On Sun, Oct 24, 2010 at 3:15 AM, julien <[email protected]> wrote: > Hi, > > I'm running cppcheck (1.45 from sourceforge since the testing package on > Debian repository is the 1.44 version) the whole project. > And i found these points : > [./build/libreoffice-3.2.99.2/berkeleydb/unxlngi6.pro/misc/build/db-4.7.25.NC/btree/bt_compact.c:1980]: > (error) Array index -1 is out of bounds > => > epg = &cp->csp[-1]; > > [./build/libreoffice-3.2.99.2/berkeleydb/unxlngi6.pro/misc/build/db-4.7.25.NC/btree/bt_delete.c:624]: > (error) Array index -1 is out of bounds > => > for (epg = &cp->csp[-1]; epg >= cp->sp; epg--) { > > Are these bugs false positive or just typos ? I never saw an index to -1 but > i don't know at all C++.
1/ the report is in berkleydb, which is an external library. So either it is a false positive, or the problem need to be reported upstream to the berkley db project. 2/ a cursory look at the code lead me to the conculsion that a/ -1 is hard coded there, and it is used as such in many places in that source file. It is unlikely a typo or a un-intended pre-compiler consequences b/ csp sound like 'current stack pointer' to me... and -1 would make sens for that kind of object. So I'd say this is most likely a false positive. Norbert > _______________________________________________ > LibreOffice mailing list > [email protected] > http://lists.freedesktop.org/mailman/listinfo/libreoffice > _______________________________________________ LibreOffice mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/libreoffice
