Closed Bug 1068011 Opened 10 years ago Closed 10 years ago

TPS broken due to 'let' changes in bug 1001090

Categories

(Testing Graveyard :: TPS, defect, P1)

35 Branch
defect

Tracking

(firefox34 unaffected, firefox35 affected)

RESOLVED FIXED
Tracking Status
firefox34 --- unaffected
firefox35 --- affected

People

(Reporter: cosmin-malutan, Assigned: cosmin-malutan)

References

Details

(Keywords: regression)

Attachments

(2 files, 2 obsolete files)

TPS extension doesn't works anymore with the latest change from bug 1001090. This requires to check all instances where we use let to make sure we don't declare the same variable twice. 04:47:20 Running test test_sync.js 04:47:20 04:47:20 JavaScript error: resource://tps/tps.jsm, line 823: TypeError: redeclaration of variable cb 04:51:21 04:51:21 TEST-UNEXPECTED-FAIL | test_sync.js | test did not complete 04:51:21 04:51:21 phase1: unknown
Please note that this bug also needs an upgrade of Mozmill to version 2.0.8! Also this is a critical regression and should be marked as such, and the causing bug needs to be in the blocking list.
Blocks: 1001090
Severity: normal → critical
Keywords: regression
OS: Linux → All
Priority: -- → P1
Hardware: x86_64 → All
Attached patch patch v1.0 (obsolete) — Splinter Review
This patch changes all 'let' keywords to 'var' in TPS and it upgrades the mozmill extension to 2.0.8 I ran a complete testrun with the latest nightly and it ran perfectly.
Assignee: nobody → cosmin.malutan
Status: NEW → ASSIGNED
Attachment #8490602 - Flags: review?(hskupin)
Attached patch patch v1.1 (obsolete) — Splinter Review
I added DONTBUILD tag.
Attachment #8490602 - Attachment is obsolete: true
Attachment #8490602 - Flags: review?(hskupin)
Attachment #8490603 - Flags: review?(hskupin)
Comment on attachment 8490602 [details] [diff] [review] patch v1.0 Review of attachment 8490602 [details] [diff] [review]: ----------------------------------------------------------------- I don't see the reason why we have to replace any let statement with var. This is something I already mentioned yesterday to Andrei for the Mozmill fix. Fix the real issue but do not do massive replacements of code. Also the Mozmill update should be its own attachment. Don't squeeze all together into a single patch.
Attachment #8490602 - Flags: review-
This is the change needed in TPS
Attachment #8490603 - Attachment is obsolete: true
Attachment #8490603 - Flags: review?(hskupin)
Attachment #8490640 - Flags: review?(hskupin)
Here is the patch to update mozmill to 2.0.8
Attachment #8490641 - Flags: review?(hskupin)
Comment on attachment 8490640 [details] [diff] [review] TPS changes v1 [checked-in] Review of attachment 8490640 [details] [diff] [review]: ----------------------------------------------------------------- Looks trivial enough.
Attachment #8490640 - Flags: review?(hskupin) → review+
Comment on attachment 8490641 [details] [diff] [review] patch v1.2 MOZMILL Review of attachment 8490641 [details] [diff] [review]: ----------------------------------------------------------------- Something what really worries me on this patch is the fact that changes, which were done a long time ago, have not been integrated with the last version bump of the Mozmill extension for TPS some weeks ago! Also again, the changes as landed for Mozmill 2.0.8 are totally not part of this patch! Cosmin, please tell us how you actually updated the mozmill extension. The current method is not going to work. ::: services/sync/tps/extensions/mozmill/chrome.manifest @@ -1,1 @@ > -resource mozmill resource/ Not sure when this came in, but it was definitely not between Mozmill 2.0.6 and 2.0.8. Looks like the last bump of Mozmill was not correctly done some weeks ago.
Attachment #8490641 - Flags: review?(hskupin) → review-
Comment on attachment 8490640 [details] [diff] [review] TPS changes v1 [checked-in] I landed this patch directly on mozilla-central. Not sure if it will make the next Nightly builds. We will see.
Attachment #8490640 - Attachment description: patch v1.2 TPS → TPS changes v1 [checked-in]
Version: unspecified → 35 Branch
(In reply to Henrik Skupin (:whimboo) from comment #8) > Something what really worries me on this patch is the fact that changes, > which were done a long time ago, have not been integrated with the last > version bump of the Mozmill extension for TPS some weeks ago! Also again, > the changes as landed for Mozmill 2.0.8 are totally not part of this patch! > > Cosmin, please tell us how you actually updated the mozmill extension. The > current method is not going to work. This is exactly how it was made in bug 982610. So here is how I done this: - I cloned mozmill from github, I checked out branch 2.0.8 - I copied the files form /mozmill/mozmill/extension into tps/extensions/mozmill - I removed components and content directories as they are not needed for the extension to work but for the mozmill python side or building. Please keep in mind that in TPS we use the extension only, not the jsBridge and not the python(server side). > ::: services/sync/tps/extensions/mozmill/chrome.manifest > @@ -1,1 @@ > > -resource mozmill resource/ > > Not sure when this came in, but it was definitely not between Mozmill 2.0.6 > and 2.0.8. Looks like the last bump of Mozmill was not correctly done some > weeks ago. It was correctly done, reviewed, tested and it didn't failed in the past 4 months we lived with it. Again it looks strange and it might not include all the changes because we don't use the whole code. I've got nothing to change from my side, if there is something wrong please point it to me and I'll change it.
Comment on attachment 8490641 [details] [diff] [review] patch v1.2 MOZMILL Review of attachment 8490641 [details] [diff] [review]: ----------------------------------------------------------------- (In reply to Cosmin Malutan from comment #10) > This is exactly how it was made in bug 982610. This cannot be true because as you can see lines of code are getting changed in your patch, which didn't exist in the Mozmill repository for a longer time. > - I removed components and content directories as they are not needed for > the extension to work but for the mozmill python side or building. Something you also had to do is the modification of some files. So most likely you missed to remove those lines the last time when editing the appropriate files. > Please keep in mind that in TPS we use the extension only, not the jsBridge > and not the python(server side). I know that, and this is not the problem here. > It was correctly done, reviewed, tested and it didn't failed in the past 4 > months we lived with it. Again it looks strange and it might not include all > the changes because we don't use the whole code. Given that those were only comments nothing affected the functionality of Mozmill itself. It's just a change which wasn't present before. So this raised confusion. > I've got nothing to change from my side, if there is something wrong please > point it to me and I'll change it. I manually checked all the files updated with this patch and it seems to be fine. I will get it landed in a bit.
Attachment #8490641 - Flags: review- → review+
Btw, the mozmill patch does not contain a valid commit message. I got this updated before landing it: https://hg.mozilla.org/mozilla-central/rev/5afba113494b We are not going to update older branches, given that we want to extract the tps framework soon from the tree.
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
(In reply to Henrik Skupin (:whimboo) from comment #9) > Comment on attachment 8490640 [details] [diff] [review] > TPS changes v1 [checked-in] > > I landed this patch directly on mozilla-central. Not sure if it will make > the next Nightly builds. We will see. I missed to give the changeset for that landing: https://hg.mozilla.org/mozilla-central/rev/641d450532be
Product: Testing → Testing Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: