Editorial: add some URL parsing examples #177
Conversation
|
@domenic feel free to add more examples in new commits. |
| <td> | ||
| <td>Failure | ||
| <tr> | ||
| <td><code>https:example.invalid</code> |
There was a problem hiding this comment.
Reminder to self: we should add this example but against about:blankno base URL since I am pretty sure (will check) that it results in https://example.invalid/.
|
I wonder if we should move this elsewhere and add an additional column to this table with respect to validity. Or perhaps Then we could also address #118. Maybe after "shorten a url’s path"? |
|
OK, added a number of my favorite examples. There were certainly more (e.g. demonstrating some of the host normalization like IDNA and IPv4/v6 parsing would have been cool; demonstrating file URLs would probably help) but this I think hammers home the main points about how a large variety of weird URLs can parse, but some things can still cause failure. Regarding validity. I think the current location is pretty good. Maybe what we need is a paragraph spelling out the fact that URL syntax is all about validity, and what validity means in practice. That paragraph could link to this example, perhaps. Adding a column for validity seems OK but we might need some more valid examples in that case, as I emphasized the invalid ones. Also I'm not sure it applies that well since we talk about validity on an absolute URL, not on an input to the URL parser, right? So e.g. the row that lists |
| <tr> | ||
| <td><code>example</code> | ||
| <td><code>https://example.com/demo</code> | ||
| <td><code>https://example.com/example</code> |
There was a problem hiding this comment.
This one was upfront because it demonstrates a rather simple thing that everyone takes for granted.
There was a problem hiding this comment.
Yeah, I just reorged into no base, then base, then failures; it seems kind of nice that way. I don't feel strongly.
|
Validity would be about the input column. Though it could be a column on its own describing the issue. |
|
I think it would be kind of confusing to say that |
|
|
|
Oh, I didn't realize; very interesting. I think that's not usually what people mean by "valid URL" but it does have a rigorous definition, so good enough for me. |
|
There is some handwaving around base URL which at some point should be resolved. Since to be a valid relative-URL string, we do need a base URL. |
|
I decided to go ahead and merge this since it's an improvement over the status quo and the validity question deserves a bit more thought and study. |
Fixes #119.