Revert 169289 "A thread-safe Media Query Parser"

Revert 169289 "A thread-safe Media Query Parser"

A bunch of builders failing compile step.

http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac%20Oilpan/builds/2093/steps/compile/logs/stdio

http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Oilpan/builds/2319/steps/compile/logs/stdio

> A thread-safe Media Query Parser
> 
> Implement a thread-safe Media Query parser based on eseidel's tokenizer (https://codereview.chromium.org/123053002/).
> 
> This CL is composed of 2 parts:
> * Rendering the MediaQuery related classes thread-safe, by eliminating AtomicStrings and replacing them with hash comparisons.
> * Replacing BisonCSSParser for a new parser for MQ parsing.
> 
> BUG=
> 
> Review URL: https://codereview.chromium.org/171383002

TBR=yoav@yoav.ws

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=169292

[Zhenyao Mo, 2014-03-15 00:36:49]

Committed patchset #1 manually as r169292.

[eseidel, 2014-03-15 00:43:58]

Why are we reverting for oilpan bots?

[Yoav Weiss, 2014-03-15 06:04:58]

The builder results above are giving me a 404

[eseidel, 2014-03-15 07:37:43]

I think we should just roll back in all these patches which were rolled out to
fix the oilpan bots, but we should wait for mads or haraken or someone involved
in oilpan to OK that.  oilpan is not supposed to block main-line development
since it's an off-by-default compile flag.

[eseidel, 2014-03-15 07:37:44]

I think we should just roll back in all these patches which were rolled out to
fix the oilpan bots, but we should wait for mads or haraken or someone involved
in oilpan to OK that.  oilpan is not supposed to block main-line development
since it's an off-by-default compile flag.

[Mads Ager (chromium), 2014-03-17 07:10:59]

On 2014/03/15 07:37:44, eseidel wrote:
> I think we should just roll back in all these patches which were rolled out to
> fix the oilpan bots, but we should wait for mads or haraken or someone
involved
> in oilpan to OK that.  oilpan is not supposed to block main-line development
> since it's an off-by-default compile flag.

Sound good to me. We will help fix build breakages.

[Zhenyao Mo, 2014-03-17 17:14:54]

If a patch fails a few tests on oilpan bots, probably that's OK, but break the
compile and totally kill the bots, I don't think that's OK.

[Zhenyao Mo, 2014-03-17 17:14:54]

If a patch fails a few tests on oilpan bots, probably that's OK, but break the
compile and totally kill the bots, I don't think that's OK.

[eseidel, 2014-03-17 17:37:16]

Oilpan is a compile-time configuration flag.  Users are not expected to test
that configuration (there are no oilpan trybots).  Thus we should not be rolling
out non-oilpan patches for any type of oilpan break, period.

To do so, is unfair to all the non-oilpan contributors who have to worry about
configurations they can't anticipate failures for (off by default and try
servers not in the default config).

I think we should remove oilpan bots from garden-o-matic if they're causing
confusion there.  We should either be fully supporting oilpan (with trybots,
build-blockers, etc.) or not at all.  A half-state is bad for the project.

[Dirk Pranke, 2014-03-17 17:39:37]

On 2014/03/17 17:37:16, eseidel wrote:
> I think we should remove oilpan bots from garden-o-matic if they're causing
> confusion there.  We should either be fully supporting oilpan (with trybots,
> build-blockers, etc.) or not at all.  A half-state is bad for the project.

I think you're probably right that we should remove oilpan from garden-o-matic
given the other factors.

[haraken, 2014-03-17 17:50:18]

On 2014/03/17 17:39:37, Dirk Pranke wrote:
> On 2014/03/17 17:37:16, eseidel wrote:
> > I think we should remove oilpan bots from garden-o-matic if they're causing
> > confusion there.  We should either be fully supporting oilpan (with trybots,
> > build-blockers, etc.) or not at all.  A half-state is bad for the project.
> 
> I think you're probably right that we should remove oilpan from garden-o-matic
> given the other factors.

I'm afraid that garden-o-matic is the only confusing factor. Oilpan bots are
still displayed in the tree, so it's possible that gardeners roll out CLs
looking at the red oilpan bots.

Probably can we just re-announce that when oilpan bots break, people are
expected to leave it as is and just ping oilpan-reviews@chromium.org?

[Zhenyao Mo, 2014-03-17 17:56:43]

Ca we also change their category from nonlayout to oilpan?  So when I
open the link
http://build.chromium.org/p/chromium.webkit/waterfall?category=nonlayout&fail...,
the bots won't show.

On Mon, Mar 17, 2014 at 10:50 AM,  <haraken@chromium.org> wrote:
> On 2014/03/17 17:39:37, Dirk Pranke wrote:
>>
>> On 2014/03/17 17:37:16, eseidel wrote:
>> > I think we should remove oilpan bots from garden-o-matic if they're
>> > causing
>> > confusion there.  We should either be fully supporting oilpan (with
>> > trybots,
>> > build-blockers, etc.) or not at all.  A half-state is bad for the
>> > project.
>
>
>> I think you're probably right that we should remove oilpan from
>> garden-o-matic
>> given the other factors.
>
>
> I'm afraid that garden-o-matic is the only confusing factor. Oilpan bots are
> still displayed in the tree, so it's possible that gardeners roll out CLs
> looking at the red oilpan bots.
>
> Probably can we just re-announce that when oilpan bots break, people are
> expected to leave it as is and just ping oilpan-reviews@chromium.org?
>
>
> https://codereview.chromium.org/196653020/

To unsubscribe from this group and stop receiving emails from it, send an email
to blink-reviews+unsubscribe@chromium.org.

[Dirk Pranke, 2014-03-17 18:39:01]

On 2014/03/17 17:56:43, Zhenyao Mo wrote:
> Ca we also change their category from nonlayout to oilpan?  So when I
> open the link
>
http://build.chromium.org/p/chromium.webkit/waterfall?category=nonlayout&fail...,
> the bots won't show.
> 

Yeah, I can look into that. I thought I had done that already, but my memory is
probably wrong and I'm distracted by much worse things at the moment.

[Zhenyao Mo, 2014-03-17 18:41:15]

On 2014/03/17 18:39:01, Dirk Pranke wrote:
> On 2014/03/17 17:56:43, Zhenyao Mo wrote:
> > Ca we also change their category from nonlayout to oilpan?  So when I
> > open the link
> >
>
http://build.chromium.org/p/chromium.webkit/waterfall?category=nonlayout&fail...,
> > the bots won't show.
> > 
> 
> Yeah, I can look into that. I thought I had done that already, but my memory
is
> probably wrong and I'm distracted by much worse things at the moment.

my typo: change from "layout" to oilpan, not "nonlayput".

[Dirk Pranke, 2014-03-26 21:24:46]

On 2014/03/17 18:41:15, Zhenyao Mo wrote:
> On 2014/03/17 18:39:01, Dirk Pranke wrote:
> > On 2014/03/17 17:56:43, Zhenyao Mo wrote:
> > > Ca we also change their category from nonlayout to oilpan?  So when I
> > > open the link
> > >
> >
>
http://build.chromium.org/p/chromium.webkit/waterfall?category=nonlayout&fail...,
> > > the bots won't show.
> > > 
> > 
> > Yeah, I can look into that. I thought I had done that already, but my memory
> is
> > probably wrong and I'm distracted by much worse things at the moment.
> 
> my typo: change from "layout" to oilpan, not "nonlayput".

Oh, I remember now ... we can't easily set the category to "oilpan" for those
bots because of the way the buildbot code is structured (all of the linux web
configs are in a single file, and the category is a global setting in that file,
not a per-bot setting :( ).

This may have to wait until we completely re-work this waterfall and switch it
to recipes. Sorry!