User tests: Successful: Unsuccessful:
Pull Request for Issue #9585
There are some behaviors in the install adapters which are different between the update and install routes which can cause an incorrect installation. This PR tries to switch to the install route if the check for the existing extension's data yields no result from the database, but only against the update route.
Also, this makes all checkExistingExtension()
method implementations consistent in loading the extension record if the ID is found.
See #9585 for additional info including reproduction instructions.
Status | New | ⇒ | Pending |
Labels |
Added:
?
|
Labels |
Added:
?
|
Category | ⇒ | Libraries |
Rel_Number | 0 | ⇒ | 9585 |
Relation Type | ⇒ | Pull Request for | |
Labels |
Category | Libraries | ⇒ | Libraries Unit Tests |
If you can provide some sample extensions to test with it will be easier for people to test. It didn't inspire much confidence when you said you hadn't tested it yourself
I think i can test this
If you can provide some sample extensions to test with it will be easier for people to test. It didn't inspire much confidence when you said you hadn't tested it yourself
I don't have anything that is triggering the condition reported in #9585 so if anyone hitting it can share something it would help. Otherwise I'm really just making a best guess based on the steps the code goes through.
@ZhukDmitry @waveywhite since you add this problem (#9585) please provide with a package to test this fix and also please test it.
Original reporters haven't tested and I can't validate the fix as I have no code with this break. If someone else is interested, have at it.
Status | Pending | ⇒ | Closed |
Closed_Date | 0000-00-00 00:00:00 | ⇒ | 2016-09-12 20:51:49 |
Closed_By | ⇒ | mbabker |
Anyone? There are legitimate bugs here that mainly affect package extensions being fixed.