Current-Users archive

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

Re: Build failures on kernel version changes



On Sun, Jan 15, 2012 at 03:06:45PM +0100, Nicolas Joly wrote:
 > > > Except that both gcc and binutils explicitely disallow passing user
 > > > specified cache file to subconfigures, except for special value
 > > > /dev/null (which is the default for top-level configure and disable
 > > > caching).
 > > 
 > > Gah!
 > > So let's just check for it and remove all config.cache files when it
 > > happens.
 > 
 > Do you want to remove all caches when configure is rerun (for whatever
 > reason) ? Or only for os version bump ?
 > 
 > In the former case, it makes the cache files useless; they are only
 > used when configure is rerun. As for the latter, how do we easily
 > detect that cache files were generated from an different version ?

Aren't they also used during configure? A lot of configure scripts do
inane things like

   Checking for socket -lsocket... no
     :
   Checking for socket in -lsocket... (cached) no
     :

although I don't know if gcc and binutils do.

(also, is there a reason they block user-specified cache files? if
not, can that be altered easily?)

-- 
David A. Holland
dholland%netbsd.org@localhost


Home | Main Index | Thread Index | Old Index