Ronny Pfannschmidt
Ronny Pfannschmidt
just took note on this again @raphaelcastaneda thanks for the analysis, moved to the main repo as we have to fix this inside of pytes
@gs202 im not sure what you are referring to here the issue is primarily about the `record_xml_attribute` fixture not using a storage/transport mechanism that that transfers via xdist
@gs202 at first glance that is a **different** issue, please open a new one
@Zac-HD im under the impression that the warning comes from fixture parsing a magic object in that library
When order doesn't matter, comparing the result of sorted is orders of magnitude simpler to comprehend. If unique values should be folded, comparing sets is order of magnitude simpler to...
There is currently an attempt to get an idea of matcher combination, so that we have a library with matchers for not just unordered, but other qualities as well, currently...
A few weeks ago i had a call with @asottile about creating a DSL for assertion /check writing that combines tools like re_asserrt, unordered, approx and a a few more...
the cached property in the stdlib used `__dict__` as storage space there would be a need for a variant that uses a slot and/or a identity weak mapping in the...
i would like to point out that "named constructors" are commonly used for such setups allowing more clear spelling ```python @attr.s class C: i = attr.ib() j: int = attr.ib()...
It's absolutely unclear how to map a straightforward simple thing like plain cached attributes into something matching declared attributes mixed with computation