Re: pr19.test

[Akim Demaille]

| From
| http://sources.redhat.com/ml/automake/2001-03/msg00183.html
| 
|     2) lines 4946-4947 of automake.in version 1.977
| 
|     # FIXME: nodist.
|     &push_dist_common ($pfx . $base . '.' . $ext);
| 
|     commenting out the last line removes foo.c (which is a temporary file)
| from the make distdir target, which fails because according to automake,
| foo.c is supposed to be redistributed. however, (gnu) make removes temporary
| files like foo.c (generated from foo.l). in my opinion, this is correct and
| should not be redistributed. the Makefile.am writer can always explicitly
| include foo.c if they really want to. from the comment in the automake.in
| file, it seems the automake peeps are aware of this. there are similar
| comments in automake.in with other generated files as well.
| 
| Here is the context:
| 
| foo.o: foo.l
| 
| make has enough information to convert foo.l -> foo.c (via lex/flex)
| then compile foo.c to foo.o. However, since foo.c is an intermediate file
| it is deleted.
| 
| automake adds foo.c to the list of dist files though, via push_dist_common.
| from the comments in the automake source, it appears this is already known,
| but i don't know if anyone plans on doing anything about it. by the way,
| same
| issue for other auto-generated files. just search automake.in for "FIXME:
| nodist"

Thanks for the details.  As I'm currently working on DIST_COMMON
etc. (I have another bunch of patches waiting at home), I'm not
inclined to fix this now (collisions...).

But still, I fail to understand the relationship between your
explanation and the failure you observe.  In what way is this related
to Cygwin?  _That_ is the real question.  The scenario you depict
seems to imply that Automake should have pr19 fail everywhere, which
is not the case.  So why do _you_ observe a failure?