Using WPT in blink guidelines

15 views
Skip to first unread message

Rick Byers

unread,
Feb 10, 2017, 12:04:08 PM2/10/17
to platform-predictability, Philip Jägenstedt, Geoffrey Sneddon
Hey Philip,
This discussion made me realize we probably need to invest more in our documentation/guidelines for blink developers.  For starters, I've marked the old docs as obsolete with a link to the new docs, and posted a CL to make a couple small tweaks.

But I guess we should think about what more we should have written down in terms of guidelines.  Eg:
  • Link to some upstream docs about writing web-platform-tests.  I'm not sure what we'd link to at the moment, so maybe wait until Geoffrey has landed the new upstream docs to replace all other WPT docs?
  • More advice on how reviewing WPTs is different from reviewing LayoutTests?
  • How to handle cases where we're landing a blink change that isn't yet reflected in specs.  Eg. does it make sense to fork a WPT into LayoutTests temporarily with a bug tracking removing it once blink again matches the spec?
  • How to work with the upstream stability_bot (or whatever we call it going forward), including avoiding causing accidental/incorrect failures on other browsers.  We're still working this out of course.
  • Something about how to handle automation (mostly "don't for now - use LayoutTests, but follow ... for progress").
  • Discuss what happens when a test change upstream causes a new failure (still TBD, but we can at least link to the bug discussing this).
WDYT?

Philip Jägenstedt

unread,
Feb 13, 2017, 6:31:11 AM2/13/17
to Rick Byers, platform-predictability, Geoffrey Sneddon
Yep, I'll see about expanding web_platform_tests.md to cover some of these things. There is a TODO about wpt_automation in writing_layout_tests.md that I should move into web_platform_tests.md.

--
You received this message because you are subscribed to the Google Groups "platform-predictability" group.
To unsubscribe from this group and stop receiving emails from it, send an email to platform-predicta...@chromium.org.
To post to this group, send email to platform-pr...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/platform-predictability/CAFUtAY8Kx7ya-w3teh1rBdbYcoyNFpkEiV8Rt9sZqHnsGyqcWg%40mail.gmail.com.

Philip Jägenstedt

unread,
Feb 14, 2017, 10:34:49 AM2/14/17
to Rick Byers, platform-predictability, Geoffrey Sneddon

https://codereview.chromium.org/2695963002 tries to address some, but not all, of the questions raised. For the various kinds of upstreaming failures, we don't know yet what will happen or how much people will have to care.

Geoffrey Sneddon

unread,
Feb 15, 2017, 5:56:15 PM2/15/17
to Rick Byers, platform-predictability, Philip Jägenstedt
On 10/02/17 17:03, Rick Byers wrote:
> # Link to some upstream docs about writing web-platform-tests. I'm not
> sure what we'd link to at the moment, so maybe wait until Geoffrey has
> landed the new upstream docs to replace all other WPT docs?

This landed today, FYI. The old docs redirect, though some just to the
homepage because they have no real corresponding page in the new docs
(mostly because they've been split up massively).

The new URL is http://web-platform-tests.org/ (yes, HTTP, :(, GitHub
Pages doesn't support HTTPS with custom domains).

/g

Rick Byers

unread,
Feb 15, 2017, 8:43:00 PM2/15/17
to Geoffrey Sneddon, Philip Jägenstedt, platform-predictability
Just read through the whole site - great work Geoffrey!

--
You received this message because you are subscribed to the Google Groups "platform-predictability" group.
To unsubscribe from this group and stop receiving emails from it, send an email to platform-predictability+unsub...@chromium.org.
To post to this group, send email to platform-predictability@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/platform-predictability/a2ec7fbf-9834-19d7-f5a5-235012fc7879%40gsnedders.com.

Philip Jägenstedt

unread,
Feb 15, 2017, 11:30:05 PM2/15/17
to Rick Byers, Geoffrey Sneddon, platform-predictability
I confess to not having read much of it, but I really like how it looks! And I've updated a bunch of links in https://codereview.chromium.org/2696323003/ to point to the things that I haven't read :)

On Thu, Feb 16, 2017 at 8:43 AM Rick Byers <rby...@google.com> wrote:
Just read through the whole site - great work Geoffrey!
On Feb 15, 2017 5:56 PM, "Geoffrey Sneddon" <m...@gsnedders.com> wrote:
On 10/02/17 17:03, Rick Byers wrote:
> # Link to some upstream docs about writing web-platform-tests.  I'm not
> sure what we'd link to at the moment, so maybe wait until Geoffrey has
> landed the new upstream docs to replace all other WPT docs?

This landed today, FYI. The old docs redirect, though some just to the
homepage because they have no real corresponding page in the new docs
(mostly because they've been split up massively).

The new URL is http://web-platform-tests.org/ (yes, HTTP, :(, GitHub
Pages doesn't support HTTPS with custom domains).

/g

--
You received this message because you are subscribed to the Google Groups "platform-predictability" group.
To unsubscribe from this group and stop receiving emails from it, send an email to platform-predicta...@chromium.org.
To post to this group, send email to platform-pr...@chromium.org.
Reply all
Reply to author
Forward
0 new messages