Re: Problem with building insight 19991222 Linux RedHat-6.1 -> mips64orionel

[Kai Ruottu <karuottu at freenet dot hut dot fi>]

Andrew Cagney wrote:
> 
> "Alexey S. Fadeyev" wrote:
> >
> > Faild to build!
> > make[2] *** No rule to make target `m16.igen' , needed by `tmp-igen'.
> > Stop.
> 
> That's an interesting looking target name - its new on me and know most
> of them :-)

 The 'mips64orionel' is quite ok, a little-endian 'mips64orion' goes
with
that name, like 'mips64orionel-elf'.
 
> It looks like the sim/mips configury is getting confused and thinks it
> needs to build 64 bit MIPS16 target!

 The NEC 'mips64vr4100' target has the MIPS16 capability (VR4111, VR4121
VR4122, VR4181), so a 64 bit MIPS16 target isn't unexpected...

 Ok, back to the original problem, the missing 'm16.igen' can be the
result
of running 'make distclean'... I remember quite well there being a bug
sometimes, the 'sim/mips/Makefile.in' having a row :

	rm -f m32* m16* itable*

instead of the current :

	rm -f m32* itable*

in the file end...

 Fortunately I had a backup for the 'm16.igen' then... Now I couldn't
find any incarnation of that bug anywhere, neither the 'ChangeLog'
mentioned anything about somebody removing it, but perhaps the
insight-19991222 had this bug... although I remember it being there
long time ago...

 But now there is another problem... All my 'mips64*-elf' Insights
built after November 11, 1999 now freeze when the 'src+asm' will be
clicked.  I have a 'mips64vr5000-elf' target one built at Nov. 11, 1999
and it works ok in this respect, but my 'mips64vr4300-elf' (01/05,
2000),
'mips64vr4100-elf' (01/17, 2000) and 'mips64orion-elf' (01/31, 2000)
have
this bug... It happens both in Linux and Win32 hosts and only with the
mips64 targets...

 So the bug was added to the Insight sources sometime between Nov.11,
1999
- Jan. 05, 2000... I will hunt the bug soon, but perhaps someone else
will
succeed earlier...

Cheers, Kai