DOM getElementsByTag order consistency - dom

When doing a getElementsByTag, I see a particular order in the returned nodelist. Can I count on that order being consistent from any particular browser vendor from one release to the next and is there consistency among all browser vendors?
The more generic question is - Is there a spec that the browser vendors follow to guarantee an order?

I just can asnwer for Firefox: the returned (e.g. elements = element.getElementsByTagName(tagName)) is a live NodeList of found elements in the order they appear in the subtree.
Source: https://developer.mozilla.org/en-US/docs/Web/API/element.getElementsByTagName
I think there is no spec to guarantee an order, because W3C doesn't mention that:
"getElementsByTagName: Returns a collection of all descendant Elements
with a given tag name.
Parameters: tagname -The name of the tag to match on. If the string
"*" is given, this method will return all elements in the document.
Return Values: A new NodeList object containing reference to all the
Elements found.
This method raises no exceptions.
Source: http://www.w3.org/TR/WD-DOM/level-one-core.html

Related

VSCode Extension: Get outline of function for custom outliner

I'm trying to create a custom outliner for VSCode (currently only for python), but I don't find measures to get the information I needed.
I like to get information in this manner this:
Array:
[0]
label: "foo"
type: "Function"
parameters: [...]
Range: [...]
innerDefinitions: [0]
[1]
label: "myclass"
type: "Class"
base_class: ""
Range: [...]
innerDefinitions:
[0]:
[...]
[1]:
[...]
Currently I try to get outline information via vscode.commands.executeCommand( 'vscode.XXX'
What I've tried:
Here is what commands I've tried and what result I received.
vscode.executeImplementationProvider
half usable: range of functionname. Other information is missing
vscode.executeHoverProvider
half usable: string of function head (including def keyword)
vscode.executeDefinitionProvider
half usable: range of complete function. Individual information must be "parsed out"
vscode.executeTypeDefinitionProvider
Never provided any result
vscode.executeDeclarationProvider
Never provided any result
vscode.executeDocumentSymbolProvider
Goes in a good direction. However
(1) Does only work on the whole document (not single function)
(2) Does only return first-level entities (i.e. class methods are not included in result)
Is there any API call I've overseen?
I wonder how the built-in outliner works, as it contains all-level information.
You need to use vscode.commands.executeCommand<vscode.Location[]>("vscode.executeDocumentSymbolProvider", uri, position)
This will give you the full outline of one file. There is no way to receive a partial outline.
Note: innerDefinitions are called children here.
Regarding the detail of the outline:
How detailed (and correct) an outline is going to be, depends on the implementation of the provider. Also, provider's information is no necessarily consistent among languages. This is very important to keep in mind!
At the moment (2021/03), the standard SymbolProvider for...
... Python will have a child for each parameter and local variable of a function. They will not be distinguishable
... C++ will contain no children for parameters. But it will have the parameter types in its name. (e.g. name of void foo(string p) will be foo(string): void.
As you can see, both act differently with their own quirks.
You could create and register a DocumentSymbolProvider yourself, that would return a level of detail you need (see VSCode Providers)
Also see: https://stackoverflow.com/a/66486297/6702598

Firestore security rules - wildcarding Collection Names?

I have a set of Collections whose names all start with ABC and I want to write a single rule that applies to all of them regardless of what follows ABC. Something like:
match /ABC*/{anyid} {
allow read, write;
}
Is this possible? In the Rules Console there are no syntax errors highlighted, but the Simulator won't allow me to access the table with:
GET /ABC123/456
Any ideas?
As far as I know it is not currently possible to match on a partial collection (or document) name. It sounds like an interesting feature request though, so I recommend filing a feature request.
In the meantime, the only thing I can think of is matching all collections, and then testing the path through resource['__name__']:
match /53829635/{document} {
match /{col}/{doc} {
allow read: if resource['__name__'][5].matches('ABC.*')
}
}
The resource['__name__'] expression returns a Path, which can be indexes as an array to get the path segments. It has a form /databases/(default)/documents/collection/document, so the subcollection is at index 5. Since that is just a string, we can use matches on it. In this case I allow reading from any subcollection whose name starts with ABC.
Update: it turns out that you can also simply access the col wildcard, instead of looking up from the path. So this would work the same:
allow read: if col.matches('ABC.*')

How to filter by both text and property in Chrome DevTool's network panel?

I want to filter Chrome DevTool's network panel by the method property and text in the URL. For example, if I am searching for the text chromequestion in the URL and only HTTP GET requests (ignore PUT, POST, DELETE, etc).
I am able to filter by text or by method:
I am not able to combine the filter to search by both text and method:
I read the documentation at https://developers.google.com/web/tools/chrome-devtools/network-performance/reference#filters and I am able to filter by multiple properties (.e.g, domain:*.com method:GET). However, I am unable to filter by text and property (e.g., method:GET chromequestion).
Unfortunately, it's not possible to do this currently. I played around in DevTools originally, but couldn't find a way. I later had a look into how the filtering was implemented, and can confirm there's a limitation preventing you from mixing the pre-defined filters and text filters.
Implementation details
This is a bit long but I thought it might be interesting for some to see how it's implemented. I will probably look into improving the implementation, either myself or I'll log it because it's limited.
There's a _parseFilterQuery function that parses the input field and categorises the entries into two arrays. The first is called filters, and it's the pre-defined filtering options, such as method:GET etc. The second is a text array filter, split up by spaces. The parser determines the difference fairly naively, by checking for the occurrence of :, and - at the start (for negation).
Scenario 1
You only input a pre-defined filter, or multiple filters. For each filter, the specific filter function, which looks at the different properties of the request object, is pushed to a network module filters array (this._filters). Later on, for each request, the function is called on it, and a match returns true, otherwise false. This will determine whether the request is shown. There's obviously a requirement for ALL filters to return true for the row to show.
Scenario 2
This is the interesting one, where you input both a pre-defined filter and a bit of text. This covers the Stack Overflow question. The _parseFilterQuery function looks at the text filters first, before the pre-defined ones. In Scenario 1, this was empty, so it was skipped.
We pass each text word to the _createTextFilter, and push each of the resulting filters to the network module filters array. However, the implementation of this is questionable. The only time the actual word passed in is used is to check whether its a negation filter for a bit of text. If the first character is -, it means the user doesn't want to see a request with the following word in the name. For example -icon means don't show any request with that in the name/page. If there is no negation, it simply returns the WHOLE input text as a regular expression, NOT the word passed in. In my case, it returns /method:GET icon/i.
The pre-defined filters are looked at next. In this case, method:GET is pushed.
Finally, it loops over the requests calling each filter on it. However, since the first filter is /method:GET icon/i, it makes ALL other filters redundant because it will NEVER pass. The text filters only apply to name and path, so method:GET in a text filter will be invalid.

Sensenet length filter Not working

I want to query the Field which are empty and which are not empty using Sensenet Odata Rest API. Their documentation mentions a filter function called 'length'. I have tried to query the field with the length operation but it fails with the error.
This is the filter I have used
$filter=length(Name) eq 2
Sense/Net 6.5.4.9496
Exception
"code": "NotSpecified",
"exceptiontype": "SnNotSupportedException",
"message": {
"lang": "en-us",
"value": "Unknown method: length"
},
Wiki Link http://wiki.sensenet.com/OData_REST_API
The length operation was included in the list of supported methods incorrectly, we apologise for that. SenseNet compiles these filters to Lucene queries and it is not possible to compose such a query in Lucene that performs an operation on a field.
(the remaining methods, like substringof or startswith can be compiled to a wildcard expression easily, so that should work)
Unfortunately 'empty' expressions are also not supported by Lucene, because of their document/term structure. So the following expression does not work either:
Description eq ''
Edit: as a workaround, developers may create a custom field index handler.
For every field you want to check for emptiness (e.g. Description), you may create a technical hidden bool field (IsDescriptionEmpty) in the content type definition. The only thing you have to create and define is a custom field index handler class. In your case it would inherit from the built-in bool field index handler and you could return a boolean index value based on whether the target field (in this case Description) is empty or not.
After this you would be able to define search exressions like the following:
+Type:File +IsDescriptionEmpty:true
Please check the wiki article below and the source code for index handler examples.
How to create a field indexhandler

RESTful reordering of nested objects

I have a RESTful API that supports two objects so as object A contains an ordered list of nested objects B:
Create object A - POST /a
Create object B and add to A - POST /a/<id>/b
Update object B in A - PATCH /a/<id>/b/<id>
What would be a RESTful way to update the order of B objects in a specific A?
Option 1: PATCH /a/<id> with json content that replaces A.Bs
A has a list of embedded Bs, namely A.Bs so you can replace that list in its entirety also changing the order on the way. This relies on the client to resubmit the entire list correctly.
Option 2: PATCH /a/<id> with json content that replaces A.B_order
Add a separate list of B ids and have the client update it. This is similar to Option 1 but does not rely on the client resubmitting all the objects. It does require the server to manage the list, updating it upon B creation, and validating the update contains all the required B ids on list order update.
Option 3: PATCH /a/<id>/b with json content that replaces A.Bs
The same as Option 1, but with a different URL
Which would be most RESTful and clear?
Any other options?
I would suggest using the proposed standard defined in RFC 6902. Specifically, the "move" op would seem to be what you're looking for.
Given Foo has Bars and Bars are resources (have ID or link), when you need to reorder foo.bars then I would suggest:
’PUT /foos/:id/bars’ with array of IDs or links in the body.
But if Bars are not resources (have no ID), then:
’PATCH /foos/:id’ with body of Foo including complete new array of Foo.bars in the ’bars’ property.
The question I would ask myself is: "What does 'order' mean in this case".
The specific instances of B don't have an order between them. They're all independent resources that are not really 'aware' of each order.
Given that, it's not really the B resource that you are changing. What are you changing?
Presumably there is a collection of B's somewhere. You do a GET request on that collection to get an ordered list of B's. Do you get that list on /a/<id> or on /a/<id>/b ?.
Wherever that ordered list is, I would also do the operation to change the order because the order is a 'property' of the collection.
So for the sake of the argument, lets assume that your 'collection of B lives on /a/<id>b. What format should that be?
Well, a good REST service will replace the entire state. So as a default I would do a PUT request on that resource and do a full replace of the entire thing.
If you don't like that idea and want to use PATCH to update only a part of the collection (and nothing else), I think I would opt for one of these:
Use a standard format, like json-patch.
Come up with your own syntax to describe this.
Option 2 will likely be a lot simpler to implement, and I would also keep the format as simple as possible.
If you use a format like HAL, a collection is probably a list of links. In that case I would use a syntax like:
{
"_links": {
"item": [
{ "href": "/a/<id>/b/ordered-item-1" },
{ "href": "/a/<id>/b/ordered-item-2" }
]
}
}
If you don't have a hypermedia-style API, you probably use id's and force the client to expand id's in urls. In that case, I'd imagine the format could look like:
{
"items": [ 1, 3, 5, 2]
}
In each case, it's a Good Idea to define your own media type for this, because this format for PATCH has special meaning to your api. For example:
application/vnd.jonathan.patch+json