Implement form submissions by kokoISnoTarget · Pull Request #188 · DioxusLabs/blitz

kokoISnoTarget · 2025-02-28T01:47:41Z

Implements basic form submissions (fixes #178)
I tried following the specification, where it made sense.

Things that are missing are:

DomEvents: submit, a click event when implicitly submitting and formdata when constructing the entry list.
<select>
multipart/form-data
Image Buttons

nicoburns · 2025-02-28T20:10:27Z

Just a note that I going to be offline for next 3-4 days. Will review when I'm back!

nicoburns · 2025-04-01T06:20:34Z

I have rebased this on top of latest main and force pushed

nicoburns · 2025-04-10T00:45:50Z

Have rebased and force pushed again. Sorry for the delay in review. Will get to this soon!

Signed-off-by: Nico Burns <nico@nicoburns.com>

nicoburns

@kokoISnoTarget I've left you some review comments. However, in the interest of 1. getting this merged and 2. avoiding conflicts, I'm going to merge this as-is and iterate on top of it.

nicoburns · 2025-04-28T15:32:17Z

+        for id in self.form_nodes.borrow().iter() {
+            doc.reset_form_owner(*id);
+        }


html5ever actually computes the form associations for us (https://docs.rs/html5ever/latest/html5ever/interface/trait.TreeSink.html#method.associate_with_form). Perhaps we should use that?

(not sure though, as we'll need to implement this ourselves for dioxus-native anyway)

I didn't actually see this, but we need to reset all form controls anyways because html5ever does not find associations using the form attr:

<button form="form1">Test</button> <form id="form1"> </form>

nicoburns · 2025-04-28T15:35:28Z

+
+#[derive(Debug, Clone)]
+pub enum DocumentResource {
+    String(String),


What is the purpose of the String variant? It never seems to be used? Perhaps url and document_resource should just be a blitz_traits::net::Request?

nicoburns · 2025-04-28T15:36:55Z

+    pub fn new_with_root(doc: &'a BaseDocument, root: usize) -> Self {
+        TreeTraverser {
+            doc,
+            stack: vec![root],


Nit: Consider pre-allocating more stack space. Perhaps a capacity of 20?

nicoburns · 2025-04-28T15:40:38Z

+
+#[derive(Clone)]
+/// An pre-order tree traverser for a [BaseDocument](crate::document::BaseDocument).
+pub struct TreeTraverser<'a> {
+    doc: &'a BaseDocument,
+    stack: Vec<usize>,
+}
+


Can we move these "traversers" into their own traversal module please?

These also duplicate the method starting with iter_children_mut, iter_subtree_mut, next_node, etc in document.rs. The approach is slightly different, but we'll probably want to deduplicate this at some point. Could you please review those methods and let me know any thoughts you have on how we might unify these APIs.

nicoburns · 2025-04-28T15:54:46Z


    pub changed: HashSet<usize>,

+    pub controls_to_form: HashMap<usize, usize>,


Can you add a doc comment on this field please? I assume the key is the NodeId of the form control and the value is the NodeId of the form itself?

Can you add a doc comment on this field please?

Sure thing.

I assume the key is the NodeId of the form control and the value is the NodeId of the form itself?

You assumed correctly.

nicoburns · 2025-04-28T16:16:37Z

+pub struct NavigationHandler(String);
+impl NavigationHandler {
+    pub fn boxed(url: impl Into<String>) -> Box<Self> {
+        Box::new(Self(url.into()))
+    }
+}
+impl NetHandler<Resource> for NavigationHandler {
+    fn bytes(self: Box<Self>, doc_id: usize, bytes: Bytes, callback: SharedCallback<Resource>) {
+        callback.call(
+            doc_id,
+            Ok(Resource::Navigation {
+                url: self.0,
+                document: bytes,
+            }),
+        );
+    }
+}


I don't think blitz-dom should handle the network call when navigating to a new document. Not least because a custom NavigationProvider may want to implement their own loading logic which is different to that implemented in the NetProvider.

The way I think this should work is that blitz-dom just calls into the navigation provider with the "request" to navigate (be that from clicking on a link or submitting a form), and then the navigation provider can do whatever it likes. In practice, the navigation provider will have it's own reference to the (same) network provider which it can use to fetch the document.

Yea, this should be removed anyways, as it's a leftover from a test implementation.

nicoburns · 2025-04-28T19:00:10Z

+enum FormMethod {
+    Get,
+    Post,
+    Dialog,


What is the "dialog" method? Can't find any references to this.

This is intended to close a dialog.
https://html.spec.whatwg.org/multipage/form-control-infrastructure.html#attr-fs-method

* remove: net navigation handler * add comment to explain Document controls_to_form * remove unnecessary DocumentResource * clippy

kokoISnoTarget force-pushed the form_submissions branch from c0564e0 to 4ecfadc Compare February 28, 2025 02:29

SpikeHD mentioned this pull request Feb 28, 2025

Add ureq backend for blitz-net #189

Open

5 tasks

kokoISnoTarget force-pushed the form_submissions branch 2 times, most recently from dfb6847 to 520400b Compare March 6, 2025 03:12

nicoburns force-pushed the main branch from a3c7fbe to 503f70c Compare March 23, 2025 02:35

kokoISnoTarget force-pushed the form_submissions branch from 520400b to 4dd97c4 Compare March 25, 2025 19:01

nicoburns force-pushed the form_submissions branch from 4dd97c4 to 2a0564b Compare April 1, 2025 06:20

nicoburns force-pushed the form_submissions branch from 2a0564b to c06b679 Compare April 10, 2025 00:45

nicoburns force-pushed the form_submissions branch from c06b679 to 58d4b73 Compare April 26, 2025 14:01

nicoburns force-pushed the form_submissions branch from 58d4b73 to 6252cfd Compare May 7, 2025 19:12

nicoburns force-pushed the form_submissions branch 2 times, most recently from 638eb10 to aeb27b4 Compare May 20, 2025 12:45

kokoISnoTarget added 4 commits May 20, 2025 14:34

Implement form submission

9f05c6d

Use custom iterator instead of node_chain

7ca20c9

Signed-off-by: Nico Burns <nico@nicoburns.com>

Simplify some things

f2b2a31

Signed-off-by: Nico Burns <nico@nicoburns.com>

fix: add check for empty name attribute

1eb34ca

nicoburns force-pushed the form_submissions branch from aeb27b4 to 1eb34ca Compare May 20, 2025 13:34

nicoburns reviewed May 20, 2025

View reviewed changes

nicoburns merged commit 5f75fe6 into DioxusLabs:main May 20, 2025
9 checks passed

nicoburns pushed a commit that referenced this pull request May 20, 2025

some fixes from #188 (#215)

cd6ba87

* remove: net navigation handler * add comment to explain Document controls_to_form * remove unnecessary DocumentResource * clippy


		pub changed: HashSet<usize>,

		pub controls_to_form: HashMap<usize, usize>,

Conversation

kokoISnoTarget commented Feb 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nicoburns commented Feb 28, 2025

Uh oh!

nicoburns commented Apr 1, 2025

Uh oh!

nicoburns commented Apr 10, 2025

Uh oh!

nicoburns left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kokoISnoTarget May 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

kokoISnoTarget commented Feb 28, 2025 •

edited

Loading

kokoISnoTarget May 20, 2025 •

edited

Loading