Bruce Momjian <bruce@momjian.us> writes:
> Tom Lane wrote:
>> Actually, I believe that the *main* problem with pgrminclude is that
>> it fails to account for combinations of build options other than those
>> that Bruce uses. In the previous go-round, the reason we were still
>> squashing bugs months later is that it took that long for people to
>> notice and complain "hey, compiling with LOCK_DEBUG no longer works",
>> or various other odd build options that the buildfarm doesn't exercise.
>> I have 100% faith that we'll be squashing some bugs like that ... very
>> possibly, the exact same ones as five years ago ... over the next few
>> months. Peter's proposed tool would catch issues like the CppAsString2
> The new code removes #ifdef markers so all code is compiled, or the file
> is skipped if it can't be compiled. That should avoid this problem.
It avoids it at a very large cost, namely skipping all the files where
it's not possible to compile each arm of every #if on the machine being
used. I do not think that's a solution, just a band-aid; for instance,
won't it prevent include optimization in every file that contains even
one #ifdef WIN32? Or what about files in which there are #if blocks
that each define the same function, constant table, etc?
The right solution would involve testing each #if block under the
conditions in which it was *meant* to be compiled.
regards, tom lane