Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

making graphs-new the default for graphs.m.o

19 views
Skip to first unread message

Robert Helmer

unread,
Nov 2, 2011, 9:38:36 PM11/2/11
to
Hello,

I'd like to make graphs 2.0 (http://graphs-new.mozilla.org) the default
for graphs.m.o this quarter (and put the old one at graphs-old.m.o for a
while, in case anyone needs it).

I have filed a tracking bug to this effect:
https://bugzilla.mozilla.org/show_bug.cgi?id=698989

If anyone knows about issues that should prevent this move, please file
a graphserver bug and mark it blocking the tracker bug:

https://bugzilla.mozilla.org/enter_bug.cgi?product=Webtools&component=Graph%20Server

Or, reply to this message if you feel it needs discussion.

One issue I know about is that incoming URLs will need to change (tbpl,
the m.d.tree-management robot, etc) - this is mostly because the new
interface displays an average of the results of each test machine, for
each revision, rather than having to specify each machine in the URL. I
will file bugs on the various other components to this effect.

There are also one or two (very old, but still totally valid) bugs on
file to move to a better URL scheme, but I don't think this should block
making graphs-new the default (it's actually why I've been holding out
making the switch, but I feel like I'm making perfect the enemy of good
here.)


Thanks!
Rob

Clint Talbert

unread,
Nov 3, 2011, 8:29:49 PM11/3/11
to
This is great. I have seen some bugs for improving/extending
graphs-new, but I haven't seen anything that makes me think we should
block on doing this move. Graphs-new is a far better experience than
the old one, and I think the improvements we have in mind can be made as
we iterate on the new one.

That said, should the bugs we've filed for
advancements/improvements/extensions to graphs-new be marked as
"depending" on this bug? Or do you plan to triage those separately once
the move is done?

Thanks Rob!

Clint

Scott Johnson

unread,
Nov 4, 2011, 11:33:35 AM11/4/11
to dev-pl...@lists.mozilla.org
This is great, but is there a place that maps the names of the tests run
in tbpl to those that are graphed on graphs-new? It just seems like some
of the tests are difficult to find in the new graph list, given what we
have (e.g. ts_places_generated_max) on the tbpl list.

~Scott


On 11/03/2011 07:29 PM, thus spoke Clint Talbert:
> _______________________________________________
> dev-platform mailing list
> dev-pl...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-platform

Justin Lebar

unread,
Nov 4, 2011, 5:08:43 PM11/4/11
to Scott Johnson, dev-pl...@lists.mozilla.org
> This is great, but is there a place that maps the names of the tests run in
> tbpl to those that are graphed on graphs-new?

The fact that such a list needs to exist is a bug. We should be using
consistent nomenclature everywhere.

Note also that TBPL has "Firefox Opt" and "Firefox PGO" while
graphserver has "Firefox" and, hidden down at the bottom, "Firefox
non-PGO".

On Fri, Nov 4, 2011 at 11:33 AM, Scott Johnson <sjoh...@mozilla.com> wrote:
> This is great, but is there a place that maps the names of the tests run in
> tbpl to those that are graphed on graphs-new? It just seems like some of the
> tests are difficult to find in the new graph list, given what we have (e.g.
> ts_places_generated_max) on the tbpl list.
>
> ~Scott
>
>
> On 11/03/2011 07:29 PM, thus spoke Clint Talbert:
>>

Robert Helmer

unread,
Nov 4, 2011, 5:15:10 PM11/4/11
to
I took a quick look and I *think* this name is in the DB already, we
just need to expose it (graphs-new just uses the JSON exposed by the
already-existing graphserver backend, with some improvements).

I definitely agree that things like this should be consistent across tools.

Is this a problem on the current graphs.m.o also? From a quick peek at
the UI it looks like it is, but I could be looking in the wrong place..

If that's the case I'd like to not block moving graphs-new to the
default, but we can definitely fix this. Would you mind filing a bug?
0 new messages