[Mesa-users] (no subject)

Warrick Ball W.H.Ball at bham.ac.uk
Tue Dec 18 09:17:43 EST 2018


Hi Samuel,

This is a known issue when compiling older versions of MESA with newer 
versions of the SDK.

My normal recommendation would be to first look up when a given public 
version was released in the FAQ:

     http://mesa.sourceforge.net/faq.html#which-svn-revisions-were-mesa-release-versions

e.g. r7624 was released Jun 3, 2015.  Then, you can hunt down the relevant 
version of the SDK.  The oldest version available is actually 20170802, so 
you can try using that but it might also be too new.

You can work around these uninitialized variables by simply ignoring them 
during compilation.  To do so, add `-Wno-maybe-uninitialized` to `FCwarn` 
in `$MESA_DIR/utils/makefile_header`.  Don't worry: suppressing the 
compiler warnings like this is safe.

Finally, you can also work through the compiler warnings, adding the 
necessary variable initializations.  They're mostly caused by variables in 
conditional blocks where the compiler is worried that MESA hasn't given 
them a value, even though MESA does give them an appropriate value if it 
uses those variable.  Initializing them to 0 (or whatever) at the 
beginning of the relevant subroutines/functions should work (without 
affecting MESA's results).  You can also look at more recent versions of 
MESA to see how the main codebase has been changed to satisfy these 
requirements for newer versions of the compiler.

Cheers,
Warrick

------------
Warrick Ball
Postdoc, School of Physics and Astronomy
University of Birmingham, Edgbaston, Birmingham B15 2TT
W.H.Ball at bham.ac.uk
+44 (0)121 414 4552


On Tue, 18 Dec 2018, mesa-users at lists.mesastar.org wrote:

> Dear MESA-Users,
>
> I am trying to install version 7624 of MESA (in part to reproduce this paper 
> http://adsabs.harvard.edu/abs/2016ApJ...823...46F),
> There doesn't seem to be a specific version of the MESASDK available for this 
> version so instead, I tried to install it on Linux using the current version 
> of the SDK but that fails with the error;
>
>                       rpar(r_burn_const_P_time_eos) + dble(time1 - time0) /
>     clock_rate
>    ^
>    Error: 'time0' may be used uninitialized in this function
>    [-Werror=maybe-uninitialized]
>    f951: all warnings being treated as errors
>    make: *** [net_burn_const_p.o] Error 1
>
> How might I fix this in order to replicate the paper or is there a different 
> version of the SDK available?
>
> Thank you,
>
> Samuel Lloyd
>
>
> Physics PhD student
> University of Edinburgh
>
> -- 
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>
> _______________________________________________
> mesa-users at lists.mesastar.org
> https://lists.mesastar.org/mailman/listinfo/mesa-users
>



More information about the Mesa-users mailing list