Re: design of SUBST_STAGE=post-install behavior?

[brook%biology.nmsu.edu@localhost (Brook Milligan)]

OBATA Akio writes:
 > I did not test, but like a following patch?

Something like that would certainly clean up the cookie in just the
post-install case.  That would solve my particular problem (for now I
have essentially done the same thing as a post-install target hack in
the package makefile, but that seems wrong).

However, there remains the broader question of what the behavior in
this case is supposed to be.  Should the post-install stage be special
in this regard?  Was there a specific intention that all stages be
handled identically, even though there is this potential problem?  Has
this case been considered before?

Cheers,
Brook