Actually, 39 is not that much, I'd have thought somewhere in the hundreds. That's perfectly doable, if no one wants it, I can when I have some time this week.
sj Ekkor: 2025. december 2. 21:50:08 CET, Andreas Schneider <[email protected]> írta: >I would rather define a USE_PDFVIEW constant as suggested by sj. > >Andreas > > >Am 02.12.25 um 17:44 schrieb Richard Shann: >> On Mon, 2025-12-01 at 22:36 +0100, Andreas Schneider wrote: >>> Wouldn't it be more straight forward to test for USE_EVINCE or >>> USE_ATRIL in the sources? >> >> well there are 39 places where USE_EVINCE occurs in the sources, but >> perhaps if I put >> #ifdef USE_ATRIL >> #define USE_EVINCE 1 >> #endif >> >> into denemo.h this would meet the need. Anyone see any snag with that? > > >Am 01.12.25 um 23:04 schrieb sj: >> I think it would be even better to just replace current USE_EVINCE usage >> with something like USE_PDFVIEW (keeping the original defines for fixing >> differences), but that seems like a pretty troublesome adventure to >> start now that the current usage is so ingrained in the codebase. >> >> I think putting in 'or defined USE_ATRIL' into every USE_EVINCE check is >> about the same burden in terms of workload (and if you mess up, the bugs >> may or may not become apparent at first for whatever weird gcc reason, >> though I'm quite unsure about that – is there any actual behavioral >> difference between the two libs that required #ifdef-ing behavior >> instead of just conditionally defining functions as NOP if no PDF viewer >> was found?). > >
