Hi,
While working on shared compilation cache for windows, I noticed I could
get a 20% build time improvement with the following in .mozconfig:
mk_add_options export COMPILE_PDB_FLAG=
mk_add_options export HOST_PDB_FLAG=
mk_add_options export MOZ_DEBUG_FLAGS=-Z7
(the downside is a
On Thu, Mar 20, 2014 at 11:46 PM, Nicholas Nethercote
n.netherc...@gmail.com wrote:
At start-up, with a new profile, Firefox creates more than 230 system
compartments.
I just measured again and got 198, which is great! The start-up
numbers on AWSY have improved by maybe 5-10 MiB since I
Mike Hommey wrote:
mk_add_options export MOZ_DEBUG_FLAGS=-Z7
-Z7 is faster than -Zi?
Do VS2013 users need to turn off -FS?
--
Warning: May contain traces of nuts.
___
dev-platform mailing list
dev-platform@lists.mozilla.org
On Thu, Apr 24, 2014 at 10:19:11AM +0100, Neil wrote:
Mike Hommey wrote:
mk_add_options export MOZ_DEBUG_FLAGS=-Z7
-Z7 is faster than -Zi?
Surprisingly, yes.
Do VS2013 users need to turn off -FS?
Maybe, although it may just be ignored if using -Z7.
Mike
On Thu, Apr 17, 2014 at 10:09 AM, Henri Sivonen hsivo...@hsivonen.fi wrote:
I am not done preparing the removal patches yet, but with my current
patch queue I can already get 149 KB off of Android ARMv7 optimized
apk size and 138 KB off of Android ARMv7 optimized libxul size.
And the numbers
We intend to ship Gecko on RAM-constrained devices. Yet, we build and
ship code that is pure bloat: code that is built with Firefox/B2G but
is used only by c-c code or code that's built with Firefox/B2G but not
used by anyone.
I have prepared a queue of patches that removes Netscape-era (circa
(CC'ing people who have worked on the ICU integration)
On Thu, Apr 24, 2014 at 2:31 PM, Henri Sivonen hsivo...@hsivonen.fi wrote:
However, especially in the context of slimming down our own set of
encoding converters, it's rather demotivating to see that at least on
desktop, we are building
2014-04-24 8:31 GMT-04:00 Henri Sivonen hsivo...@hsivonen.fi:
I have prepared a queue of patches that removes Netscape-era (circa
1999) internationalization code that efforts to implement the Encoding
Standard have shown unnecessary to have in Firefox. This makes libxul
on ARMv7 smaller by
On 4/24/2014 9:20 AM, Benoit Jacob wrote:
2014-04-24 8:31 GMT-04:00 Henri Sivonen hsivo...@hsivonen.fi:
I have prepared a queue of patches that removes Netscape-era (circa
1999) internationalization code that efforts to implement the Encoding
Standard have shown unnecessary to have in Firefox.
On Thu, Apr 24, 2014 at 09:20:06AM -0400, Benoit Jacob wrote:
2014-04-24 8:31 GMT-04:00 Henri Sivonen hsivo...@hsivonen.fi:
I have prepared a queue of patches that removes Netscape-era (circa
1999) internationalization code that efforts to implement the Encoding
Standard have shown
(2014/04/24 21:49), Till Schneidereit wrote:
(CC'ing people who have worked on the ICU integration)
On Thu, Apr 24, 2014 at 2:31 PM, Henri Sivonen hsivo...@hsivonen.fi wrote:
However, especially in the context of slimming down our own set of
encoding converters, it's rather demotivating to
Telemetry experiments is enabled as of this morning's Nightly.
Our first experiment is already live: it reorders the tiles in
about:newtab so that the most-frecent tile is in the center instead of
the top-left. The experiment should deploy to 25% of nightly users for a
5-day period from now
On 4/24/14, 1:51 AM, Nicholas Nethercote wrote:
There are
still 16 compartments whose name includes devtools, for example.
Filed https://bugzilla.mozilla.org/show_bug.cgi?id=1001131 for lazily
loading devtools modules where it makes sense.
___
Hi everyone,
As many of you are already aware, we have been working hard towards making
it possible to express our Web facing Javascript APIs in WebIDL. These
.webidl files can be found under dom/webidl in mozilla-central. Over the
past few years we have been trying to be more aware and
So I tried it. My objdir size did increase by 40%. Although I can't confirm
any speedups as the build time very much depend on the environment I am
building in and the other stuff that I am doing on the box. I will try to
so an untouched build in a cooled environment in a couple of days.
On Thu,
(I want to avoid entangling the dom/webidl plan with this discussion, which
is why I forked the thread)
On Thu, Apr 24, 2014 at 3:22 PM, Ehsan Akhgari ehsan.akhg...@gmail.comwrote:
Following up on this, people asked us to not abuse the superreview flag for
this purpose
If this is abuse,
Those asides are precisely the reason it's abuse :)
We should update the list, but from a quick skim I think there aren't
more than 2-3 names on that list that need removing. Part of the
problem might be solved by introducing an superreviewer emeriti
list.
Gavin
On Thu, Apr 24, 2014 at 3:36 PM,
On Thu, Apr 24, 2014 at 4:03 PM, Ehsan Akhgari ehsan.akhg...@gmail.comwrote:
* Are there obvious places that people should inspect for code that's
being built but not used? Some libs that got imported for WebRTC
maybe?
Nothing big comes to my mind. Perhaps hunspell on b2g?
On 2014-04-24, 6:36 PM, Bobby Holley wrote:
(I want to avoid entangling the dom/webidl plan with this discussion,
which is why I forked the thread)
On Thu, Apr 24, 2014 at 3:22 PM, Ehsan Akhgari ehsan.akhg...@gmail.com
mailto:ehsan.akhg...@gmail.com wrote:
Following up on this, people
On 2014-04-24, 7:24 PM, Mike Hommey wrote:
On Thu, Apr 24, 2014 at 07:03:09PM -0400, Ehsan Akhgari wrote:
On 2014-04-24, 8:31 AM, Henri Sivonen wrote:
However, especially in the context of slimming down our own set of
encoding converters, it's rather demotivating to see that at least on
On Thu, Apr 24, 2014 at 3:58 PM, Gavin Sharp ga...@gavinsharp.com wrote:
Those asides are precisely the reason it's abuse :)
We should update the list
What is the list good for, exactly? There doesn't seem to be any consistent
usage of it anymore. In the areas that I work on (JS, XPConnect,
(moving dev-b2g to bcc because cross-group threads are evil)
We do have fairly clear rules: http://www.mozilla.org/hacking/reviewers.html
(The definitions of Significant and API are somewhat subjective, though
it's impossible to come up with completely objective definitions - IIRC
there were
On Thu, Apr 24, 2014 at 9:38 PM, Henri Sivonen hsivo...@hsivonen.fi wrote:
And the numbers are in with the complete set of removals:
apk size reduction: 193 KB
libxul size reduction: 181 KB
This is good stuff! Thanks.
Nick
___
dev-platform mailing
On 2014-04-24, at 05:31, Henri Sivonen hsivo...@hsivonen.fi wrote:
* Are there obvious places that people should inspect for code that's
being built but not used? Some libs that got imported for WebRTC
maybe?
https://bugzilla.mozilla.org/show_bug.cgi?id=1001114
I’m told that sipcc is 3M
Sounds good to me.
On Thu, Apr 17, 2014 at 6:49 PM, Boris Zbarsky bzbar...@mit.edu wrote:
We've been shipping :scope in querySelector(All) for a while now, enabled
in nightly/aurora, disabled in beta/release.
The spec is no longer trying to do wild and wooly stuff with it, so I
believe our
As we consider the value of super-reviews, we should include the
relationship between module owners and peers.
Long review queues are a burden for reviewers. Slow review turnarounds
force patch authors to juggle multiple patches to stay productive, but
this has a high context-switch overhead.
My improvements were closer to 5%. I use VS2013's amd64_x86 cross-compiler. I
didn't touch the -FS flag (I imagine it just becomes meaningless).
Debug build, before: build 17:06, objdir 3.83GB
Debug build, -Z7:build 16:06, objdir 5.22GB
Opt build, before: build 17:47, objdir 3.01GB
Opt
27 matches
Mail list logo