Compile error on current egcs_1_00 branch

Mark Mitchell mmitchell@usa.net
Mon Apr 20 12:06:00 GMT 1998


>>>>> "Andreas" == Andreas Jaeger <aj@arthur.rhein-neckar.de> writes:

    Andreas> Please add this for egcs1.0.3.

Somehow, without any indication from Jeff, we seem to have decided
that we're building an egcs-1.0.3.  (Jeff and others, you're welcome
to correct me if Jeff has indeed indicated that we're building 1.0.3.)

I'm strongly against building a 1.0.3.  The effort required for us to
roll out a subsequent 1.0 release will be substantial.  In addition,
we will again begin haggling over which mainline changes are
sufficiently important to make it into 1.0.3.

I, for one, am beseiged by bug reports about template problems in
1.0.2 that have been fixed months ago on the mainline.  However, I
don't propose putting these into 1.0.2; instead, I suggest that we
soon begin work on 1.1, which will contain *all* the mainline
improvements.   The right thing to do is to branch soon for 1.1, and
then to begin the effort of stabilizing 1.1.1.  

The argument seems to be that 1.0 is stable, and so it's safe to
make minor changes to it to fix our own pet peeves.  But, 1.0 only
attained this stability after the considerable testing and bug-fixing
efforts we all put into it *after* it was branched from the mainline.
We have to bite the bullet and to the same for 1.1.

Our resources are scarce, and Jeff's are probably scarcer.  Let's put
1.0 to bed, and get to work on 1.1.  Our user community will be much
more impressed with us when we release a 1.1 that fixes many C++
problems, improves code generation in a variety of ways, and so forth
than they will be if we produce another minor incremental improvement
over the status quo.

-- 
Mark Mitchell <mmitchell@usa.net>
http://home.earthlink.net/~mbmitchell
Consulting Services Available



More information about the Gcc-bugs mailing list