Vincent van Ravesteijn - TNW wrote:
If this does the job for others, then someone can commit. I won't have
time until late tonight.
I have backported it to branch, and I can commit. I'd like to hear Vincent's
opinion first, though, since his approach (removing updateMacro calls) was
quite different.
Well, both should be fixed, the logic in updateLabels is bypassed by this call
to updateMacros in setParent.
I'm a bit sceptical that someone did the effort to write a FIXME and to change updateLabels according to this, just not to implement a for-loop of one line of code. Why ? Maybe we miss something and we introduce another bug.
By the way, doing both like in Richard's patch would always be ok I guess.
I strongly suspect that with my patch in place almost every other call
to setBuffer() can be removed. I'm going to put the most radical
possible fix in my local tree and see if I get a crash. Then we'll know.
rh
