https://bugs.kde.org/show_bug.cgi?id=497977
--- Comment #15 from Paul Floyd <pjfl...@wanadoo.fr> --- (In reply to Vincent Lefèvre from comment #14) > This is still confusing, because with an unsigned size_t, a negative value > is not possible. > Or you should say: "possibly a negative value converted to size_t". You're still thinking of this from the perspective of the malloc implementation. Users don't care about that. They just call malloc. Thanks to implicit conversion the user isn't even limited to just positive and negative values. malloc will happily take floating point and complex and imaginary numbers as well. None of those are likely, the point I'm making is that just about anything will get implicitly converted. Above all ***we have no idea what the user wrote in their call to malloc***· All that we are saying is that it is fishy and possibly signed negative. I just had a look on GitHub for issues related to this message. One bug, the fix was to correctly initialize a flag and I can't follow the connection between that flag and the error One testcase Two look like a junk values One using a char for size which probably wrapped then then was implicitly converted One due to wrapping on a 32bit system Another one due to wrapping, using an int instead of a long A junk value due to reading from an outdated json format Wrong value due to string not null teminated Another problem with wrapping The majority look to me like signed integer values with a problem related to wrapping (initial type too small). -- You are receiving this mail because: You are watching all bug changes.
