file-newer-than-file-p [was: [COMMIT] Check absolute source file ...]

[Stephen J. Turnbull]

Aidan Kehoe writes:
 >  Ar an t-aonú lá is triochad de mí na Nollaig, scríobh Stephen J. Turnbull: 
-------------- next part --------------

 >  > [...] which is precisely what (1) ... (4) give. Do you disagree, or do
 >  > you just think the whole idea of trying to extend the relation to
 >  > non-existent files is bogus and the caller should check explicitly?
 > 
 > Yes, that. In an ideal world #'file-newer-than-file-p would throw a
 > file-error if either argument didn?t exist or the corresponding metadata
 > couldn?t be read. Doing this now would be introducing a needless
 > incompatibility though, it?s much more realistic to live with it.

OK, good.

 >  > [...] BTW, do you agree with me (at #### above) that we want to use the
 >  > .el file, not the .elc?
 > 
 > Well, we have to use one or the other, and using the .elc has the advantage
 > that we can pass the path on to make-docfile without further processing. 
 > make-docfile.c much prefers .elc files, since the docstrings are
 > automatically generated in a format it can read.

OK.  I guess it doesn't really matter, since the only reason the .elcs
would be bogusly newer than the DOC file is if somebody rm'ed some
deliberately.  "make beta-clean" wouldn't be a problem because that
rm's the DOC file, too.