Automatically closed - issue fixed for 2 weeks with no activity.
- ๐บ๐ธUnited States chrisfromredfin Portland, Maine
IT'S ALL IN $filters NOW!!!!1
-
chrisfromredfin โ
committed 8d991315 on 2.0.x authored by
phenaproxima โ
Issue #3505152: Svelte: Remove the searchString writable and pass the...
-
chrisfromredfin โ
committed 8d991315 on 2.0.x authored by
phenaproxima โ
- First commit to issue fork.
- ๐บ๐ธUnited States chrisfromredfin Portland, Maine
I agree with this behavior change. We are trying to make each source a bit more independent, and this gets us there. We may want to, in the future, do a feature request to persist the pageSize per source but I am going to wait for someone to ask for it. It may not be a big enough deal.
-
chrisfromredfin โ
committed 015f313f on 2.0.x authored by
phenaproxima โ
Issue #3504997: Clean up some variable references in the Svelte code
-
chrisfromredfin โ
committed 015f313f on 2.0.x authored by
phenaproxima โ
- First commit to issue fork.
- @phenaproxima opened merge request.
- ๐บ๐ธUnited States phenaproxima Massachusetts
While I'm at it, I'm going to try to remove
searchText
. It exists in a strange feedback loop with the equally poorly-namedsearchString
, and only one should exist at all. - @phenaproxima opened merge request.
- Issue created by @phenaproxima
-
chrisfromredfin โ
committed 1db616d5 on 2.0.x authored by
phenaproxima โ
Issue #3504960: Remove stateContext from the Svelte code
-
chrisfromredfin โ
committed 1db616d5 on 2.0.x authored by
phenaproxima โ
- First commit to issue fork.
- ๐บ๐ธUnited States phenaproxima Massachusetts
The tests pass on my local machine, so I have reason to believe they'll pass on CI too.
- @phenaproxima opened merge request.
- Issue created by @phenaproxima
- ๐บ๐ธUnited States chrisfromredfin Portland, Maine
I think with automated tests passing for a known reason, and the code changes being grokable, and me manual testing, I think this one looks good!
-
chrisfromredfin โ
committed d1bb6215 on 2.0.x authored by
utkarsh_33 โ
Issue #3502935: Refactor the Search component to treat filter changes...
-
chrisfromredfin โ
committed d1bb6215 on 2.0.x authored by
utkarsh_33 โ
- First commit to issue fork.
- ๐บ๐ธUnited States phenaproxima Massachusetts
Turns out the tests were only failing because we weren't always passing the filters to
load()
inProjectBrowser.svelte
. That makes sense -- we probably always want to pass them explicitly. I am facing issues in solving the test fails.I checked the URLs formed in both 2.0.x and this branch are same.
- ๐บ๐ธUnited States rbrandon
Similar to XB we have had to backport this to workaround some dependency issues and it is working well. We have several libraries that all depend on shared libraries. We were running into cases where the libraries were included in different aggregates and the dependencies were included in the minimal representative subset for both and ending up on the page twice.
Since the optimized aggregate urls don't track explicitly what other dependencies have been loaded like with ajax it would have required a lot of tinkering with weights and making duplicate libraries like with XB to get what the dependency sorting gives us out of the box.
- ๐บ๐ธUnited States phenaproxima Massachusetts
This is making a ton of sense. There are a few more steps to go, but this is really solid so far!
- ๐บ๐ธUnited States phenaproxima Massachusetts
This generally looks good. I think we can remove
filtersWrapper
, it seems like a pointless extra layer around what$filters
is already doing.We do need to fix the announce problem that @narendrar found. I would suggest that we change the backend
MultipleChoiceFilter
class so that it accepts a "template string" that can be used for announcements, and then have the announce function in Svelte use that. Make it part of the filter's definition -- how does it want to announce a state change to the user? All the filters should be taking this sort of accessibility into account, but let's fix it for the MultipleChoiceFilter type first, and the other filter types later.