Re: [RFA] dwarf2read.c: complain() -> complaint()

[Jim Blandy <jimb at redhat dot com>]

Andrew Cagney <ac131313@redhat.com> writes:

> > In a recently submitted patch, I added a new complaint to dwarf2read.c
> > using the old (deprecated) interface.  Andrew asked me to use the new
> > interface.  I've generated a patch for dwarf2read.c which converts all
> > calls to complain() to complaint().
> > I've written a script to do this transformation.  It finds all of the
> > deprecated_complaint structs and counts the uses of these
> > structs. (Surprisingly, some were unused!)  For cases where there's
> > more than
> > one occurrence, it creates a new function as suggested in
> > complaints.h.  For the rest, it performs the obvious transformation.
> > The script also uses GNU indent to perform localized reindentations
> > of
> > the affected text.
> > If this patch is accepted (and this approach is deemed acceptable),
> > I'll
> > generate patches for the other files which still use complain().
> 
> Definitly fine with the theory.  Two reservations (really just one).
> 
> - how is the result with -Wformat?  The reason behind switching from
> complain() to complaint was to get the parameters checked and hence
> find some nasty address printing bugs.
> 
> - I suspect it will need a visual audit to check for any cases of:
> 	"%08lx", (long) core_addr_variable
> 
> Looking through this specific patch, though, turned up no cases where
> this occured. So you must have that those problems covered.
> 
> Symtab maintainers?

It looks fine to me.

I was uncomfortable with the idea of having to create wrapper
functions at first --- the idea being that each occurrence of a
complaint string is its own independent complaint --- but it seems
like it'll do the right thing by default more often than making people
create complaint structures.  So I'll go with that.