Robert Billings wrote:
> I tried yesterday to compile 1.4_ALPHA, it was going quite well till close 
> to the end.
> (I hope it was near the end, it was over nine hours into compiling. Yes I 
> know that
> my LC II is slow.) Anyway the last instruction and the following error I 
> got are as follows:
>  (sorry about any typos, and weird spacing)
> ------------------------------
> cc -02 -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes 
> -Wpointer-arith -msoft-float -I. -I../../../../arch -I../../../..nostdinc 
> -DHZ=3D"0x3c" -DM68040 -DM68030 -DM68020 -DFPSD -DFPU_EMULATE -DLKM 
> -DDIAGNOSTIC -DNS Dmaxusers=3D16 -D_kernel -Dmac68k -c 
> ./../../../nfs/nfs_serv.c
> cc1: warnings being treated as errors
> ./../../../nfs/nfs_serv.c: In function 'nfsrv_statfs':
> ./../../../nfs/nfs_serv.c:3113: warning: 'tval' might be used 
> uninitialized in this function.
> *** Error code 1
> stop
> ------------------------------
> 
> I haven't yet had a chance to look at the offending function to try and 
> figure out
> what the problem is. Like I said, it took nine hours to get that far. I'm 
> not rushing
> back into this untill I'm reasonably sure that it will compile.
> 
> Has anyone else seen this?

no, but i think that the -current build environment has a slightly
different set of compiler options.  try rerunning the above compile line,
but add a -Wno-unitialized in the -W section.  that should turn off
warnings about unitialized variables.  if you want to fix it, you can
probably edit the file in question and give the variable some initial
value (like 0 or null, depending on what 'tval' is).

i hope this helps.

later.

colin