It's unclear whether we should close PR 17055 in bugzilla as WONTFIX
for versions prior to 4.0.0, or backport the solution to the 3.4
and possibly 3.3 branches. My guess is that non-regression bug fixes
(such as this ICE on valid) are at the discretion of the release manager.
The following testing at least confirms that there are no technical
problems with a backport.
The following patch has been tested against the gcc-3_4-branch on
i686-pc-linux-gnu, with a full "make bootstrap", all default languages,
and regression tested with a top-level "make -k check" with no new
failures.
Ok for 3.4.3 or just close the PR?