Response consistency fixes #50
Conversation
|
@marc-maurer would you be willing to give this a look-see? :) Thank you! |
|
Seems fine |
|
Going to tackle #53 in this as well, changing to WIP |
|
I realise this probably fixed my last 3 commits.... |
|
That's ok, I'll pull in your stuff when I get a moment to get back to this. No biggie |
|
You can probably just revert all 3 when trying to merge. The reason I picked up on them is we upgraded to php 7.4 and that shows errors with the nulled / broken responses |
…pull up resources by user/updates by resource; also check if listResources has any results and if the array is empty (i.e. no more results) then send a 404 instead of a plain empty array off to infinity; closes #48.
jacobsandersen
force-pushed
the
patch/48
branch
from
a4e1bb8 to
6fd2d73
Compare
…ft_versions; fixes #53
jacobsandersen
force-pushed
the
patch/48
branch
from
6fd2d73 to
8479ff8
Compare
|
It only took me 498 days but I have now fixed #53 and put it in this PR as originally intended. @marc-maurer If you're still available could you review this again? Nothing has really changed since your last review except for |
I will take a look later today :) |
jacobsandersen
force-pushed
the
patch/48
branch
from
7a47544 to
297cee8
Compare
|
Great! Thank you Marc, I appreciate it. I'll need to take a look later and ensure I'm using strict equality where necessary. I am sure there are other areas where I am using I have made the changes you suggested |
There was a regression in #38 that caused nonexistent resources and authors to show a normal response with every field nulled out. This patch fixes that.
Furthermore, when using the
getResourcesByAuthor,getResourceUpdates, andlistResourcesroutes if you used a negative page then you would get a 404. However, if you used a page too big you would get an empty array. Also, if the user or resource did not exist (in the case of the former two routes), you would get an empty array. For all the cases you would typically get an empty array, you now get a 404.To achieve the latter part, there are two new database checks using optimized row existence queries (i.e. making use of
EXISTS, notCOUNT).This fixes #48.
This fixes #53.