ReSpec: linter

[Marcos Caceres]

Hi ReSpec Users,

As part of the W3C's transition to TLS docs on /TR/, we've added a
simple linter to ReSpec to check that the URLs in your `respecConfig`
are using TLS.

The linter is disabled by default right now. You can enable it by
adding the following to your `respecConfig`:

```
lint: true,
```

You can access the warnings through the ReSpec pill on the top-right,
as per usual. If no warnings show up there, you are all good :)

The linter also warns if you don't have a Privacy and/or Security
Considerations section in your spec, with a helpful link to a
questionnaire that can help you write such sections.

Once the W3C transitions to TLS on /TR/ (in ~2 weeks), we will enable
the linter by default to help you easily fix such warnings.

Kind regards,
Marcos

[Wendy Seltzer]

Post by Marcos Caceres
Hi ReSpec Users,

+cc PING

Post by Marcos Caceres
As part of the W3C's transition to TLS docs on /TR/, we've added a
simple linter to ReSpec to check that the URLs in your `respecConfig`
are using TLS.
The linter is disabled by default right now. You can enable it by
```
lint: true,
```
You can access the warnings through the ReSpec pill on the top-right,
as per usual. If no warnings show up there, you are all good :)
The linter also warns if you don't have a Privacy and/or Security
Considerations section in your spec, with a helpful link to a
questionnaire that can help you write such sections.

Thanks Marcos, that's great!

--Wendy

Post by Marcos Caceres
Once the W3C transitions to TLS on /TR/ (in ~2 weeks), we will enable
the linter by default to help you easily fix such warnings.
Kind regards,
Marcos

[Shane McCarron]

BTW I have suggested we also add some a11y checking to this. I don't yet
know what that will be. I am asking the APA and ARIA groups for input on
it. But I wonder if there shouldn't also be a check for an a11y
considerations section. I have raised an issue against specberus to
discuss that with the relevant players. [1] Please feel free to comment
here or against that issue.



[1] https://github.com/w3c/specberus/issues/465

Post by Wendy Seltzer

Post by Marcos Caceres
Hi ReSpec Users,

+cc PING

Post by Marcos Caceres
As part of the W3C's transition to TLS docs on /TR/, we've added a
simple linter to ReSpec to check that the URLs in your `respecConfig`
are using TLS.
The linter is disabled by default right now. You can enable it by
```
lint: true,
```
You can access the warnings through the ReSpec pill on the top-right,
as per usual. If no warnings show up there, you are all good :)
The linter also warns if you don't have a Privacy and/or Security
Considerations section in your spec, with a helpful link to a
questionnaire that can help you write such sections.

Thanks Marcos, that's great!
--Wendy

Post by Marcos Caceres
Once the W3C transitions to TLS on /TR/ (in ~2 weeks), we will enable
the linter by default to help you easily fix such warnings.
Kind regards,
Marcos

--
Policy Counsel and Domain Lead, World Wide Web Consortium (W3C)
https://wendy.seltzer.org/ +1.617.863.0613 (mobile)

[i***@w3.org]

Post by Shane McCarron
BTW I have suggested we also add some a11y checking to this. I don't
yet know what that will be. I am asking the APA and ARIA groups for
input on it. But I wonder if there shouldn't also be a check for an
a11y considerations section. I have raised an issue against specberus
to discuss that with the relevant players. [1] Please feel free to
comment here or against that issue.

Perhaps we could check for some basic i18n stuff too. For example:
- html tag should have lang attribute
- char encoding must be present and must be utf-8
- char encoding must be within 1024 bytes of page start
- meta content-language should not be used
- xml:lang attributes should not appear in the page


ri

[Shane McCarron]

Good ideas... except for xml:lang. Rdfa can use that in some
circumstances.

Post by i***@w3.org

Post by Shane McCarron
BTW I have suggested we also add some a11y checking to this. I don't
yet know what that will be. I am asking the APA and ARIA groups for
input on it. But I wonder if there shouldn't also be a check for an
a11y considerations section. I have raised an issue against specberus
to discuss that with the relevant players. [1] Please feel free to
comment here or against that issue.

- html tag should have lang attribute
- char encoding must be present and must be utf-8
- char encoding must be within 1024 bytes of page start
- meta content-language should not be used
- xml:lang attributes should not appear in the page
ri

[Marcos Caceres]

Post by i***@w3.org

BTW I have suggested we also add some a11y checking to this. I don't
yet know what that will be. I am asking the APA and ARIA groups for
input on it. But I wonder if there shouldn't also be a check for an
a11y considerations section. I have raised an issue against specberus
to discuss that with the relevant players. [1] Please feel free to
comment here or against that issue.

- html tag should have lang attribute

We default this to "en" if missing.

Post by i***@w3.org
- char encoding must be present and must be utf-8

Can check with: "document.charset", so doable.

Post by i***@w3.org
- char encoding must be within 1024 bytes of page start

We got bit by this recently, so we now move the the ***@charset to be
first child in the head on save... so, at least, generated files won't
have any issues.

Post by i***@w3.org
- meta content-language should not be used

Yeah, easy one. Have you seen any spec do this?

Post by i***@w3.org
- xml:lang attributes should not appear in the page

W3C doesn't publish XHTML anymore, so this one probably won't matter.

Good suggestions. The "document.charset === 'utf-8'" is probably the
lowest hanging fruit.

[i***@w3.org]

Post by Marcos Caceres

Post by i***@w3.org

BTW I have suggested we also add some a11y checking to this. I don't
yet know what that will be. I am asking the APA and ARIA groups for
input on it. But I wonder if there shouldn't also be a check for an
a11y considerations section. I have raised an issue against specberus
to discuss that with the relevant players. [1] Please feel free to
comment here or against that issue.

- html tag should have lang attribute

We default this to "en" if missing.

i guess this is ok for most TR stuff, and the validator should now catch
exceptions.

Post by Marcos Caceres

Post by i***@w3.org
- char encoding must be present and must be utf-8

Can check with: "document.charset", so doable.

Post by i***@w3.org
- char encoding must be within 1024 bytes of page start

first child in the head on save... so, at least, generated files won't
have any issues.

sounds good

Post by Marcos Caceres

Post by i***@w3.org
- meta content-language should not be used

Yeah, easy one. Have you seen any spec do this?

no i haven't - just precautionary

Post by Marcos Caceres

Post by i***@w3.org
- xml:lang attributes should not appear in the page

W3C doesn't publish XHTML anymore, so this one probably won't matter.
Good suggestions. The "document.charset === 'utf-8'" is probably the
lowest hanging fruit.

yes, and useful since i heard of a WD published quite recently that
wasn't in utf-8 (!)

thanks Shane and Marcos
ri

[Marcos Caceres]

Post by i***@w3.org

Post by Marcos Caceres
Good suggestions. The "document.charset === 'utf-8'" is probably the
lowest hanging fruit.

yes, and useful since i heard of a WD published quite recently that
wasn't in utf-8 (!)

Filed feature request:
https://github.com/w3c/respec/issues/873