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

Re: 3.10.0 FTBFS



On 30-08-2007, Samuel Mimram <samuel.mimram@ens-lyon.org> wrote:
> Hi,
>
> Mike Furr wrote:
>> It looks like OCaml 3.10 is failing to build on the non-native arches. I
>> think I tracked down the problem and have reported it upstream[1].
>> However, I don't have time to make/test a patch for svn at the moment.
>> One quick fix should be to just change the variable mentioned in the bug
>> report to default to 'gcc' for any linux arch.
>
> Please also notice that we have a funky (and apparently unrelated) build
> error on arm:
>
> boot/ocamlrun ./ocamlc -nostdlib -c -g -w a -I camlp4/boot -I
> camlp4/build -I utils -I parsing -I camlp4 -I stdlib -o
> camlp4/boot/Camlp4.cmo camlp4/boot/Camlp4.ml
> Exit code 255 while executing this command:
>   boot/ocamlrun ./ocamlc -nostdlib -c -g -w a -I camlp4/boot -I
> camlp4/build -I utils -I parsing -I camlp4 -I stdlib -o
> camlp4/boot/Camlp4.cmo camlp4/boot/Camlp4.ml
> make[2]: *** [camlp4out] Error 255
>

I take a closer look to the situation. It seems to stop compiling on a
very big file (18kLOC of ocaml) without any warning. Moreover 255 is not
a classical exit code. Most of the time exit code over 128 are exit code
combined with a signal interruption...

My opinion: the compilation of this file take too much time on arm and
the process got interrupted. 

After asking on IRC, it seems that the buildd watch the output of the
compilation. If nothing happens for a certain time, it kills the
processs. I think, we are in this case!

Any comments?

Regards,
Sylvain Le Gall

ps: i think i have a patch for the 1st FTBFS bug, it stands in the svn 



Reply to: