> We see to have 8 addtional testsuite failures in our make compared to
> MacPorts. Adding the missing InfoTest to the current make.info....
> Updating the package to 3.82 and disabling the make.patch (since MacPorts uses
> none for 3.82), I get much improved results of...
Those patches used to be absolutely necessary on Mac OS X, as they added support for building Objective C code; omitting them would result in a "make" executable that fails to build lots of OS X specific code. Luckily, in 3.82, upstream finally integrated at least parts of those changes. I am not quite sure why I never picked up 3.82; my guess is that I simply missed the fact that they finally made a new release 4.5 years after 3.81 ;)
Anyway, I just put a 3.82 into CVS. There is a hypothetical chance that it might break a few packages that rely on the other Apple specific features added by that patch -- we'll see. If anybody notices any regressions, let mw know.
As for the remaining two test failures: They are caused by the HFS+ file system which preserves case in filenames, but ignores it for purposes of accessing files. Since those tests are about distinguishing "makefile", "Makefile", etc., it is no wonder they don't work :). It's not a serious issue either, though. I tweaked the test to silence this false failure report.