Print potential errors after running a command only once

`parse_cmd_output` calls `parse_log_for_error` which outputs potential
errors by default via `log.info`.
The former then outputs those via `log.warning` again.
Fix this by passing `stdout=False` to the latter.
Also minor rewording and variable renaming.

Author: Flamefire

easybuild/tools/run.py

@@ -626,13 +626,15 @@ def parse_cmd_output(cmd, stdouterr, ec, simple, log_all, log_ok, regexp):
 
     # parse the stdout/stderr for errors when strictness dictates this or when regexp is passed in
     if use_regexp or regexp:
-        res = parse_log_for_error(stdouterr, regexp, msg="Command used: %s" % cmd)
-        if len(res) > 0:
-            message = "Found %s errors in command output (output: %s)" % (len(res), "\n\t".join([r[0] for r in res]))
+        res = parse_log_for_error(stdouterr, regexp, stdout=False)
+        if res:
+            errors = "\n\t" + "\n\t".join([r[0] for r in res])
+            error_str = "error" if len(res) == 1 else "errors"
             if use_regexp:
-                raise EasyBuildError(message)
+                raise EasyBuildError("Found %s %s in output of %s:%s", len(res), error_str, cmd, errors)
             else:
-                _log.warning(message)
+                _log.warning("Found %s potential %s (some may be harmless) in output of %s:%s",
+                             len(res), error_str, cmd, errors)
 
     if simple:
         if ec: