#POSTGIS IRC Log - 2008-08-18

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
02:14:47 Oldbie: Hi
02:14:56 Oldbie: what is this "Only KML 2 is supported"?
02:29:38 Oldbie: Any idea?
07:06:22 darkblue_B: hi mloskot
07:15:25 mloskot: hi
07:16:43 darkblue_B: mloskot: can you setup Oracle spatial on an Ubuntu Linux?
07:18:25 mloskot: yes, I can
07:18:44 mloskot: I have it running under ubuntu on my Virtual Machine
07:19:13 darkblue_B: I would be willing to pay you something if you did that on a machine here
07:19:57 mloskot: remotely?
07:20:03 darkblue_B: I can help
07:20:07 darkblue_B: yes :- )
07:20:14 darkblue_B: sorry no plane tckets :P
07:20:55 darkblue_B: its a long airplane flight anyway - not worth it
07:21:40 mloskot: I can do it remotely, but you really doubt you can do it as well?
07:21:44 mloskot: darkblue_B: yourself?
07:21:47 mloskot: it simple :- )
07:22:28 darkblue_B: i doubt that!
07:22:53 mloskot: darkblue_B: when you want me to try?
07:23:23 darkblue_B: anytime you have time..
07:23:46 darkblue_B: I can use Western union, if that goes to your city
07:25:57 mloskot: darkblue_B: I can't do it now, but later tonight. Send me details on mateusz@loskot.net
07:26:26 mloskot: credentials, what oracle version ( express to download ) or other editions ( where can I find them? )
07:26:42 mloskot: installation requirements, should I do default installation or in custom location, etc.
07:27:02 mloskot: what is the hardware there, how much SWAP memory, RAM and disk space
07:27:22 : * mloskot is off for 10 mins
07:27:30 darkblue_B: ok... I have 2 meetings starting in about 7 hours
07:40:33 mloskot: pramsey: last Friday, I've confirmed the prepared geometry leak
07:40:37 mloskot: I'm working now to fix it.
07:40:54 mloskot: But I test it without using postgis, just geos + my test program
07:41:04 pramsey: the leak is in GEOS?
07:41:08 mloskot: yes
07:41:33 mloskot: I'd report it in Saturday, but big storm destroyed our network here in my block
07:41:43 : * mloskot is generating valgrind report now, need a few mins
07:42:06 pramsey: http://lists.refractions.net/pipermail/postgis-devel/2008-August/003436.html
07:42:07 sigq: Title: [postgis-devel] Leaking Prepared Geometries ( at lists.refractions.net )
07:42:38 pramsey: I see one huge leak and two big ones.
07:42:47 mloskot: AFAIR, it was close to your leak, near index/noding stuff
07:42:59 : * mloskot is booting geos vm
07:49:51 mloskot: pramsey: and it leaks not only for big polygons but for small too
07:49:55 mloskot: 100 points star polygon leaks
07:50:16 mloskot: here is my report
07:50:16 mloskot: http://rafb.net/p/3Z6dN535.html
07:50:17 sigq: Title: Nopaste - No description ( at rafb.net )
07:51:59 pramsey: yeah, that *a* memory leak, but not *the* memory leak.
07:53:06 pramsey: my test losts 50% of its memory : )
07:53:07 mloskot: In my leak, vector of segments ( segChains ) leaks
07:53:29 mloskot: So, we have two leaks occurring very close to each other, have I got it?
07:53:47 pramsey: no, I'm pretty sure the big leak is in postgis
07:53:54 pramsey: not in GEOS
07:54:11 pramsey: the report here is for the run of my test case in postgis http://lists.refractions.net/pipermail/postgis-devel/2008-August/003436.html
07:54:12 sigq: Title: [postgis-devel] Leaking Prepared Geometries ( at lists.refractions.net )
07:54:43 pramsey: 8,910,300 leaked in one of them
07:54:55 mloskot: Your report includes number of geos related leaks.
07:55:03 mloskot: Perhaps postgis does not to enough cleanup
07:55:04 pramsey: yes, GEOS leaks too
07:55:22 pramsey: just not nearly as much
07:55:30 mloskot: who is BWJ ?
07:55:43 pramsey: benjubb@refractions.net
07:55:56 mloskot: I see
07:55:59 mloskot: Here is his comment
07:55:59 mloskot: http://trac.osgeo.org/geos/browser/trunk/source/noding/MCIndexSegmentSetMutualIntersector.cpp#L48
07:56:00 sigq: Title: /trunk/source/noding/MCIndexSegmentSetMutualIntersector.cpp - GEOS - Trac ( at trac.osgeo.org )
07:56:08 mloskot: and looks like he agreed to leave it leaking
07:56:25 mloskot: looks like the ownership of segChains returend in line 39 is unclear
07:56:35 mloskot: or tricky to track and find where it should be destroyed
07:57:52 pramsey: perhaps mbdavis could clarify the lifecycle of those objects.
07:58:53 mloskot: it would be far clearer if SpatialIndex::insert method is documented saying who ownes passed objects
07:58:53 mloskot: http://trac.osgeo.org/geos/browser/trunk/source/headers/geos/index/SpatialIndex.h#L50
07:58:54 sigq: Title: /trunk/source/headers/geos/index/SpatialIndex.h - GEOS - Trac ( at trac.osgeo.org )
08:19:28 darkblue_B: ok mloskot I sent you the emal
08:19:58 darkblue_B: email.. you will see, you have to generate a public key to be able to login
08:20:13 darkblue_B: . .when you have time.. I see you are workng
08:23:38 CIA-6: mloskot * r2158 /trunk/source/geom/prep/ ( 13 files ): geom/prep: Refactored kamikaze casts to more readable form. Fixed svn:keywords.
08:24:26 mloskot: darkblue_B: ok, I'll check it
08:24:28 mloskot: later today
08:26:22 CIA-6: mloskot * r2159 /trunk/source/headers/geos/geom/prep/ ( 12 files ): Fixed svn:keywords.
08:27:18 CIA-6: mloskot * r2160 /trunk/source/headers/geos/util.h: Fixed UNREFERENCED_PARAMETER macro.
08:28:30 CIA-6: mloskot * r2161 /trunk/source/ ( 13 files in 5 dirs ): Fixed svn:keywords.
08:30:17 CIA-6: mloskot * r2162 /trunk/tests/unit/ ( 9 files in 5 dirs ): tests/unit: added tests cases to PreparedGeometryFactoryTest, refactored casting utils and geometry comparators, small cleanup.
08:32:11 CIA-6: mloskot * r2163 /trunk/tests/bigtest/ ( GeometryTestFactory.cpp TestSweepLineSpeed.cpp ): tests/bigtest: Pointed out number of memory leaks but not fixing them, waiting for comments.
15:29:20 CIA-6: mloskot * r2164 /trunk/tests/unit/capi/ ( GEOSCoordSeqTest.cpp GEOSSimplifyTest.cpp ): tests/unit/capi: Fixed number of memory leaks.
15:44:59 CIA-6: mloskot * r2165 /trunk/tests/unit/algorithm/CGAlgorithms/isCCWTest.cpp: tests/unit/algorithm/CGAlgorithms: Fixed number of memory leaks.
16:04:24 CIA-6: mloskot * r2166 /trunk/tests/unit/algorithm/CGAlgorithms/isPointInRingTest.cpp: tests/unit/algorithm/CGAlgorithms: Fixed memory leaks.
16:17:24 CIA-6: mloskot * r2167 /trunk/tests/unit/algorithm/ConvexHullTest.cpp: tests/unit/algorithm/ConvexHullTest.cpp: Fixed memory leaks.
16:59:59 KnOppIx_: hello
17:00:04 KnOppIx_: any here to hellp me ???
17:00:39 springmeyer: KnOppIx_: don't ask to ask, just ask