Bug#988986: cppcheck: Symbol `_ZTVN8tinyxml210XMLPrinterE' has different size in shared object, consider re-linking
Package: cppcheck Version: 2.3-1 Severity: important Dear Maintainer, running cppcheck prints the following message: Symbol `_ZTVN8tinyxml210XMLPrinterE' has different size in shared object, consider re-linking Also reproducible with cppcheck version 2.4 from experimental. -- System Information: Debian Release: 11.0 APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 5.10.0-7-amd64 (SMP w/4 CPU threads) Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8), LANGUAGE=en_GB Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) LSM: AppArmor: enabled Versions of packages cppcheck depends on: ii libc6 2.31-12 ii libgcc-s1 11.1.0-2 ii libpcre3 2:8.39-13 ii libstdc++611.1.0-2 ii libtinyxml2-8 8.1.0+dfsg-1 ii libz3-4 4.8.10-1 ii python3 3.9.2-3 ii python3-pygments 2.7.1+dfsg-2 cppcheck recommends no packages. Versions of packages cppcheck suggests: pn clang pn clang-tidy pn cppcheck-gui -- no debconf information
Bug#988986: cppcheck: Symbol `_ZTVN8tinyxml210XMLPrinterE' has different size in shared object, consider re-linking
reassign 988986 tinyxml2 severity 988986 serious retitle 988986 ABI change in tinyxml2 8.1.0+dfsg-1 found 988986 8.1.0+dfsg-1 thanks Hi Chow, it seems that there's an ABI change in tinyxml2 8.1.0+dfsg-1. cppcheck 2.3-1 was uploaded to unstable on 2020-12-17, compiled against tinyxml2 8.0.0+dfsg-2, and migrated to testing. On 2021-05-20, tinyxml2 8.1.0+dfsg-1 was uploaded to unstable. Running "touch foo.cpp; cppcheck foo.cpp" works as expected in testing, but shows "_cppcheck: Symbol `_ZTVN8tinyxml210XMLPrinterE' has different size in shared object, consider re-linking" in unstable. $ echo _ZTVN8tinyxml210XMLPrinterE | c++filt vtable for tinyxml2::XMLPrinter Indeed, the diff both between versions of tinyxml shows that the vtable of XMLPrinter has been extended. Looks to me as this ABI change requires at least a library transition. Thinking a bit more about it ... This also changes the vtable of derived class (XMLPrinter is not final) in an incompatible way , which can be seen with the example in the contrib folder. Building against 8.0.0+dfsg-2 and running with 8.1.0+dfsg-1 leads to a crash. I wonder why upstream did not bump the SOVERSION. Best regards, Joachim
