Conversation
Codecov Report
@@ Coverage Diff @@
## master #427 +/- ##
=======================================
Coverage 98.92% 98.92%
=======================================
Files 38 38
Lines 3160 3160
=======================================
Hits 3126 3126
Misses 34 34
Continue to review full report at Codecov.
|
|
@acinader I opened montymxb/parse-server-test#3 a while ago to run the tests against Parse Server 3.0+. @montymxb Can you take a look at it? |
|
@dplewis well I can reproduce the fact that tests fail with parse 3 that don't fail with 2. I have no idea what is going on. There is so much output from the tests and my phpunit experience is very limited (so far). Also, I got an email from packagist that this package needs to be conformed to changes in github. So, I don't have time to dig in today, but I will over the next couple of days to a) figure out how to test against 3 and have all pass b) update packagist to use github hooks or whatever is required.... |
|
ideally, we can just use docker and not maintain a repo for this. I wanna mess around with it. Not a fan of using forever for this, just too complicated and seems unnecessary off the cuff. But again, I need to get into it to see what's what. When the tests have completed, I'd like all the scaffolding torn down. |
|
Docker is a good idea for running the server :) |
|
Hopefully you guys, @dplewis @flovilmart can help me get some momentum going :)? Here's the first example of a test that passes against parse-server 2.8 but fails against master/3.13. for this test Here's the verbose output from 2.8 And here's the output from master: |
|
Smells like a regression on the server Sent with GitHawk |
|
regression!?! this is a job for BISECT! hmmmm. 8/15? Something is fishy here.... now let's see if i can repro on the js sdk..... |
|
of course I can't! cause if no one could sign up for like 6 months, we'd prolly know. in addition to which, i have had many many users sign up using our php sdk against parse-server 3+. So this is some kind of tooling issue.....and the tooling is really hard to use anywhoo, so gonna try a different approach. brb |
|
@acinader not sure what's the issue, this fails when attempting to upgrade the use from anonymous, which is not supported on JS. |
|
well, that is just one test @flovilmart there's a few dozen that fail against 3 that work against 2. I'll look at a few more to see if I can figure out what is going on. |
|
oy. I tried a few things on the php side to maintain the functionality of deleting the anonymous authData when a user sets her username, but I haven't found a working solution. I introduced this issue in #417 So, any suggestion? I'm keen to get out a new release.
|
|
oh, and if I rollback #417 then i'm down to just a small handful of tests that fail that I'll need to pursue. |
|
ok, i think I have all tests passing. now just have to figure out what is going on with packagist. I don't think I am going to bite off switching to docker for testing right now. @dplewis you wanna figure out how we can switch npm over from @montymxb ? i've just been using and i think that just merging your branch into your master and then using is just fine for now imho until we can simplyfy the testing. what do you think? |
|
That's what I've been using, its a solution for now |
|
@dplewis I think this is ready. I think i've mucked up packagist a bit. we may need to go to 1.5.1 to unmuck it. https://packagist.org/packages/parse/php-sdk it thinks that 1.5.0 is released and it doesn't have the stuff in this branch. my bad. |
|
It looks good to me! I'll leave packagist to you |
No description provided.