[PATCH] Set entryval_error to NULL if entryval is set

[Tom Tromey]

>>>>> "Yao" == Yao Qi <yao@codesourcery.com> writes:

>> The bug fix part looks ok to me, but I think you need a special test
>> case in order to really test entry values.  Or to put it another way,
>> what happens if this test is run with a compiler that doesn't emit entry
>> value information?

Yao> Patch in gdb.trace/collection.exp is to test arguments and/or entry
Yao> values are shown properly, without the internal error.  Testing the
Yao> contents of entry values is not the point of this test and the patch.

I've re-read the patch and I think that I was misunderstanding what it
was doing.  In particular I missed the "set print entry-values only" bit.

The patch is ok.

thanks,
Tom