Re: float8 regression failure (HEAD, cygwin)

Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>
>> One other thought I had was that we could have
>> pg_regress always allow a fallback to the canonical result file.
>>
>
> Hm, that's a good thought.  Want to see how painful it is to code?
>
>

Would this do the trick?

cheers

andrew


Index: pg_regress.c
===================================================================
RCS file: /cvsroot/pgsql/src/test/regress/pg_regress.c,v
retrieving revision 1.16
diff -c -r1.16 pg_regress.c
*** pg_regress.c    27 Jul 2006 15:37:19 -0000    1.16
--- pg_regress.c    1 Aug 2006 14:04:20 -0000
***************
*** 914,919 ****
--- 914,952 ----
          }
      }

+     /*
+      * fall back on the canonical results file if we haven't tried it yet
+      * and haven't found a complete match yet.
+      */
+
+     if (strcmp(expectname, testname) != 0)
+     {
+         snprintf(expectfile, sizeof(expectfile), "%s/expected/%s.out",
+                  inputdir, testname, i);
+         if (!file_exists(expectfile))
+             continue;
+
+         snprintf(cmd, sizeof(cmd),
+                  SYSTEMQUOTE "diff %s \"%s\" \"%s\" > \"%s\"" SYSTEMQUOTE,
+                  basic_diff_opts, expectfile, resultsfile, diff);
+         run_diff(cmd);
+
+         if (file_size(diff) == 0)
+         {
+             /* No diff = no changes = good */
+             unlink(diff);
+             return false;
+         }
+
+         l = file_line_count(diff);
+         if (l < best_line_count)
+         {
+             /* This diff was a better match than the last one */
+             best_line_count = l;
+             strcpy(best_expect_file, expectfile);
+         }
+     }
+
      /*
       * Use the best comparison file to generate the "pretty" diff, which
       * we append to the diffs summary file.