David Lawrence [:dkl] Reporter
	Description • 10 years ago

Ryan pinged me in IRC about the Release Tracking Report not picking up a bug that is should be showing so I took a look at the SQL being generated.

<RyanVM|sheriffduty> I'm trying to figure out why bug 1024778 isn't coming up in the results when I choose "approval-mozilla-aurora", "+", "2014-05-27 and 2014-07-07" and check the "cf_status_firefox32:not fixed" option below it

So I selected the same and bug 1024778 did not show up in the list even though it seemed like it should.

The SQL generated to create the list of bug ids:

SELECT DISTINCT b.bug_id
FROM bugs b
     INNER JOIN flags f ON f.bug_id = b.bug_id 
     INNER JOIN bugs_activity a ON a.bug_id = b.bug_id
     LEFT JOIN tracking_flags_bugs AS tfb ON tfb.bug_id = b.bug_id
     LEFT JOIN tracking_flags AS tf ON tfb.tracking_flag_id = tf.id
WHERE (a.fieldid = 69)
      AND (a.bug_when >= '2014-05-27 00:00:00')
      AND (a.bug_when < '2014-07-01 00:00:00')
      AND (a.added LIKE '%approval-mozilla-aurora+%')
      AND (f.type_id = 720)
      AND (f.status = '+')
      AND (tf.name = 'cf_status_firefox32'
           AND COALESCE(tfb.value,'') NOT IN ('fixed', 'verified'))

bug_id
995407
998785
1010986

1024778 is absent. But looking at the bug itself it had all the proper attributes to match (although now it has status-firefox32:fixed set)

Any thoughts on what I am missing?

dkl

	:glob ✱ Assignee
	Comment 1 • 10 years ago

that query returns bugs where status-firefox32 is set to a value which is not 'fixed' or 'verified'.

before ryan set status-firefox32 to fixed on [2014-06-24 13:35:04 PDT], that field was not set to any value, so it was not returned by the report.

ryan - my understanding is the bug should have had status-firefox32 set to 'affected' for it to show up on your radar.  is this correct?

if not, we'll have to update the report to make "not set" values returned when searching for "not fixed".

	Ryan VanderMeulen [:RyanVM]
	Comment 2 • 10 years ago

(In reply to Byron Jones ‹:glob› from comment #1)
> if not, we'll have to update the report to make "not set" values returned
> when searching for "not fixed".

That's the behavior I'd expect, yes.

	:glob ✱ Assignee
	Comment 4 • 9 years ago

- treat unset tracking flags as 'not fixed'
- line-for-line port from yui to jquery
- show field description (status-firefox40) instead of field name (cf_status_firefox40)

	David Lawrence [:dkl] Reporter
	Comment 5 • 9 years ago

Comment on attachment 8670838 [details] [diff] [review]
1029800_1.patch

Review of attachment 8670838 [details] [diff] [review]:
-----------------------------------------------------------------

Looks good. I did notice some errors in the console when clicking [All|None][Invert]

TypeError: cb is null
TypeError: cb is null
TypeError: el is null

I was able to fix the errors locally with:

@@ -82,8 +82,11 @@ function onFlagChange() {
 // form

 function selectAllFields() {
-  for(var i = 0, l = fields_data.length; i < l; i++) {
-    var cb = Dom.get('field_' + fields_data[i].id + '_cb');
+  var productData = productEl.value == '0'
+                    ? getFlagByName(flagEl.value)
+                    : getProductById(productEl.value);
+  for(var i = 0, l = productData.fields.length; i < l; i++) {
+    var cb = $('#field_' + productData.fields[i] + '_cb')[0];
     cb.checked = true;
     onFieldToggle(cb, fields_data[i].id);
   }
@@ -91,8 +94,11 @@ function selectAllFields() {
 }

 function selectNoFields() {
-  for(var i = 0, l = fields_data.length; i < l; i++) {
-    var cb = Dom.get('field_' + fields_data[i].id + '_cb');
+  var productData = productEl.value == '0'
+                    ? getFlagByName(flagEl.value)
+                    : getProductById(productEl.value);
+  for(var i = 0, l = productData.fields.length; i < l; i++) {
+    var cb = $('#field_' + productData.fields[i] + '_cb')[0];
     cb.checked = false;
     onFieldToggle(cb, fields_data[i].id);
   }
@@ -100,12 +106,17 @@ function selectNoFields() {
 }

 function invertFields() {
-  for(var i = 0, l = fields_data.length; i < l; i++) {
-    var el = Dom.get('field_' + fields_data[i].id + '_select');
-    if (el.value == '+') {
-      el.options[1].selected = true;
-    } else {
-      el.options[0].selected = true;
+  var productData = productEl.value == '0'
+                    ? getFlagByName(flagEl.value)
+                    : getProductById(productEl.value);
+  for(var i = 0, l = productData.fields.length; i < l; i++) {
+    var el = $('#field_' + productData.fields[i] + '_select')[0];
+    if (el) {
+      if (el.value == '+') {
+        el.options[1].selected = true;
+      } else {
+        el.options[0].selected = true;
+      }
     }
   }
   serialiseForm();

We can address those separately or fix on commit. r=dkl

	:glob ✱ Assignee
	Comment 6 • 9 years ago

(In reply to David Lawrence [:dkl] from comment #5)
> Looks good. I did notice some errors in the console when clicking
> [All|None][Invert]
> 
> TypeError: cb is null
> TypeError: cb is null
> TypeError: el is null

none of these errors are new, but i'll fix them here.

> I was able to fix the errors locally with:

thanks -- although developing fixes for issues you find during review is unnecessary; feel free to just say you found a problem and let the dev fix it.

	:glob ✱ Assignee
	Comment 7 • 9 years ago

went with a much easier fix than the one you suggested (just skip elements that weren't found).

To ssh://gitolite3@git.mozilla.org/webtools/bmo/bugzilla.git
   d055408..c0de3b6  master -> master