Remove excessive serde tests and fix a bug #434
Conversation
| // See https://github.com/serde-rs/serde/issues/1905 | ||
| #[serde(rename = "$value")] | ||
| items: Vec<MyEnum>, | ||
| } |
There was a problem hiding this comment.
This is a much more complex test than the one that replaces it (https://github.com/tafia/quick-xml/blob/d34b77972ccaad505c3348761382fe7d01ca49ce/tests/serde-de.rs#L1261=), is it really a full replacement?
There was a problem hiding this comment.
Actually, yes. The mapping should be independent and composable, so we shouldn't bother of internals of the MyEnum type (or Item type in Project). For now we don't have a tests that ensures that mapping is really additive in all cases (which is why #435 is happened), but
- that check should be added explicitly and for all combinations
- in that test mapping is independent and composable
I'll plan to add necessary tests that would ensure that mapping is always composable (may be in form of doctests in #369)
|
I used the merge conflict cleanup feature just to try it out - you can squash that commit if you want. |
Codecov Report
@@ Coverage Diff @@
## master #434 +/- ##
==========================================
- Coverage 51.51% 51.44% -0.07%
==========================================
Files 25 25
Lines 13310 13312 +2
==========================================
- Hits 6856 6849 -7
- Misses 6454 6463 +9
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
…able_name::variable_size::simple` test
…name::variable_size::field_before_list` tests
…::top_level::simple` test
failures (1):
struct_::non_closed::missing_field
Fixes: - struct_::non_closed::missing_field
Mingun
force-pushed
the
serde-tests-cleanup
branch
from
acdb8c5
to
85a569d
Compare
|
The history becomes too confusing because of the back-and-forth merges. I'd rather do a rebase. |
This PR removes tests which have better alternatives in
seqsubmodule since #387.Also, while working on namespaces support, I accidently found a bug in a deserializer, so I've added a test and a fix for it.