Improve "exclude_empty" processing.

Agree, using empty() is incorrect here. And it's definitely more clear to read.
#3 looks great, RTBC.

1) exclude_empty should never be absent. It comes from the database, which does not accept NULL values. If it is absent, the source of the error is elsewhere. Webform's policy is NOT to safety check data which should never be invalid. It leads to excess code and allows sloppiness elswhere.

2) I like the use of implode().

3) I do not like deleting the comment about grids, which provides the motivation for the code. We need to remember that we are writing these comments not for us, who may remember the issues involved, but for future webform maintainers.

I would support a patch that only simplifies this with implode().

excluded_components is an optional parameter. This why the formal argument is set to NULL, a valid case that the code needs to handle.

D7 and webform generally don't put types in doc blocks. I don't object, but they shouldn't use namespaces. excluded_components can be NULL or an array of integers, or an array of strings.

@DanChadwick, If you will not contradict yourself,

Don't insult me. I'm volunteering my time.

And the problems is not long in coming. You not check for !empty() and, otherwise, check for NULL and for key existence.

I have no idea what you are saying. empty(), like is_set(), does not raise exceptions when the parameter is undefined. empty() returns TRUE for ("0"), which is sometimes desired but often not.

You're changing the API after it's been in use. If someone passes NULL, your change will fail.

I don't see much gain in this beyond the implode() improvement and adding a webform-standard doc block. The right place to make these changes would be in the D8 branch.

There isn't any current bug. We're fiddling with code for no functional improvement. This is not a high reward area of endeavour. I would suggest focusing your energies on either new D7 features, D7 issues, or the D8 branch. I know I am.

Your patch looks good. Nice improvement. Needs testing, which I haven't done yet.

Enough with the ad hominem comments. Let's keep this about the code.

+++ b/includes/webform.submissions.inc
@@ -636,22 +636,18 @@ function webform_submission_render($node, $submission, $email, $format, $exclude
-    if ($email && $email['exclude_empty']) {
+    if (!empty($email['exclude_empty'])) {

Even thought exclude_empty should never be absent, in #2499749: Notice: Undefined index: exclude_empty it is absent. Yes, this notice caused by incorrect usage in Webform2PDF module, but I think this condition change can really prevent such notices, and it doesn't lead to excess code at all.

So again I think it's good to go.

Re #18 -- As a matter of policy, we don't work around incompatibilities in other modules -- they get their own fixes. The code in question is good because $email may be null, not because $email['exclude_empty'] may be undefined.

@BROkEN -- I resigned as primary maintainer. That's up to quicksketch now.

Committed to 7.x-4.x.

Closing to clear out the old Webform 8.x-4.x branch. See #2827845: [roadmap] YAML Form 8.x-1.x to Webform 8.x-5.x.