Re: RFA: Prevent an ICE when redeclaring a static function as weak

[Jeff Law <law at redhat dot com>]

On 02/22/2016 02:49 AM, Nick Clifton wrote:
> Hi Jeff,
>
> > >     Redefining a previously defined static function as both public and
> > >     weak triggers an ICE in ipa-visibility.c:
> > >
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49899
>
>
> > Does this fix a regression?
>
> How far back am I allowed to go ? :-)
10 years?


> > Essentially I'm trying to figure out if we want this for gcc-6 or gcc-7.
>
> Technically I guess the patch ought to be for gcc-7.  It just feels wrong
> to have a patch to fix a known bug, but not to be able to apply it because
> the bug has always existed...
The development stages are essentially a method by which we reduce churn 
in the source tree as we prepare for a release.  Reality is the more 
changes that are made, the greater the chance of introducing new bugs.

So we go from "free-for-all" development, to a very loose "fix any bug 
you can find" to "just regression fixes" to "just documentation fixes". 
 Each step in that process reduces the amount of changes we're willing 
to make and thus reduces the chances for introducing a new bug and 
increases our chances to get a release out the door on-time.

At this stage, a bug that has been in the source tree, for several years 
is typically not going to be considered all that important (it has after 
all been sitting for years and doesn't appear to be causing a large 
number of problems given the lack of BZ activity).

My inclination would be to defer to gcc-7.  Richi, Jakub or Joseph, as 
release managers, would have the final say though.

jeff