Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: notify onValidationError on paste if validation failed #1462

Merged
merged 9 commits into from Apr 15, 2024

Conversation

zewa666
Copy link
Contributor

@zewa666 zewa666 commented Apr 9, 2024

this is both a new feature and a bug fix.
Now when pasting if the target cell has an editor with a validator, and that one returns invalid, the onValidationError event will be triggered.

But far more sever is that the paste on target cells with editors was currently broken.
See when going to https://ghiscoding.github.io/slickgrid-universal/#/example22, setting any row into edit mode, and pasting a number from Excel lets say into the column Completion. (check the devtools).

This regression somewhat was related to this PR #1429

Copy link

stackblitz bot commented Apr 9, 2024

Review PR in StackBlitz Codeflow Run & review this pull request in StackBlitz Codeflow.

Copy link

codecov bot commented Apr 9, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.8%. Comparing base (e902a87) to head (7dfdb97).

Additional details and impacted files
@@           Coverage Diff            @@
##           master   #1462     +/-   ##
========================================
+ Coverage    99.8%   99.8%   +0.1%     
========================================
  Files         199     199             
  Lines       21703   21707      +4     
  Branches     6992    6993      +1     
========================================
+ Hits        21639   21643      +4     
  Misses         64      64             

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@@ -127,15 +127,15 @@ export class SlickCellExternalCopyManager {

// if a custom getter is not defined, we call serializeValue of the editor to serialize
if (columnDef) {
if (columnDef.editor) {
if (columnDef.editor?.model) {
Copy link
Owner

@ghiscoding ghiscoding Apr 9, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

actually I think it would be best to call the one that the SlickGrid is aware of which is editorClass since SlickGrid doesn't have any knowledge whatsoever of editor.model (you won't find editor.model anywhere in SlickGrid core file, it is assigned later by the wrapper which can be VanillaGridBundle, AngularSlickgrid, AureliaSlickgrid, ...). Note that there a few of them to replace

Copy link
Contributor Author

@zewa666 zewa666 Apr 9, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thats interesting bc the examples all assign the model directly in the column definitions.

EDIT: ah ok misread, you're not using it in the core itself

hmm that also means though, while internalEditor is only deprecated but not gone we should take that first before editorClass?

Copy link
Owner

@ghiscoding ghiscoding Apr 9, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hmm that also means though, while internalEditor is only deprecated but not gone we should take that first before editorClass?

hmm I don't so, I would think that using editorClass should work, internalEditor is only used by the wrappers. I understand that it might be confusing, it might be just in my head but because I want to keep my slickgrid.core.ts file as close as possible to the original SlickGrid project. Some properties are not meant to be used within SlickGrid core file, for example any props related to export services and any plugin props and so on.. when I wasn't sure if I was going to still use 6pac/slickgrid in here, I added Generics on SlickGrid that could be extended and so when I was using it in the wrapper, I could simply provide my custom GridOption and Column interfaces from the wrapper and SlickGrid was happy with it but it's just an extra layer that I would rather not really use too much (in fact I still have a lot of places where I have grid: SlickGrid instead of the full typed ref of grid: SlickGrid<GridOption, Column> because I find it too long and I don't have to do that anymore since I merged SlickGrid in here and typically the props I want to use are already in the GridOption/Column base interfaces)... so long story short, it's kind of "only me" knowing this little thing but I'm not sure I really want to go with the full Generics+extend interface way either because I think it would overcomplicate things. I don't want to extend GridOption 3 times (1x for SlickGrid, 1x for Slickgrid-Universal, 1x. for wrapper Angular-Slickgrid... I'm already doing an extend on the wrapper and because of that I always need to release Slickgrid-Universal + all wrappers on the same day because if I don't, TypeScript complains for some reasons that I've never figured out.

export class SlickGrid<TData = any, C extends Column<TData> = Column<TData>, O extends BaseGridOption<C> = BaseGridOption<C>> {
// Public API

Copy link
Owner

@ghiscoding ghiscoding Apr 9, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hmm that also means though, while internalEditor is only deprecated but not gone we should take that first before editorClass?

the internalEditor will be removed completely in the next major and its goal was simply to hide the complexity of Editor creation to the user. I didn't want to introduce a breaking change so it's still in place and still work but will be removed completely in next major. SlickGrid always had a editor grid option (it never had the filter option though) and it was used like this editor: Slick.longText (which is simple uninstantiated Class), but since I use a few external libraries (ms-select, autocomplete, flatpickr, ...), I needed a way to potentially provide extra options and I strongly felt that the developer would still prefer to use editor without necessarily caring about how it was used internally which is why at the end I came up with that internalEditor thing, it's something that was meant to be used only within the wrapper and hidden to the user...but in some cases the user want to have access to the editor instance and modify some props (e.g. add some items to the ms-select collection) and in that case the user needed to reference the instance (not the class itself that SlickGrid editor had) and that was confusing that the developer was creating it with editor but later had to reference it with internalEditor... so this is the long story of how it all started, hope it helps understanding why I'm trying to make it simpler ;)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok got it. yeah I make use of the internalEditor a couple of times in my ForeignKey Filter/Editor so I understand what you're trying to simplify here.

@ghiscoding
Copy link
Owner

This regression somewhat was related to this PR #1429

yeah so we don't have much Cypress tests for copy+paste I guess and that is for sure not something I would think to test myself which is why it wasn't caught. There was this recent SlickGrid PR, which we should probably replicate, that added some Cypress tests for copy+paste in SlickGrid lib. We should probably do the same in here

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 9, 2024

cool realPress was the command I was missing. yep will add some Cypress tests now that I know how to replicate it 💯

not today as its getting quite late over here but will get things sorted this week

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 14, 2024

alright I got my change now verified with Cypress as well. I'll add others as I touch the code base again. Pretty sure there will be more to come as I continue with my projects POC to Prod conversion

@@ -8,6 +8,8 @@ import {
import { SlickCustomTooltip } from '@slickgrid-universal/custom-tooltip-plugin';
import { Slicker, type SlickVanillaGridBundle } from '@slickgrid-universal/vanilla-bundle';
import { ExampleGridOptions } from './example-grid-options';
import { BindingEventService } from '@slickgrid-universal/binding';
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I like separating external vs internal imports, so would you mind moving this import further to the top? Thanks :)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

how about a linter rule for that? ;)

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

oh does that actually exists? Dang ESLint is so powerful :P

Copy link
Contributor Author

@zewa666 zewa666 Apr 14, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll create an additional PR after this one as it currently would involve 319 problems, pretty much all auto-fixable though ;)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

its not only separation but also unifies blank lines between those and ordered imports. So with the extra PR you get a chance to tweak those configs as you prefer.

I'll set it up tomorrow as I need to get some rest today

package.json Outdated
@@ -97,5 +97,8 @@
"funding": {
"type": "ko_fi",
"url": "https://ko-fi.com/ghiscoding"
},
"dependencies": {
"cypress-real-events": "^1.12.0"
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm assuming this should go in Dev Dependencies instead

@ghiscoding
Copy link
Owner

I'm not sure what's happening with Example 12 but it keeps failing with this PR, does it have any reference to internalEditor or something?

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 14, 2024

meh ... example12 was the first target I wanted to look into using for E2E tests until I realized there was no copy&paste configured. I was fixing the validator condition as it didn't sound right according to the description and because I had errors when trying an inline edit. So I guess the tests build upon that bad assumption. I'll revert that again.

Edit: yep thats it. But do you agree that it should be an OR instead of AND check there?

@ghiscoding
Copy link
Owner

meh ... example12 was the first target I wanted to look into using for E2E tests until I realized there was no copy&paste configured. I was fixing the validator condition as it didn't sound right according to the description and because I had errors when trying an inline edit. So I guess the tests build upon that bad assumption. I'll revert that again.

Edit: yep thats it. But do you agree that it should be an OR instead of AND check there?

I'm not exactly sure what you mean, would you mind explaining more? I remember adding a validation of the "Task [number]" with a RegEx, is that at all related?

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 15, 2024

@ghiscoding all good, I'll create a new PR to clearly depict what I meant. As mentioned meanwhile I rolled back the change I've done here to example12 which makes the tests now pass.

So this PR should be done from my side

container: tmpDiv, // a dummy container
column: columnDef,
position: { top: 0, left: 0 }, // a dummy position required by some editors
grid: this._grid
});
}) as Editor;
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@zewa666 ahh sorry, I missed this earlier, I think we can remove the any on the column and make the editorClass parsed as EditorConstructor that was added not long ago. I'm guessing we might have to explicitly parse it though

const editor = new (columnDef.editorClass as EditorConstructor)({ })`

unless this other suggestion works, which would be better

const editor = new columnDef.editorClass({ }) as EditorConstructor`

Note that there's 2 of these to change

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not sure if thats gonna work as editorClass can be both the ctor and an editor according to the type

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yeah ok I'll take a look at that later, I wanted to check on Stackblitz but it looks like it's blocked by the VPN on work laptop, so anyway I can review that on a separate PR if it could be improved (or not)

@ghiscoding ghiscoding merged commit 38b465c into ghiscoding:master Apr 15, 2024
6 checks passed
ghiscoding pushed a commit that referenced this pull request Apr 16, 2024
- this is extension to another PR #1462 that now uses `editorClass` and should be typed to `EditorConstructor` whenever possible (instead of `any` to get rid of the typing issue)
@ghiscoding
Copy link
Owner

ghiscoding commented Apr 16, 2024

@zewa666 your PR did not include the changes that I mentioned from SlickGrid PR change, I'm assuming that we should probably add these changes right? Which is to reapply the previous focus after waiting 100ms after the paste

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 16, 2024

nope I thought you mentioned to only replicate the copy&paste tests

@ghiscoding
Copy link
Owner

ahh my bad, I meant that code change as well, but still I'm wondering if it's useful or not? You are definitely using it more than I do so... should we replicate it? What do you think?

ghiscoding added a commit that referenced this pull request Apr 17, 2024
* chore(common): add `EditorConstructor` type whenever possible
- this is extension to another PR #1462 that now uses `editorClass` and should be typed to `EditorConstructor` whenever possible (instead of `any` to get rid of the typing issue)
@zewa666
Copy link
Contributor Author

zewa666 commented Apr 17, 2024

need to check out that one today, especially with multi cell pastes

@ghiscoding
Copy link
Owner

ghiscoding commented Apr 18, 2024

need to check out that one today, especially with multi cell pastes

@zewa666 Have you had a chance to check this one? There's no rush per say, I would just like to know because I'm planning to push a new version this weekend. It could be postpone if you don't have time to check, no worries. Cheers

I have 1 more possible enhancement to look into and once I push that new release then it might be the last one before the next major. I'm thinking to create a next branch and start working on next release PRs. I was also nearly done with the transition from Flatpickr to Vanilla-Calendar-Pro but it doesn't look like the person is in a hurry to accept my PRs to fix couple small things, so I'm looking at forking it and maybe completely rewriting it to my taste even though that's more work (I think it's missing a couple of methods and I don't like how we set the settings in that lib which is quite different compared to Flatpickr, so I want to make it closer to Flatpickr to ease the transition). Anyway, I like that lib nonetheless and I find that the UI and UX are so much better with the new lib, I think it's worth the switch and it's much smaller :) See Vanilla-Calendar-Pro and PR #1466

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 18, 2024

nope not yet but will do so tomorrow.

I've also seen the new date lib, which could be a real improvement. especially bc flatpickr lacks some really standard stuff like a clear button in the textfield (we had to hack one by ourselves).

with regards to the next branch, I think thats a good idea to have one early preview available (could also be npm published via next flags).

the couple smaller PRs now where the ones my teammates and me found during our first sprint. next one starts in two weeks but I feel like we've already covered a huge area. So yeah I'll check out and if it makes sense try to squeeze in the other PRs feature so that its rdy for your next release.

@zewa666
Copy link
Contributor Author

zewa666 commented Apr 19, 2024

Hmmm ... I'm not 100% sure this is exactly a safe thing to do as focus stealing in general is regarded a bad practice. Imagine your validation lets another text box pop up where you need to write in a confirmation text. Now this PR would steal the focus and destroy your workflow. While the example is made up, generally we dont know what the user might do.

Now in this exact scenario I think I would be pro this feature but I'd introduce it during the breaking change release as focus redirects can lead to undesired side-effects. So should we open an issue with a link to said PR for the sake of not forgetting it once we're around the breaking change timing?

@ghiscoding
Copy link
Owner

if you think it's more of a problem than a solution then we can just forget about the change. However if you think it's still worth it, then we can add it to the list of future changes or open an issue, up to you

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants