[petsc-dev] Pushing non-working code

[Jed Brown]

On Sat, Feb 2, 2013 at 6:34 PM, Anton Popov <popov at uni-mainz.de> wrote:

> Have you ever tried to compile PETSc with -Wconversion option (or similar)
> switched on?
>
> I just tried it recently with our code, and it completely spammed compiler
> output by the stuff originating from PETSc headers.
> I find this option really useful, because it already helped me to find
> truncated conversion bugs.
> However, enormous volume of PETSc warning makes my job really tough to
> find that couple of lines that my code triggers.
> Yes, I know, there is grep, but I still want to use my IDE.
>

I set up -Wconversion for one of my projects (years ago) and needed to add
a couple GCC diagnostic ignored "-Wconversion" pragmas. (Use C99 _Pragma so
you can put it in a macro.) I'm not sure if it's feasible to use this
internally in PETSc, but I'll have a look.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130202/eee8cbaa/attachment.html>