<div dir="ltr">This would substantially improve our user experience with PETSc. Is MPI consistently tested before other external packages?<div><br></div><div>A</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Fri, Apr 4, 2014 at 1:15 PM, Jed Brown <span dir="ltr"><<a href="mailto:jed@jedbrown.org" target="_blank">jed@jedbrown.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
*******************************************************************************<br>
UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details):<br>
-------------------------------------------------------------------------------<br>
Must give a default value for known-mpi-shared-libraries since executables cannot be run<br>
*******************************************************************************<br>
<br>
<br>
Matt, is there something consistent and maintainable that we could do so<br>
that all the checks for issues like this run immediately, before<br>
compilation tests, so that users can get error messages in a reasonable<br>
amount of time? Few things are more frustrating than waiting so long to<br>
get a trivial error message, then start the entire process over again.<br>
<br>
And/or, what about caching results so that re-running after adding this<br>
flag would be fast? (It scares me too, but I consider configure time to<br>
be a major usability issue.)<br>
</blockquote></div><br></div>