[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

make package-recursive



in message
<20040823_(_dot_)_231846_(_dot_)_7ede85c647d0486f_(_dot_)_10_(_dot_)_0_(_dot_)_3_(_dot_)_20_(_at_)_bugsgrief_(_dot_)_net>, wrote
horio shoichi thusly...
>
> On Mon, 23 Aug 2004 09:59:52 +0700 "User &"
> <pirat_(_at_)_access_(_dot_)_inet_(_dot_)_co_(_dot_)_th> wrote:
>
> > there should some way to tell make that if ports have been made
> > package, the next time that ports should not have been made
> > again in the make package-recursive from some other ports.

Concur.


> Unfortunately, no. Change CFLAGS and remake, change some of make
> options and remake, ... They give (for the most part) single same
> package name.

So, package-recursive is useful, rather not wasteful, only if a port
depends on previously un-packaged ports.


> This problem (no straightforward way to indicate what exactly the
> outcome is) is inherent even in compilations (or file naming
> conventions we have).  Since make package has undergone build
> process, and doesn't know if it is being repackaged with the same
> set of files, it probably took the safest path, which is the right
> thing, I guess.

How about use of digest checksum(s) stored in a port's file to
compare w/ that of port's package before start of package-ing?

If package name is different or the checksums don't match (say,
different options, CFLAGS & such), package should be built.


  - Parv

--