Quote:
Originally Posted by DNSB
Of course, in the spirit of Kovid working around bugs in QT, etc., have you no desire to work around a bug in calibre? 
|
I already do, to some extent; I manually patch the source code in my machines’ dev environments to use my preferred cover page code. I’m tempted to try tracking down a long-time bug that Kovid doesn’t see as one, though: the “depth” value calibre calculates for NCX files is too high by one.
(Kovid’s explanation is that the value is not allowed to be zero, as it would be for an empty NCX, thus he sets empty to 1, a single level to 2, etc. My rebuttal was to point out that an NCX is not allowed to be empty in the first place; at least one node is required by the spec. He responded that an empty NCX doesn’t break on any known readers, so there. Neither of us is going to persuade the other, and it’s his software, so he wins.)