Segmentation fault when printing large numbers

[Marc Glisse]

On Wed, 28 Apr 2021, Jeff Epler wrote:

> The attached program creates a rather large number, then attempts to print
> it.  The printing step encounters a segmentation fault.
>
> Typical error traceback as seen by gdb:
> Program received signal SIGSEGV, Segmentation fault.
> __memmove_avx_unaligned_erms () at
> ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:500
> 500 ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S: No such file
> or directory.
> (gdb) where
> #0  __memmove_avx_unaligned_erms () at
> ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:500
> #1  0x000055555555cb3c in __gmp_asprintf_memory (d=0x7fffffffe2e0,
> str=0x7ffeff624010 "", len=18446744071562067968) at asprntffuns.c:48
> #2  0x000055555555d042 in __gmp_doprnt_integer (funs=0x555555583d80
> <__gmp_asprintf_funs_noformat>, data=0x7fffffffe2e0, p=0x7fffffffe340,
> s=0x7ffeff624010 "") at doprnti.c:125
> #3  0x0000555555556ad0 in __gmp_doprnt_integer_ostream (o=...,
> p=0x7fffffffe340, s=0x7ffeff624010 "") at osdoprnti.cc:60
> #4  0x0000555555556a09 in operator<< (o=..., z=0x7fffffffe3d0) at
> osmpz.cc:46
> #5  0x0000555555556934 in operator<< <__mpz_struct [1], __mpz_struct [1]>
> (o=..., expr=...) at ./gmpxx.h:2181
> #6  0x0000555555556428 in main (argc=2, argv=0x7fffffffe528) at bigg.c:16
>
> It appears that the character-size of strings at some layer
> (__gmp_doprnt_integer?) is restricted to INT_MAX bytes of output.

Using int for the length in that file is certainly a bad sign. Generally, 
formatted I/O hits limitations in GMP, see for instance 
https://gmplib.org/list-archives/gmp-bugs/2015-November/003795.html and 
various other reports. It may be possible to fix some of the printing 
functions without any ABI break, but not others...

-- 
Marc Glisse