem is that I am trying to do MD5 hashes of maps
06:47:45 mdev: and the micro bit they are off kills it
06:48:02 FrankW: I would suggest you explicitly round away some of the insigniifcant decimals before computing your hash.
06:48:20 hobu: that would mean interpreting the mapfile in his scenario, I think
06:48:46 mdev: I tried rounding down to 10 ( rather than 12 ) decimal places so there wouldn't be this going on
06:48:57 mdev: but then I think mapserver returns it to 12 and does stuff with it anyway
06:49:36 mdev: the inconsistency is very small and doesn't happen every time or to every extent ( x or y ) value but it is completely inconsistent from load to load
06:49:57 mdev: maybe 1 out of 5 runs are identical
06:50:35 mdev: i did mapsave to save the mapfile and made diffs between them and found that the extent will get rounded slightly different at the decimal end
07:01:07 CIA-6: dmorissette * r9637 /branches/branch-5-4/docs/include/announcements.inc: Added 5.6.0 announcement
07:26:09 FrankW: danmo: Shall I issue the MS 5.6 announcement via osgeo channels?
07:26:24 danmo: I sent it to news_item@...
07:26:35 FrankW: Ah, so you did!
07:26:41 danmo: so the answer is yes, please ( just surprised you didn't see the news_item email )
07:26:43 FrankW: It goes into a different folder and I didn't notice it.
07:27:06 danmo: k. cool. thanks for checking
07:29:46 danmo: BTW, I saw that you bet me to blogging... after thinking about it for many many months, I finally setup a blog ( around the same time you launched yours ), but just haven't got into blogging officially yet... http://dmorissette.blogspot.com/
07:29:47 sigabrt: Title: Geo Gears, Nuts & Bolts ( at dmorissette.blogspot.com )
07:30:21 aboudreault: !!! :P
07:31:57 FrankW: danmo: cool, with regard to the blog. I trust you will get registered with planet.osgeo.org?
07:32:08 danmo: once I start blogging, yes
07:43:08 danmo: FrankW: thanks for posting the announcement
07:51:23 bitnerd: is everyone blogging now??