------- Comment #1 from burnus at gcc dot gnu dot org 2008-01-19 16:03 -------
Confirm. I think it is a fallout of Paul's typespec patch (PR 34429 et alia).
In resolve_branch (which prints the error message), label->defined ==
ST_LABEL_BAD_TARGET.
The problem is probably that ST_GET_FCN_CHARACTERISTICS is not taken care of
correctly. check_statement_label has:
switch (st)
{
case ST_END_PROGRAM: [... some more ST_END*]
case_executable:
case_exec_markers:
type = ST_LABEL_TARGET;
case ST_FORMAT:
type = ST_LABEL_FORMAT;
default:
type = ST_LABEL_BAD_TARGET;
The question is now whether one simply adds the ST_GET_FCN_CHARACTERISTICS to
the cases or whether one adds it to, e.g., "#define case_executable" ?
--
burnus at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |pault at gcc dot gnu dot org
Status|UNCONFIRMED |NEW
Ever Confirmed|0 |1
Keywords| |rejects-valid
Last reconfirmed|0000-00-00 00:00:00 |2008-01-19 16:03:40
date| |
Summary|Spurious error in snapshot |[4.3 Regression] Spurious
|of 01/18/08: Statement at |error in snapshot of
|(1) is not a valid branch |01/18/08: Statement at (1)
|target statement for the |is not a valid branch target
|branch statement at (2) |statement for the branch
| |statement at (2)
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34872
