On Monday 07 May 2007 18:08, Davide Viti wrote: > > - Should not first a "cleaner" solution be implemented for the fact > > that the width of the sigils varies? I think subtracting a position > > currently only "works" because most times the sigil is not 2 > > positions while we do now reserve 2 positions for it. This function > > would be the correct place to implement that. > > there's quite a lot of code duplication in newt.c when it comes to > computing window sizes. sigill length is not the only variable: title > length must be taken into account as well (and there are various > subcases depending on the longest of the lengths of the strings inside > the window) No, what I mean is that we currently reserve 4 positions for the sigil, whatever the actual width of the sigil is (and it is often only 3). This could well be the real reason that the imbalance occurs in the first place. So IMO we should should first make sure we correctly calculate the width of the title before we start hacking in fixes in other places. As you have now also shown that other imbalances occur with fixes in this place in the code, I still feel this patch should not be applied as it is now.
Description: PGP signature