This is the mail archive of the
gcc-regression@gcc.gnu.org
mailing list for the GCC project.
Re: GCC build failed for native with your patch on 2003-01-22T05:00:02Z.
- From: Geoff Keating <geoffk at geoffk dot org>
- To: gcc-regression at gcc dot gnu dot org
- Cc: Zack Weinberg <zack at codesourcery dot com>, Roger Sayle <roger at eyesopen dot com>
- Date: 22 Jan 2003 10:10:14 -0800
- Subject: Re: GCC build failed for native with your patch on 2003-01-22T05:00:02Z.
- References: <200301220626.h0M6QrLb014379@devtools1.apple.com>
Hi Zack,
> insn-attrtab.c: In function `internal_insn_latency':
> insn-attrtab.c:18461: warning: ISO C forbids empty initializer braces
you're working on this, right?
Also, you should know that once this is fixed, there's a bootstrap
miscomparison problem on Darwin, that was introduced at about the same
time as your patch; it seems certain that it's one of:
+2003-01-21 Roger Sayle <roger@eyesopen.com>
+
+ * stmt.c (same_case_target_p): New function to determine whether
+ two case labels branch to the same target. Split out from...
+ (group_case_nodes): ... here. Use same_case_target_p instead.
+ (strip_default_case_nodes): Remove explicit case nodes
+ that branch to the default destination.
+ (expand_end_case_type): Call strip_default_case_nodes after
+ group_case_nodes, to simplify the case-list before we count it.
+ Only generate table_label RTX when actually needed. Try to share
+ thiscase->exit_label and thiscase->data.case_stmt.default_label
+ when a switch has no explicit default case. Simplify test for
+ constant index.
+
...
+2003-01-21 Zack Weinberg <zack@codesourcery.com>
+
+ * genautomata.c: Space savings in generated code:
+ (output_dfa_insn_code_func): Split out the table-enlargement
+ path to an out-of-line static function, dfa_insn_code_enlarge.
+ (output_internal_insn_latency_func): Use a lookup table for the
+ default latencies.
+ (output_print_reservation_func): Use a lookup table for the
+ strings.
+
--
- Geoffrey Keating <geoffk@geoffk.org>