On 04.08.2011 11:07, Julian Foad wrote: > Peter Samuelson wrote: >> [Peter Samuelson] >>> (It's news to me, in fact, that it was even possible to build a >>> recent version of neon in such a way that it doesn't support the >>> features we use.) >> >> Followup from IRC: we determined that the reason Doug's neon library >> didn't support XML functions is, he used 'configure --without-webdav'. >> If you don't say --without-webdav, the neon configure script will >> indeed find an appropriate XML library or die trying. >> >> Now ... it would be possible to detect, explicitly, that neon was built >> --without-webdav, and tailor our error appropriately. But it'd be ugly >> code. Is it worth it? Do other users run into this problem? > > I'd say, don't look for this specific problem, just change our error > message to be more honest. 'configure' has already printed > > "checking neon library version... 0.29.6" > > and the problem with the current error message > > "cannot find Neon" > > was simply that it contradicts the earlier message. So, on failure to > compile and link a test program, just make the message say something > like > > "failed to compile and link a test program" > > instead. That's enough to make the user look at 'config.log' for more > details.
<bikeshed> checking neon usability... no </bikeshed> This is similar to the standard header checks, e.g. checking stdio.h usability... yes checking stdio.h presence... yes checking for stdio.h... yes Regards, Rainer
