ERROR 1: CPLMalloc(-35710063): Silly size requested. Segmentation fault: 11

[peifer]

I simply used ogrinfo in combination with GML_SKIP_RESOLVE_ELEMS=NONE: ogrinfo -al -so -ro somefile.gml

I tried both: GDAL 1.10.1 from kyngchaos.com and self-compiled GDAL 1.11dev from trunk. Crash report attached.

[Even Rouault]

Could you attach such a GML file ?

[peifer]

Here my command line arguments in full length:

ogrinfo -al -so -ro download/B_BE_2014-prelim.gml --config GML_SKIP_RESOLVE_ELEMS NONE --config GML_GFS_TEMPLATE AQD_Zone_as_featureMember.gfs

When trying repeatedly I get either the segfault immediately, or the machine stays at 100% CPU for some time, then finishes w/o segfault.

The gml file is here: ​http://tinyurl.com/oqpzfck, gfs template attached

[Even Rouault]

I couldn't replicate your issue with a normal run, or within Valgrind. But looking at the stack trace you've provided, I could see a potential cause for the crash, so...

trunk r27035 "GML: temptative fix for crash on certain xlink:href with GML_SKIP_RESOLVE_ELEMS=NONE (#5417)"

[peifer]

As mentioned: I can't consistently reproduce the error either. 12 runs resulted in 7 crash reports, which are always about the same source code lines in resolvexlinks.cpp.

I will sync with trunk and let you know. Thanks.

[peifer]

For the same set of 8 GML files, I continue getting with GDAL 1.10.1 from kyngchaos.com something like:

ERROR 1: CPLMalloc(-38843007): Silly size requested.
ERROR 1: CPLMalloc(-39848911): Silly size requested.
ERROR 1: CPLMalloc(-35745567): Silly size requested.
ERROR 1: CPLMalloc(-39849247): Silly size requested.
ERROR 1: CPLMalloc(-35749247): Silly size requested.
ERROR 1: CPLMalloc(-35742847): Silly size requested.
ERROR 1: CPLMalloc(-35673743): Silly size requested.
ERROR 1: CPLMalloc(-35738511): Silly size requested.

Whereas the updated trunk version now results into a total of 4597 messages like "Warning 6 ..." and 4 messages "More than 1000 errors or warnings have been reported. No more will be reported from now." But this is expected behaviour for the given data files, I would assume. So I take the liberty and close the ticket. Thanks again for the quick fix.