This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH COMMITTED: Don't break tests for enum in range


I'm not sure this is your patch, but my latest Ada run on x86_64-linux
had three new failures that might be VRP related:

c37213k.adb: In function 'C37213K':
c37213k.adb:324: error: insn 75 outside of basic blocks has non-NULL bb field
+===========================GNAT BUG DETECTED==============================+
| 4.3.0 20070606 (experimental) (x86_64-unknown-linux-gnu) GCC error:      |
| verify_flow_info failed                                                  |
| Error detected around c37213k.adb:324                                    |

c61010a.adb: In function 'C61010A':
c61010a.adb:246: error: insn 25 outside of basic blocks has non-NULL bb field
+===========================GNAT BUG DETECTED==============================+
| 4.3.0 20070606 (experimental) (x86_64-unknown-linux-gnu) GCC error:      |
| verify_flow_info failed                                                  |
| Error detected around c61010a.adb:246                                    |

,.,. CD92001 ACATS 2.5 07-06-06 23:17:29
---- CD92001 Check object attribute: X'Valid.
   * CD92001 Expected 'Valid =FALSE for Default item  1.
   * CD92001 Expected 'Valid =FALSE for Default item  2.

This is with r125496, r124624 did not fail those tests.

Laurent

On Wed, 2007-06-06 at 07:05 -0700, Ian Lance Taylor wrote:
> I noticed that gcc was in some cases breaking tests for whether a
> value was in range for an enum.  An example is this, in C++:



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]