Hi Marcel, On Wednesday, 2011-12-21 16:18:46 +0100, Marcel Metz wrote:
I've tested the patch with 'make check', but would like to point out that the reviewer should take special care of the change in 'sc/source/core/tool/chgtrack.cxx:3140-3160'. From my understanding of the code the tools Table does not allow multiple entires with the same key so that this assertion for double keys is not required but maybe I'm missing something.
The assertion tells if there was a dup, in that case something was going wrong and memory leaks. Anyway, aparently you were confused by the two different tables/maps and mixing iterator from different map, see http://cgit.freedesktop.org/libreoffice/core/commit/?id=c1db721af6286ae4c3164075337d81e943845938 Note that instead of map.erase(key) you can use map.erase(iterator) if you already obtained one and don't access it afterwards, that saves an internal map.find(key) Other than that and the correction by Michael Stahl (kudos), good patch! Btw, did you check if Calc changetracking and accepting/rejecting actions still works? Thanks Eike -- LibreOffice Calc developer. Number formatter stricken i18n transpositionizer. GnuPG key 0x293C05FD : 997A 4C60 CE41 0149 0DB3 9E96 2F1A D073 293C 05FD
Attachment:
pgpYJK7Afeudv.pgp
Description: PGP signature