MiraMonVector: Fixing chromium bug 68809

[AbelPau]

What does this PR do?

Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809
A copy of a pointer was freed instead of the pointer.

What are related issues/pull requests?

https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809

Tasklist

• Make sure code is correctly formatted (cf pre-commit configuration)
• All CI builds and checks have passed

[rouault]

thanks!

[rouault]

@AbelPau This PR doesn't fix https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809. I can still reproduce the leak with latest master including this merged pull request:

$ valgrind --leak-check=full fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720 
==2045802== Memcheck, a memory error detector
==2045802== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==2045802== Using Valgrind-3.15.0 and LibVEX; rerun with -h for copyright info
==2045802== Command: fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720
==2045802== 
==2045802== 
==2045802== HEAP SUMMARY:
==2045802==     in use at exit: 708,976 bytes in 5,464 blocks
==2045802==   total heap usage: 10,934 allocs, 5,470 frees, 1,873,812 bytes allocated
==2045802== 
==2045802== 16,129 (2,128 direct, 14,001 indirect) bytes in 1 blocks are definitely lost in loss record 5,459 of 5,461
==2045802==    at 0x483DD99: calloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==2045802==    by 0x50AEC28: VSICalloc (cpl_vsisimple.cpp:454)
==2045802==    by 0x5EB8FD9: MM_CreateEmptyHeader (mm_gdal_functions.c:243)
==2045802==    by 0x5EB905A: MM_CreateDBFHeader (mm_gdal_functions.c:271)
==2045802==    by 0x5EB611F: MMCreateMMDB (mm_wrlayr.c:6355)
==2045802==    by 0x5EB17C8: MMCreateRecordDBF (mm_wrlayr.c:4411)
==2045802==    by 0x5EB207B: MMAddFeature (mm_wrlayr.c:4712)
==2045802==    by 0x5EA45DA: OGRMiraMonLayer::MMWriteGeometry() (ogrmiramonlayer.cpp:1938)
==2045802==    by 0x5EA3B86: OGRMiraMonLayer::MMProcessGeometry(OGRGeometryHS*, OGRFeature*, char) (ogrmiramonlayer.cpp:1694)
==2045802==    by 0x5EA3C2D: OGRMiraMonLayer::ICreateFeature(OGRFeature*) (ogrmiramonlayer.cpp:1723)
==2045802==    by 0x610C3C6: OGRLayer::CreateFeature(OGRFeature*) (ogrlayer.cpp:786)
==2045802==    by 0x63DA76A: LayerTranslator::Translate(OGRFeature*, TargetLayerInfo*, long long, long long*, long long&, int (*)(double, char const*, void*), void*, GDALVectorTranslateOptions const*) (ogr2ogr_lib.cpp:6466)

[AbelPau]

@AbelPau This PR doesn't fix https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809. I can still reproduce the leak with latest master including this merged pull request:

$ valgrind --leak-check=full fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720 
==2045802== Memcheck, a memory error detector
==2045802== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==2045802== Using Valgrind-3.15.0 and LibVEX; rerun with -h for copyright info
==2045802== Command: fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720
==2045802== 
==2045802== 
==2045802== HEAP SUMMARY:
==2045802==     in use at exit: 708,976 bytes in 5,464 blocks
==2045802==   total heap usage: 10,934 allocs, 5,470 frees, 1,873,812 bytes allocated
==2045802== 
==2045802== 16,129 (2,128 direct, 14,001 indirect) bytes in 1 blocks are definitely lost in loss record 5,459 of 5,461
==2045802==    at 0x483DD99: calloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==2045802==    by 0x50AEC28: VSICalloc (cpl_vsisimple.cpp:454)
==2045802==    by 0x5EB8FD9: MM_CreateEmptyHeader (mm_gdal_functions.c:243)
==2045802==    by 0x5EB905A: MM_CreateDBFHeader (mm_gdal_functions.c:271)
==2045802==    by 0x5EB611F: MMCreateMMDB (mm_wrlayr.c:6355)
==2045802==    by 0x5EB17C8: MMCreateRecordDBF (mm_wrlayr.c:4411)
==2045802==    by 0x5EB207B: MMAddFeature (mm_wrlayr.c:4712)
==2045802==    by 0x5EA45DA: OGRMiraMonLayer::MMWriteGeometry() (ogrmiramonlayer.cpp:1938)
==2045802==    by 0x5EA3B86: OGRMiraMonLayer::MMProcessGeometry(OGRGeometryHS*, OGRFeature*, char) (ogrmiramonlayer.cpp:1694)
==2045802==    by 0x5EA3C2D: OGRMiraMonLayer::ICreateFeature(OGRFeature*) (ogrmiramonlayer.cpp:1723)
==2045802==    by 0x610C3C6: OGRLayer::CreateFeature(OGRFeature*) (ogrlayer.cpp:786)
==2045802==    by 0x63DA76A: LayerTranslator::Translate(OGRFeature*, TargetLayerInfo*, long long, long long*, long long&, int (*)(double, char const*, void*), void*, GDALVectorTranslateOptions const*) (ogr2ogr_lib.cpp:6466)

Oh, ok. Then I'll take another look on Monday.

[AbelPau]

@AbelPau This PR doesn't fix https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809. I can still reproduce the leak with latest master including this merged pull request:

$ valgrind --leak-check=full fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720 
==2045802== Memcheck, a memory error detector
==2045802== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==2045802== Using Valgrind-3.15.0 and LibVEX; rerun with -h for copyright info
==2045802== Command: fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720
==2045802== 
==2045802== 
==2045802== HEAP SUMMARY:
==2045802==     in use at exit: 708,976 bytes in 5,464 blocks
==2045802==   total heap usage: 10,934 allocs, 5,470 frees, 1,873,812 bytes allocated
==2045802== 
==2045802== 16,129 (2,128 direct, 14,001 indirect) bytes in 1 blocks are definitely lost in loss record 5,459 of 5,461
==2045802==    at 0x483DD99: calloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==2045802==    by 0x50AEC28: VSICalloc (cpl_vsisimple.cpp:454)
==2045802==    by 0x5EB8FD9: MM_CreateEmptyHeader (mm_gdal_functions.c:243)
==2045802==    by 0x5EB905A: MM_CreateDBFHeader (mm_gdal_functions.c:271)
==2045802==    by 0x5EB611F: MMCreateMMDB (mm_wrlayr.c:6355)
==2045802==    by 0x5EB17C8: MMCreateRecordDBF (mm_wrlayr.c:4411)
==2045802==    by 0x5EB207B: MMAddFeature (mm_wrlayr.c:4712)
==2045802==    by 0x5EA45DA: OGRMiraMonLayer::MMWriteGeometry() (ogrmiramonlayer.cpp:1938)
==2045802==    by 0x5EA3B86: OGRMiraMonLayer::MMProcessGeometry(OGRGeometryHS*, OGRFeature*, char) (ogrmiramonlayer.cpp:1694)
==2045802==    by 0x5EA3C2D: OGRMiraMonLayer::ICreateFeature(OGRFeature*) (ogrmiramonlayer.cpp:1723)
==2045802==    by 0x610C3C6: OGRLayer::CreateFeature(OGRFeature*) (ogrlayer.cpp:786)
==2045802==    by 0x63DA76A: LayerTranslator::Translate(OGRFeature*, TargetLayerInfo*, long long, long long*, long long&, int (*)(double, char const*, void*), void*, GDALVectorTranslateOptions const*) (ogr2ogr_lib.cpp:6466)

@AbelPau This PR doesn't fix https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=68809. I can still reproduce the leak with latest master including this merged pull request:

$ valgrind --leak-check=full fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720 
==2045802== Memcheck, a memory error detector
==2045802== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==2045802== Using Valgrind-3.15.0 and LibVEX; rerun with -h for copyright info
==2045802== Command: fuzzers/gdal_vector_translate_fuzzer clusterfuzz-testcase-minimized-gdal_vector_translate_fuzzer-6575332009246720
==2045802== 
==2045802== 
==2045802== HEAP SUMMARY:
==2045802==     in use at exit: 708,976 bytes in 5,464 blocks
==2045802==   total heap usage: 10,934 allocs, 5,470 frees, 1,873,812 bytes allocated
==2045802== 
==2045802== 16,129 (2,128 direct, 14,001 indirect) bytes in 1 blocks are definitely lost in loss record 5,459 of 5,461
==2045802==    at 0x483DD99: calloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==2045802==    by 0x50AEC28: VSICalloc (cpl_vsisimple.cpp:454)
==2045802==    by 0x5EB8FD9: MM_CreateEmptyHeader (mm_gdal_functions.c:243)
==2045802==    by 0x5EB905A: MM_CreateDBFHeader (mm_gdal_functions.c:271)
==2045802==    by 0x5EB611F: MMCreateMMDB (mm_wrlayr.c:6355)
==2045802==    by 0x5EB17C8: MMCreateRecordDBF (mm_wrlayr.c:4411)
==2045802==    by 0x5EB207B: MMAddFeature (mm_wrlayr.c:4712)
==2045802==    by 0x5EA45DA: OGRMiraMonLayer::MMWriteGeometry() (ogrmiramonlayer.cpp:1938)
==2045802==    by 0x5EA3B86: OGRMiraMonLayer::MMProcessGeometry(OGRGeometryHS*, OGRFeature*, char) (ogrmiramonlayer.cpp:1694)
==2045802==    by 0x5EA3C2D: OGRMiraMonLayer::ICreateFeature(OGRFeature*) (ogrmiramonlayer.cpp:1723)
==2045802==    by 0x610C3C6: OGRLayer::CreateFeature(OGRFeature*) (ogrlayer.cpp:786)
==2045802==    by 0x63DA76A: LayerTranslator::Translate(OGRFeature*, TargetLayerInfo*, long long, long long*, long long&, int (*)(double, char const*, void*), void*, GDALVectorTranslateOptions const*) (ogr2ogr_lib.cpp:6466)

@rouault I've fixed it here: #9966