|
@yoshuawuyts | |||||
|
Ohh, looking at jetbrains a bit more: the test coverage window is really cool! jetbrains.com/help/rider/Uni…
If the language server protocol had a way to provide coverage info to the IDE, that would make for a really good view.
Very curious which other things jetbrains provide :O
|
||||||
|
||||||
|
yosh
@yoshuawuyts
|
2. velj |
|
Want: test integration for @code that resembles Heroku's TAP integration UI. VS Code already performs tight integration for linters/compiler warnings. Might as well lean in and provide the same for test runs as well.
|
||
|
|
||
|
yosh
@yoshuawuyts
|
2. velj |
|
There should be better ways of providing test output possible than dumping the contents into a terminal.
I mean; lists work great when working in a terminal. But if you have a DOM available there should be ways to better share this information, and improve UX in the process.
|
||
|
|
||
|
yosh
@yoshuawuyts
|
2. velj |
|
Jetbrains provides something similar; though it doesn't feel like much attention has been paid to UX: jetbrains.com/help/rider/Ref…
But yeah; it'd be interesting to see attempts at improving this. Wish testing in VS Code was entirely frictionless.
|
||
|
|
||
|
yosh
@yoshuawuyts
|
2. velj |
|
Okay, taking a step back: if I could have it my way I'd make it so tools would help me figure out what to tackle next. What I feel is always missing is *overview*.
For code that's: missing tests, docs, todos, open issues, PRs, deps.
|
||
|
|
||
|
yosh
@yoshuawuyts
|
2. velj |
|
In browsers it's: a11y, load timings, caniuse metrics, test coverage, CSS utilization, fps, db usage, cache headers, img compressions, and all the lighthouse stuff.
*something* should be possible there to automate answering the q: "What can I improve?"
|
||
|
|
||
|
yosh
@yoshuawuyts
|
2. velj |
|
It feels like so many tools out there exist doing a singular task, with varying degrees of success. But so few programming tools consider a complete workflow.
It feels there's so much room for improvement.
|
||
|
|
||
|
Heinz N. Gies
@heinz_gies
|
2. velj |
|
I think you could get Test failure as diagnostic messages with the LSP, that just require the maintainer to run tests and stuff. (Which could also be painful for longer running tests)
|
||
|
|
||