ecpg regression test failures caused by window functions patch

It hadn't occurred to me to try the ecpg tests before committing the
window functions patch :-(. It looks like those grammar additions have
resulted in whitespace changes in a lot of the test outputs. Would you
confirm that there's nothing seriously wrong and update the output
files?

It strikes me that it might be a good idea to change cat2_str to not
insert a space when obviously not necessary, perhaps along the lines
of

cat2_str(char *str1, char *str2)
{
char * res_str = (char *)mm_alloc(strlen(str1) + strlen(str2) + 2);

strcpy(res_str, str1);
- strcat(res_str, " ");
+ if (strlen(str1) != 0 && strlen(str2) != 0)
+ strcat(res_str, " ");
strcat(res_str, str2);
free(str1);
free(str2);
return(res_str);
}

It looks like this would reduce the vulnerability of the ecpg tests to
whitespace changes caused by "insignificant" grammar changes.

regards, tom lane