Comments (4)
@gsnedders Where does the web-platform-tests.org documentation even live?
from wpt.
A few comments:
-
I think in general if a rule doesn't add significant value, we shouldn't make it. More rules to enforce are more rules to read, learn, and understand, and this reduces the chance that people will follow other more important rules. In particular, this makes me want to avoid making more rules such as (5) and (7). That said, I think there are some real file length limits that WPT users face, but I think they probably mostly relate to the full path length and not the length of any component of it.
-
Regarding point 4: The 600x600 comes from a CSS WG resolution.
-
Regarding point 8 (although I'm not sure if you were asking):
maxDifference
andtotalPixels
values are available from many test runners, including from wpt.fyi, and including the wpt.fyi links in the Checks tab of a PR. (Just because they can be found doesn't necessarily mean that the test should permit that amount of fuzziness, though.) -
Regarding point 3: "null end tag" is an SGML feature that allows syntax like
<title/Document Title/
as a shorthand for<title>Document Title</title>
. It is not part of the HTML parsing algorithm and I would be surprised if it is present in WPT tests. Are you talking about empty-element TAGS (an XML feature)? -
Regarding points 3 and 5: I think in many cases it's good for tests to vary across the normal ways that web content varies; this sometimes allows them to catch additional issues that they wouldn't otherwise have caught. I'd like to avoid being too prescriptive when it doesn't add value. And specifically on point 5 I don't want to encourage lots of edits to existing tests for stylistic reasons, which risks breaking the tests unnecessarily; sometimes "optional" tags matter, such as when a
<body>
is followed by a<script>...</script>
and the script needs to run inside the body rather than the head.
from wpt.
David,
Thank you for taking the time to write your detailed comments.
More rules to enforce are more rules to read, learn, and understand, and this reduces the chance that people will follow other more important rules.
We have review of tests and there are test reviewers. Professionalisation of tests and testing insure quality of tests and testing. I have never rejected tests unless they were false, bad or fundamentally incorrect. I always have tried to rehabilitate inadequately (or poorly) formated tests as long as they were testing a testable spec statement to begin with.
point 3: The null end tag I was referring to is, e.g.:
<meta charset="utf-8" />
when
<meta charset="utf-8">
is already good, already functional and sufficient.
point 5 (Optional end tags): I have been formally asked to remove all optional end tags and optional tags (<html>
, <head>
, <body>
) from my tests. And so, I thought the wpt Writing Tests documentation should state such requirement.
If CSS tests do not need optional end tags and optional tags and if minimal - reduced as much as possible - filesize is a recommandable goal in itself and by itself, then such requirement makes sense. As I understand, it takes several hours to several computers to go through all the wpt tests for the 4 browsers. Only necessary filesize makes sense in that regard.
Varying HTML content (tags and elements) makes sense for HTML testing, when testing HTML.
point 7 (Length of tests filenames): Only a very small minority of CSS tests exceed 60 characters. I wish I knew how to find them (with a grep command) and list them in order to establish precisely their percentage from the whole set of CSS tests. I estimate (just a hunch) that less than 0.1% use, exceed more than 60 characters.
A filename should not be a sentence describing a test. A long (> 48 characters) filename does not help review, does not replace a title text and does not replace useful comments or a text assert. I do not see how anyone gets help from a long filename.
A long filename will trigger horizontal overflow, horizontal scrolling (and scrollbar) in various situations (tables) or webpages and/or creates ellipsis ("…") characters in various places. If a long filename creates ellipsis ("…") characters, then we don't even get to read such filename anyway.
My proposal is that, from now on, the lint checking tool detects and reports a filename exceeding 48 characters and asks the author to limit to 48 characters or less.
point 8 (Where to get maxDifference and totalPixels values.)
maxDifference and totalPixels values are available from many test runners
I never found such information all by myself. And recently, Sebastian Zartner was wondering and asking where he could find those values.
Just because they can be found doesn't necessarily mean that the test should permit that amount of fuzziness, though.
Of course. Agreed.
I will be away from web and internet for a week.
from wpt.
on point 5 I don't want to encourage lots of edits to existing tests for stylistic reasons, which risks breaking the tests unnecessarily
First, my proposal wrt point 5 was not for existing CSS tests but for upcoming CSS tests.
Second, my proposal is not for stylistic reasons, is not about stylistic reasons.
Third, even if we would remove </p>
, </li>
, </dt>
, </option>
, </tr>
, </td>
, </body>
, </head>
, </html>
from existing CSS tests, - which is not my goal on point 5 - , why would CSS tests break? I do not see how or why CSS tests could or would break. All these are optional tags, even in HTML4.
from wpt.
Related Issues (20)
- Add support for testing native find-in-page search results
- Are css/css-contain/contain-layout-baseline-005.html css/css-contain/contain-layout-button-001.html correct? HOT 8
- [css-color] css/css-color/parsing/color-computed-relative-color.html assumes conversion to hwb is lossless HOT 5
- infrastructure/testdriver/click_iframe_crossorigin.sub.html can fail in Firefox HOT 1
- REGRESSION(c3a572d): TypeError: can't compare offset-naive and offset-aware datetimes
- lint for tests with link rel=match and testharness.js
- Do not run tests for PRs still in draft mode HOT 1
- Fa
- Firefox Android runs missing since May 12
- tools/unittests (Python 3.12) are failing
- Lovely
- Azure Pipeline run queued and not restarted HOT 17
- iss
- tytg HOT 1
- WebVTT rendering tests fail widely because of UA-specific cue backgrounds HOT 2
- Incorrect test added for text-transform HOT 8
- UAX 14 not strictly required by css-text HOT 2
- wpt.live lists files instead of tests HOT 1
- [python style] gentestutilsunion.py gives the WebKit style checker hiccup
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from wpt.