Bill Allombert on Fri, 28 Sep 2012 11:47:18 +0200

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

Re: Please test pari-2.5.3 prerelease 1

On Fri, Sep 28, 2012 at 11:18:54AM +0200, L. Cangelmi wrote:
> However, I point out three minor questions:
> [1] In compilation (make gp) of 2.5.0, 2.5.1, 2.5.2, 2.5.3pre1 and 2.6.0
> on Ubuntu 10.04 (with gcc-4.4.3, with or without gmp-4.3.2) I get the
> following warnings for galois.c (but there were not bugs in test-galois)
> (note also that they were not present for pari-2.3.5 on Ubuntu 9.10 with
> gcc-4.4.1, and are not present for pari-2.5.3pre1 on MacOS X 10.3.9):
> ===
> /usr/bin/gcc  -c -I. -I../src/headers -fPIC -O3 -Wall
> -fno-strict-aliasing -fomi
> t-frame-pointer    -o galois.o ../src/modules/galois.c
> ../src/modules/galois.c: In function ?read_obj?:
> ../src/modules/galois.c:327: warning: ignoring return value of
> ?fread?, declared
>  with attribute warn_unused_result

Hello Leonardo,
Thanks for your tests and comments!

Ubuntu changes the stdio.h header file to add the attribute "warn_unused_result"
to fread for "security" reason. This is a known issue and is relatively easy to
fix but I will not touch galois.c. (This file need to be entirely rewritten anyway).

> ===
> [2] The last line of the output of kerntest of 2.5.1, 2.5.2 and 2.5.3pre1
> on Ubuntu 10.04 (with gcc-4.4.3, with or without gmp-4.3.2) is:
> ===
> gcc bug?:6000000000000000  c000000000000000
> ===
> While for 2.5.3pre1 on MacOS X 10.3.9 it is
> gcc bug?:60000000  c0000000  00000000.
> Is it a warning to take into account?

No, this is the expected value in both case. 'make test-kernel' report OK.

> [3] It seems to me that the file INSTALL.pdf (i.e. appa.tex) should be
> updated: in the section about kernels, we should read that the default
> kernel is auto-gmp (not auto-none).

Actually the default is auto-auto. I will fix it in PARI 2.5.4.
I do not think it is worth the trouble to do a 2.5.3pre2.