stage 1 bootstrap fail on Debian armv7

[Dennis Clarke]

On 12/27/18 3:22 AM, Marc Glisse wrote:
> On Thu, 27 Dec 2018, Dennis Clarke wrote:
> 
>> I was surprised to see this sort of thing in stage 1 of a bootstrap
>> of gcc 8.2.0 where the local host gcc was 6.3.0-18+deb9u1 :
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86724
> 
>> Well that is interesting and perhaps I should try to not leap from 6.3.x
>> all the way to 8.2.0 ?  I'll give 7.4.0 a shot but curious if anyone has
>> seen anything like the above.
> 
> Please search bugzilla and try a fresh checkout of trunk (or
> gcc-8-branch) to see if things are already known.
> 

right right right ... this is all very familiar now .. dumb me ...
I should have checked my own notes from months ago when I ran into this
exact issue :

beta $ diff -c ./gcc/graphite.h_bugid_86724 ./gcc/graphite.h
*** ./gcc/graphite.h_bugid_86724        Wed Jan  3 10:03:58 2018
--- ./gcc/graphite.h    Fri Oct 19 18:51:27 2018
***************
*** 37,42 ****
--- 37,44 ----
   #include <isl/schedule.h>
   #include <isl/ast_build.h>
   #include <isl/schedule_node.h>
+ #include <isl/id.h>
+ #include <isl/space.h>

   typedef struct poly_dr *poly_dr_p;




see https://www.mail-archive.com/gcc-bugs@gcc.gnu.org/msg581075.html



Okeee dookee ... easy fix.

Thanks for that and I'll go back to watching this little arm7 ASUS
TinkerBoard baking its little rockchip to 48C.

Dennis