[cfe-commits] r159875 - in /cfe/trunk: include/clang/Basic/DiagnosticGroups.td include/clang/Driver/Options.td lib/Driver/Tools.cpp lib/Frontend/TextDiagnosticPrinter.cpp test/Driver/warning-options.cpp test/Misc/show-diag-options.c test/Misc/war

[Ted Kremenek]

On Jul 6, 2012, at 10:01 PM, Richard Smith <richard at metafoo.co.uk> wrote:

> …but it's arguable whether that really counts, since it's likely many of these warnings should have actual flags anyway…
> 
> ... and indeed we no longer have any protection against extension warnings being added with no explicit flag (other than -Wpedantic, which doesn't seem fine-grained enough to be useful). Would it be worthwhile to teach diagtool to additionally list diagnostics which are contained directly in the 'pedantic' group and include that in our CHECKs, with the intention of eventually emptying that list too?

Agreed 100%.  Jordan added 'diagtool tree', which allows us to monitor the transitive closure of stuff under -Wpedantic.  We could write a FileCheck test based on that output.  Something more specific might be useful tool (and include that under list-warnings).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120706/71cd12fb/attachment.html>