audacitysvnlib-srclibnyquistnyquistxlispxlread.c(338) : warning C4715: 'readone' : not all control paths return a value
audacitysvnlib-srclibnyquistnyquistxlispxlbfun.c(620) : warning C4716: 'xcontinue' : must return a value
audacitysvnlib-srclibnyquistnyquistxlispxlimage.c(400) : warning C4715: 'cvoptr' : not all control paths return a value
audacitysvnlib-srclibnyquistnyquistxlispxlbfun.c(613) : warning C4716: 'xtoplevel' : must return a value
audacitysvnlib-srclibnyquistnyquistxlispxlbfun.c(606) : warning C4716: 'xcleanup' : must return a value
A couple of them I could make a stab at but I have no way of testing any changes as I have no idea how to exercise the code.
I brought this here to the forum so that the Devs could stay focused on P2s. It seems the folks who know Nyquist are more likely to see it here anyway. I suspect this has been addressed in some newer version of the code as I think these would be fatal errors in some (newer) compilers. I could “fix” the problems (at least get rid of the warnings) but have no idea how to test the repairs. About half of them are pretty obvious as to “a” fix, but is it right–only a Nyquist guru could tell/test.
As for me adding bugs to Bugzilla, I personally feel that only active Devs should be adding bugs. Problems seen/discussed here on the forum or via @feedback should be triaged by someone knowledgeable and only added where appropriate. As the old saying goes, “I have just enough knowledge to be dangerous”!
At the same time, I find a heavily tweaked Audacity to be extremely useful and will do what I can to “pay forward” by testing and (very) rarely offering code privately to a Dev who might be interested.
Speaking of compiler warnings, as an aside, it would not surprise me if some, if not quite a few, of the moonphase bugs would quietly disappear if ALL the (thousands) of warnings were addressed. I know this would be a monumental effort!
Bugs are already triaged before adding to Bugzilla in the sense that only known people (including you) have the rights to add bugs. However Bugzilla was set up as the bugs database so I strongly feel it should be used as such, so that there is only one place to look for acknowledged bugs (which compiler warnings are), rather than have cases where bugs might only be on -devel, or -quality or in the Forum). So the best approach IMO would be to add to Bugzilla and link to the report here (possibly on the -nyquist list as well), asking for comments. If you get no comments, then the bug will soon get lost if it’s only in the the ephemera of Forum discussions.
That’s quite possibly another way for the code to get lost without trace, especially if it doesn’t directly relate to a high priority bug.
Which is why that’s not a productive way to go about fixing them. Code review and clean up targeted to relevant code would be better.