Re: [Bug-glpk] Segfault in jday function

[xypron . glpk]

Hello Andrew,

so you do not want to support Linux because it is not 100% Posix? Neither all 
those other Unix derivatives that will indicate an error in gettimeofday?

Best regards

Heinrich Schuchardt

http://www.xypron.de

Am 08.07.13 um 06:51 schrieb Andrew Makhorin

> The POSIX standard says:
> 
> 
> 
> RETURN VALUE
> 
>       The gettimeofday() function shall return 0 and no value shall be
> 
>       reserved to indicate an error.
> 
> 
> 
> ERRORS
> 
>       No errors are defined.
> 
> 
> 
> So I see no need to check the return code. Implementation should not
> 
> change the semantics defined by the standard.
> 
> 
> 
> 
> 
> 
> 
> > which errors may occur depends on the POSIX system you use, just compare
> 
> > 
> 
> > http://publib.boulder.ibm.com/infocenter/iseries/v5r4/index.jsp?topic=%2Fapis%2Fgettod.htm
> 
> > 
> 
> > http://www.manpagez.com/man/2/gettimeofday/
> 
> > 
> 
> > http://illumos.org/man/3C/gettimeofday
> 
> > 
> 
> > Best regards
> 
> > 
> 
> > Heinrich Schuchardt
> 
> > 
> 
> > On 01.07.2013 21:54, Andrew Makhorin wrote:
> 
> > > Hi Heinrich,
> 
> > >
> 
> > >> if gettimeofday fails it returns -1,
> 
> > >> if gmtime fails it returns NULL.
> 
> > >>
> 
> > >> To avoid segmentation faults, please, provide error handling as follows:
> 
> > >>
> 
> > >
> 
> > > Thank you for suggestion.
> 
> > >
> 
> > > However, under which circumstances gettimeofday may fail? Most likely
> 
> > > (as I replied to Jiri) that it was a bug specific to mingw64 version of
> 
> > > gettimeofday.
> 
> > >
> 
> > >
> 
> > > Andrew Makhorin
> 
> > >
> 
> > >