Re: STL and -Weffc++ (was: Bug in optimizer...)

["Matt Austern" <austern at isolde dot mti dot sgi dot com>]

I suppose my question would be whether the warnings are reasonable
(i.e. they're pointing out potential errors) or whether they simply
represent someone's not-univerally-shared stylistic opinion.

If the former, then we should probably rewrite the code to get rid of
them.  If the latter, then probably the best approach is to put in a magic
pragma that turns off warnings for the duration of the header.  You
might notice that we have, in fact, done that for a few of the remarks
that SGI's own compiler gives with the -fullwarn option.

			--Matt

--- Forwarded mail from Jason Merrill <jason@cygnus.com>

To: pfeifer@dbai.tuwien.ac.at (Gerald Pfeifer), egcs-bugs@cygnus.com
Cc: austern@sgi.com
Subject: Re: STL and -Weffc++ (was: Bug in optimizer...)
From: Jason Merrill <jason@cygnus.com>
Date: 28 Jan 1998 11:47:37 -0800

>>>>> Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> writes:

> On Wed, 28 Jan 1998, Peter Leven wrote:
>> [...] the standard template library does not pass the -Weffc++ tests.
>> This makes it rather annoying to use the option with one's one code,
>> since one must find the messages relating to one's own code in between
>> the messages for the standard library.

> Indeed that is _quite_ annoying.

> Jeff (et al.), are you willing to accept patches that fix this?

No, it's not worth diverging from the SGI sources.  Also, some of the
effc++ warnings cannot be avoided without unreasonable changes.

> We could also Cc: those patches to someone@sgi.com (any appropriate
> address, anyone?) to get it fixed in the long term, but for egcs this
> should be fixed ASAP, IMHO.

Matt Austern <austern@sgi.com> is the guy I've been dealing with.

Jason


--- End of forwarded mail from Jason Merrill <jason@cygnus.com>