Peter,

>  > On the other hand, I wonder why this cannot be dealt with in
>  > usr/src/Targetdirs.  I don't really understand which directories are
>  > listed/created there and which are created only by other Makefiles.
> 
> Don't feel bad, I don't think anyone really understands this ;-)

I'm glad that I'm not alone here :-)

> In general, I think the theory is that Targetdirs should create all of the
> directories and that the various Makefiles should not have to.  Among
> other things, this centralized the knowledge of the permissions associated
> with various directories -- though since we stopped requiring root builds,

Btw., it might be beneficial to rip out support for root builds altogether:
this would unclutter the Makefiles considerably.  On the other hand, the
companion consolidation still requires them (as did sfw initially, which I
considered a showstopper).

> the permissions of the directories no longer really matter.  In practice,

They still do, at least for open builds:

Inconsistencies between pkgdefs and proto area:

Unit   T File Name                      Reloc/Sym name       perm owner group 
inode lnk maj min package(s)
------------------------------------------------------------------------------------------------------------
filea: d etc/lp                         -                     775 -     -       
   0  1  -  -    SUNWpsr
fileb: d etc/lp                         -                     755 -     -     
122720  9  -  -    proto
    differ: perm 

and a few more warnings.  This problem may be limited to an open build,
where the directories aren't created by Targetdirs, but by copying them
over from the closed tarball.  I tried to fix this by listing them in
Targetdirs with the correct permissions, but haven't fully succeeded yet
since one needs to make sure that copying the closed files/directories
doesn't undo Targetdirs.  It's probably best to remove all those empty
directories from the closed tarball and make sure they are correctly
created by Targetdirs.

        Rainer

-----------------------------------------------------------------------------
Rainer Orth, Faculty of Technology, Bielefeld University
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to