[PATCH] PR19515, driver crash while handling a clang crash

[Robinson, Paul]

r207454, thanks!
--paulr

> -----Original Message-----
> From: Chad Rosier [mailto:mcrosier at codeaurora.org]
> Sent: Monday, April 28, 2014 3:09 PM
> To: Robinson, Paul
> Cc: cfe-commits at cs.uiuc.edu
> Subject: Re: [PATCH] PR19515, driver crash while handling a clang crash
> 
> LGTM.
> 
> > After a compiler crash (which prints the traceback and so forth),
> > Driver::generateCompilationDiagnostics calls BuildInputs to get a
> > list of inputs to pass to the preprocessor.  BuildInputs runs through
> > the command-line arguments.  If the original command line included
> > "-lstdc++" this was translated into "-Z-reserved-lib-stdc++" which
> > is marked with options::LinkerInput, so BuildInputs includes it in
> > the list of inputs (treating it kind of like an object file).
> >
> > generateCompilationDiagnostics then wants to filter out inputs the
> > preprocessor doesn't care about.  The first thing it does is check
> > for stdin, like so:
> >
> >   if (!strcmp(it->second->getValue(), "-"))
> >
> > Sadly, "-Z-reserved-lib-stdc++" is Value-less.  The Values are a
> > SmallVector, which in this case is empty, so attempting to retrieve
> > one goes boom.
> >
> > There's another test immediately afterward, which is
> >
> >   if (types::getPreprocessedType(it->first) == types::TY_INVALID)
> >
> > and swapping the order of those tests makes the problem go away.
> > --paulr
> >
> > _______________________________________________
> > cfe-commits mailing list
> > cfe-commits at cs.uiuc.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
> >
> 
> 
> --
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
> hosted by The Linux Foundation
>