Closed panosangelopoulos closed 4 years ago
Merging #27 into master will not change coverage by
%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #27 +/- ##
=======================================
Coverage 97.23% 97.23%
=======================================
Files 12 12
Lines 615 615
=======================================
Hits 598 598
Misses 17 17
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 74771f4...c3dd88a. Read the comment docs.
You are totally right! But I think we should drop completely the exclude fields
. I didn't understand why you revert it on the second commit. Maybe would be better to do this o v0.4
, since it would drop a settings. I plan to discuss some v0.4
features on #23, I'm just not into it yet because my time is very short right now, but I'm really looking forward to it.
@PedroBern thanks for approving it! I can completely drop the exclude
fields but first I would like to discuss it with you. Of course, if it's ok then I can add a new commit where I will remove it.
I thank you for the PR! You know what you are doing, you can do all the necessary changes to make it work. I've just created the v0.4 branch, please make the PR for it. Also, would be great if you can update the documentation about the new changes :)
Sure thing, @PedroBern! Changed the base branch and update the docs accordingly.
Currently, the library only supports exclude fields, but this seems extremely dangerous because whenever a new field is added to the user model, we have to remember to exclude it in the settings, it is more than likely that we forget this and the field is exposed to public by accident.
To prevent this, with the current PR we will support
fields
on theMeta
class of theUserNode
, which is read from the new settingUSER_NODE_FIELDS
.The whitelist approach is much safer as using only exclude fields.