This is the mail archive of the
mailing list for the GCC project.
Re: (Really, summarizing remaining warnings)
- To: robertlipe at usa dot net
- Subject: Re: (Really, summarizing remaining warnings)
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Thu, 11 Mar 1999 13:00:26 -0500 (EST)
- Cc: egcs-patches at cygnus dot com, law at cygnus dot com
> From: Robert Lipe <firstname.lastname@example.org>
> Kaveh R. Ghazi wrote:
> > 2a. 50-70 missing initializer
> > 2b. 50-70 (near initialization for `???')
> > (For some reason this became a two line warning recently so
> > the warn_summary script counts it twice.) Anyway, these all (or 98%)
> > appear in toplev.c and are due to the definitions of the
> > TARGET_SWITCHES and TARGET_OPTIONS macros. One simply needs to add
> I did this for i386 a few months ago but intentionally didn't add
> strings for flags that weren't in *.texi to call attention to the
> missing doc.
How about fixing the texi file then? :-)
> > 1. ~100 `???' might be used uninitialized in this function
> > I would welcome volunteers for 1 and 3.
> Do we think these are now all "real" and not just GCC being unable
> to see that it's untrue?
There will always be some false positives, because gcc is
documented as unable to detect some cases. Some of the completely bogus
ones have been fixed. I don't know if the false positives we get today
are *only* the documented cases, or we still see some undocumented false
Kaveh R. Ghazi Engagement Manager / Project Services
email@example.com Qwest Internet Solutions