Fixing gcc 3.3 compile failures

Kris Kennaway kris at
Thu Jul 17 20:15:58 PDT 2003

OK, now that the latest 5.x package build is well underway, we can
start work on fixing the compile failures seen with gcc 3.3.

These are the ports that have become broken on the latest build
(everything after July 14):

Here is the full list of broken ports from the build in progress:

If you are running 4.x, you can also help to develop fixes for these
ports by installing the gcc33 port and setting


in your environment.

Most of the new compile failures are caused by 3 or 4 types of failure
mode (all of which have to do with stricter standards compliance in
the new compiler suite).  I haven't yet looked at how to fix most of
them: if you figure out a patch for a class of failures, please post
it in response to this email so we can all see how to do it.

To start things off:
... error: `assert' undeclared (first use this 

This class of error is fixed by adding 

#include <assert.h>

to the top of the file.

NOTE: There are at least two recent failure modes that have been
introduced into 5.x that are not due to gcc33.

1) The recent texinfo port changed the default info generation style
to not produce split files, which breaks a lot of pkg-plists.  A patch
was submitted to the ports@ list a few days ago to allow
to better handle info files, but I haven't yet evaluated or tested it.
I hope to have the time over the weekend.

2) Recent *.mk changes by ru have apparently broken a number of ports:
nm: ../generic/panic.o: No such file or directory
nm: ../generic/regexp.o: No such file or directory
nm: ../generic/tclAsync.o: No such file or directory

This will be probably be corrected once ru returns from vacation in a
few weeks.  Please don't attempt to do anything with these ports for

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url :

More information about the freebsd-current mailing list