Re: Intent to implement: WebGL 2.0
On Wed, May 7, 2014 at 6:19 PM, Benoit Jacob jacob.benoi...@gmail.com wrote: The Implementations are free to return a context that implements a higher version part violates the above requirement 1. in your email, The WebGL working group wants web pages to opt in to the WebGL2 specific parts of the functionality explicitly. For that reason, a necessary modification to this plan is to remove the sentence, Implementations are free to return a context that implements a higher version [...]. I agree with everything else. It seems like you want to be able to do that going forward so you don't have to maintain a large matrix forever, but at some point say you drop the idea that people will want 1 and simply return N if they ask for 1. -- http://annevankesteren.nl/ ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
2014-05-08 5:53 GMT-04:00 Anne van Kesteren ann...@annevk.nl: It seems like you want to be able to do that going forward so you don't have to maintain a large matrix forever, but at some point say you drop the idea that people will want 1 and simply return N if they ask for 1. Yes, that's what we agreed on in the last conversation mentioned by Ehsan yesterday. In the near future (for the next decade), there will be webgl-1-only devices around, so allowing getContext(webgl) to automatically give webgl2 would create accidentaly compatibility problems. But in the longer term, there will (probably) eventually be a time when webgl-1-only devices won't exist anymore, and then, we could decide to allow that. 2014-05-08 5:53 GMT-04:00 Anne van Kesteren ann...@annevk.nl: Are we forever going to mint new version strings or are we going to introduce a version parameter which is observed (until we decide to prune the matrix a bit), this time around? Agreed: if we still think that a version parameter would have been desirable if not for the issue noted above, then now would be a good time to fix it. If we're doing the latter, maybe we should call the context id 3d this time around... WebGL is low-level and generalistic enough that it is not specifically a 3d graphics API. I prefer to call it a low-level or generalistic graphics API. (*plug*) this might be useful reading: https://hacks.mozilla.org/2013/04/the-concepts-of-webgl/ Benoit -- http://annevankesteren.nl/ ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
On Thu, May 8, 2014 at 12:56 PM, Benoit Jacob jacob.benoi...@gmail.com wrote: WebGL is low-level and generalistic enough that it is not specifically a 3d graphics API. I prefer to call it a low-level or generalistic graphics API. Fair, forgot about that argument. webgles or some such might be better then. Or enshrine webgl2 forever, but that seems rather weird. (*plug*) this might be useful reading: https://hacks.mozilla.org/2013/04/the-concepts-of-webgl/ Ta -- http://annevankesteren.nl/ ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
On 08/05/14 12:56, Benoit Jacob wrote: (*plug*) this might be useful reading: https://hacks.mozilla.org/2013/04/the-concepts-of-webgl/ Comedy. I just read that article, and thought this article is awesomely useful. I then looked at the comments, and it turned out that the first comment is from me a year ago, saying this article is awesomely useful. :-) Gerv ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Disabling strict warnings as errors in xpcshell
I think what we should do is confirm that strict warnings as errors is actually turned on for xpcshell, and if so, where this happens. Indeed. Can you set a breakpoint in toggleWerror and see what trips it? Gavin On Thu, May 8, 2014 at 7:02 AM, Eddy Bruel ejpbr...@mozilla.com wrote: To be clear, I don't actually know where the werror flag for xpcshell tests is set. What I've observed is: -If we Cu.import a JSM that requires module X, and then also require module X from head_dbg.js after that, all is well. -If we require module X from head_dbg.js first, and then also require module X from the JSM after that, we get strict errors. What I've gathered from reading the code is: -Each JSContext has a werror flag, which is set to false by default. -Each JSScript inherits the werror flag from the JSContext in which it is compiled. -loadSubScript uses the JSContext of its caller. -Cu.import loads JSM's in a separate JSContext, which doesn't have the flag set. -Our CommonJS loader uses loadSubScript under the hood, and caches modules. The conclusion I drew from that is that apparently, the werror flag is set for the context in which head_dbg.js is compiled: if the module gets loaded from that context first, it inherits the werror flag, causing strict warnings to be turned into errors. Conversely, if the JSM gets to load the module first, it gets loaded from the JSM's context, which doesn't have the flag set, and everything is fine. This is the best explanation I could come up with based on the behavior I've observed. It could be that I missed something, and the werror flag isn't actually set for xpcshell tests, but if that's the case, they are still coming from *somewhere*, and this only strenghtens my argument that it's extremely confusing what causes them to be enabled. So my main point remains, and this is still an issue for us. I think what we should do is confirm that strict warnings as errors is actually turned on for xpcshell, and if so, where this happens. CC'ing bholley, because he probably knows where to look. On 08/05/14 03:45, Gavin Sharp wrote: To elaborate: - Bug 524781 is still open - I don't see any reference to -werror or -S in runxpcshelltests.py Gavin On Wed, May 7, 2014 at 4:48 PM, Gavin Sharp ga...@gavinsharp.com wrote: When xpcshell tests are run, they flip a bit on the initial JSContext that's off by default that tells spidermonkey make the strict warning messages into error messages. Do you have a pointer to where this happens? I've never heard of this, and couldn't find it MXRing. Gavin On Wed, May 7, 2014 at 4:39 PM, Fitzgerald, Nick nfitzger...@mozilla.com wrote: On 5/7/14, 4:21 PM, Gavin Sharp wrote: What does get rid of strict warnings as errors for xpcshell tests mean in practice? It means that our non-standard spidermonkey strict mode (not the actual strict mode) console warnings would continue to simply be console warning messages rather than console error messages in xpcshell tests. I don't understand how you're getting into the situation of accidentally turn[ing] on strict warnings as errors. Eddy can explain this better than me because he's been deep in these trenches the last couple weeks, but I'll give it a shot. When xpcshell tests are run, they flip a bit on the initial JSContext that's off by default that tells spidermonkey make the strict warning messages into error messages. Depending on how you load JS code, you might share the JSContext or you might not; for example, loadSubScript shares the JSContext, while Cu.import doesn't. Eddy has been making changes to the debugger server so that it will run in workers so we can debug workers. He has been replacing Cu.import calls with calls to a module loader that uses loadSubScript underneath the hood. So now code that used to be evaluated with this bit flipped off (because it is off by default and it was getting its own JSContext) is being evaluated with the bit on (because it is inheriting the JSContext from the xpcshell test). The result is error messages which cause devtools tests to fail. ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform On 08/05/14 03:45, Gavin Sharp wrote: To elaborate: - Bug 524781 is still open - I don't see any reference to -werror or -S in runxpcshelltests.py Gavin On Wed, May 7, 2014 at 4:48 PM, Gavin Sharp ga...@gavinsharp.com wrote: When xpcshell tests are run, they flip a bit on the initial JSContext that's off by default that tells spidermonkey make the strict warning messages into error messages. Do you have a pointer to where this happens? I've never heard of this, and couldn't find it
Re: Intent to implement: WebGL 2.0
On Thursday, May 8, 2014 5:25:49 AM UTC-4, Henri Sivonen wrote: Making the Web little-endian may indeed have been the right thing. Still, at least from the outside, it looks like the WebGL group didn't make an intentional wise decision to make the Web little-endian but instead made a naive decision that coupled with the general Web developer behavior and the dominance of little-endian hardware resulted in the Web becoming little-endian. http://www.khronos.org/registry/typedarray/specs/latest/#2.1 still says The typed array view types operate with the endianness of the host computer. instead of saying The typed array view types operate in the little-endian byte order. Don't build big endian systems anymore. *Maybe* that's cunning politics to get a deliberate little-endianization pass without more objection, but from the spec and a glance at the list archives it sure looks like the WebGL group thought that it's reasonable to let Web developers deal with the API behavior differing on big-endian and little-endian computers, which isn't at all a reasonable expectation given everything we know about Web developers. This is a digression, and I'm happy to discuss the endianness of typed arrays/webgl in a separate thread, but this decision was made because it made the most sense, both from a technical perspective (even for big endian machines!) and from an implementation perspective. You seem to have a really low opinion of Web developers. That's unfortunate, but it's your opinion. It's not one that I share. The Web is a complex platform. It lets you do simple things simply, and it makes complex/difficult things possible. You need to have some development skill to do the complex/difficult things. I'd rather have that than make those things impossible. - Vlad ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
On 2014-05-08, 5:51 AM, Anne van Kesteren wrote: On Wed, May 7, 2014 at 6:19 PM, Benoit Jacob jacob.benoi...@gmail.com wrote: The Implementations are free to return a context that implements a higher version part violates the above requirement 1. in your email, The WebGL working group wants web pages to opt in to the WebGL2 specific parts of the functionality explicitly. For that reason, a necessary modification to this plan is to remove the sentence, Implementations are free to return a context that implements a higher version [...]. I agree with everything else. It seems like you want to be able to do that going forward so you don't have to maintain a large matrix forever, but at some point say you drop the idea that people will want 1 and simply return N if they ask for 1. Yeah, this is the intent I think. Cheers, Ehsan ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
update-timer runs during mochitests, causing intermittent failures. We should disable it.
Diagnosing bug 1006075 led me to believe that our mochitest-browser framework instantiates and runs the update-timer service (http://dxr.mozilla.org/mozilla-central/source/toolkit/mozapps/update/nsIUpdateTimerManager.idl, http://dxr.mozilla.org/mozilla-central/source/toolkit/mozapps/update/nsUpdateTimerManager.js). Just my luck, the fx-team TBPL test runners for Linux/Debug take just the right amount of time to run the test suite, that the update timer service usually invokes the Add-on Manager's background update at the same time as the test suite is trying to test it, leading to unpredictable behaviour. The problem was detected because I put in a sanity check (since backed out) to throw an error if a background update starts while another is in progress. As far as I can tell, very few tests depend on the update manager; there is a small test suite specifically for the update manager, but I don't see anything else that should indirectly depend on having it run. I propose that we disable the update timer manager in at least the mochitest-browser test suite and xpcshell (if it isn't already); I'm open to suggestions for other suites where it is on and shouldn't be. I'm also interested in suggestions for the best way to implement the disable. - irving - ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
On 2014-05-08, 8:55 AM, Anne van Kesteren wrote: On Thu, May 8, 2014 at 12:56 PM, Benoit Jacob jacob.benoi...@gmail.com wrote: WebGL is low-level and generalistic enough that it is not specifically a 3d graphics API. I prefer to call it a low-level or generalistic graphics API. Fair, forgot about that argument. webgles or some such might be better then. Or enshrine webgl2 forever, but that seems rather weird. One issue which we discussed at the meeting yesterday is how we should handle the fact that a version dictionary member would be optional anyway, and what the default should be. I think it would be very hard to pick a sane default when that is not passed in because of the goal of preventing content accidentally working, so I think we might need to end up with webgl + N forever... :/ Which is not great, but if this ends up being the only bad part of the solution, I think I would be happy with it. Cheers, Ehsan ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Test failures in mochitest-browser caused by doing work inside DOM event handlers
I've recently fought my way through a bunch of intermittent test failures in the Add-on Manager mochitest-browser suite, and there's a common anti-pattern where tests receive a Window callback, usually unload, and proceed to do significant work inside that event handler (e.g. opening/closing/focusing other windows; see one detailed case at https://bugzilla.mozilla.org/show_bug.cgi?id=608820#c281). The 'unload' event is signalled before the window is completely unloaded; I don't know the fine details but the stack is in a state where other window operations sometimes fail. Specifically, I found it a bit surprising that the run_next_test() function in the async test harness starts the next test immediately on top of the current JS stack, inside whatever callbacks are currently in progress. I just filed bug 1007906 proposing that we modify the run_next_test function in the mochitest framework to always schedule the next test for a future spin of the event loop, to allow the stack to unwind - the xpcshell test harness already does this, see http://dxr.mozilla.org/mozilla-central/source/testing/xpcshell/head.js#1443. We should also update the MDN documentation about writing mochitests to strongly advise making all DOM and Window callback listeners as small as possible; my preference is to advocate using Promises as callback listeners, because stack is always unwound before the .next handler is invoked. - irving - ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Removing telemetry probes while keeping them reported for ESR
On 07/05/2014 9:51 AM, Henri Sivonen wrote: I would like to retire some telemetry probes in Firefox 32, but at the same time, I would like the Telemetry Dashboard to keep reporting data for those probes throughout the lifetime of Thunderbird 31 ESR. To make sure that the probes don't disappear from the Telemetry Dashboard during the ESR lifetime, should I leave the probe definitions in Histograms.json in m-c and set their expires_in_version to 32 instead of just removing the entries from Histograms.json in m-c? Right, the best way to do this is to set the expires_in_version to 32 and don't delete the histogram from Histograms.json until the histogram isn't needed anywhere. The Telemetry server also uses the definitions from this file. Note that if you set expires_in_version to 32, you won't collect any data from 32 / 32.0a1 / 32.0b1 etc. Also from what I gather, Thunderbird ESRs have been replaced with regular Thunderbird releases. ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Intent to implement: WebGL 2.0
On Thu, May 8, 2014 at 1:40 PM, Ehsan Akhgari ehsan.akhg...@gmail.com wrote: On 2014-05-08, 8:55 AM, Anne van Kesteren wrote: On Thu, May 8, 2014 at 12:56 PM, Benoit Jacob jacob.benoi...@gmail.com wrote: WebGL is low-level and generalistic enough that it is not specifically a 3d graphics API. I prefer to call it a low-level or generalistic graphics API. Fair, forgot about that argument. webgles or some such might be better then. Or enshrine webgl2 forever, but that seems rather weird. One issue which we discussed at the meeting yesterday is how we should handle the fact that a version dictionary member would be optional anyway, and what the default should be. I think it would be very hard to pick a sane default when that is not passed in because of the goal of preventing content accidentally working, so I think we might need to end up with webgl + N forever... :/ Which is not great, but if this ends up being the only bad part of the solution, I think I would be happy with it. Based on what has been discussed in this thread so far, specifically that: * We expect that UAs and hardware will make these features available in chunks rather than a feature here and a feature there. * We want devs to opt in to getting access to new features so that they don't accidentally depend on a particular chunk. (FWIW, this sounds like an awesome thing to me given the hardware dependency here). * We expect each new version of WebGL to be fully backwards compatible. I don't see any particular problems with the createContext(webgl/webgl2/webgl3) approach. A well written page will detect what level is supported and fall back to older versions as needed. A poorly written page will just do createContext(webgl2) and not fall back to webgl1. Whatever we do we can't get such a page to run properly on a device which doesn't support webgl2, no matter what syntax we use. A really poorly written page might do createContext(webgl2) and then only use the webgl1 feature set. Given the constraint that we want people to opt-in to additional functionality over the base webgl1 profile, there's always a risk that people will attempt to opt in to more than they need and then refuse to work when those opt-ins fail. What bad code patterns is it that people are worried that this proposal will lead to? The only thing that I could see which in some sense leads to a materially simpler API is to simply do createContext(webgl) and ask developers to do feature detection on the returned context. However that abandons the opt-in design which I think would be worse. No matter what, I think the situation here is dramatically different from CSS quirks mode and JS use strict. In both those situations there are backwards incompatible modes. That does not appear to be the case here. / Jonas ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Proposed W3C Charter: CSS Working Group
The W3C is proposing a revised charter for: Cascading Style Sheets (CSS) Working Group http://lists.w3.org/Archives/Public/public-new-work/2014May/.html http://www.w3.org/Style/2013/css-charter.html deadline for comments: May 29 Mozilla has the opportunity to send comments or objections through May 29. Please reply to this thread if you think there's something we should say. As with charters for other long-lived groups, the details in the list of deliverables largely matter only in that they might constrain what the group is allowed to work on without a new charter; the dates tend to be largely fiction. My intent (unless I get other comments) is to support the charter, but point to comments that have been made on the working group mailing list since the charter was sent out for review (some of which are about documents omitted from the list) so that the W3C has a chance to incorporate those comments (which it probably wouldn't if no AC representative mentioned them). -David -- 턞 L. David Baron http://dbaron.org/ 턂 턢 Mozilla https://www.mozilla.org/ 턂 Before I built a wall I'd ask to know What I was walling in or walling out, And to whom I was like to give offense. - Robert Frost, Mending Wall (1914) ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Proposal to revise W3C Process
W3C recently published the following proposal to revise the W3C process, which is effectively in the Proposed Recommendation stage right now: http://www.w3.org/2014/05/Process-20140506/ The document contains a description of the major changes in the Status of this Document section. The most substantive change is to merge the Last Call and Candidate Recommendation stages of the process, and begin the Advisory Committee review at Candidate Recommendation. There are also substantive changes to the rules for what constitutes adequate implementation experience. It has been developed mostly in public, in a community group: http://www.w3.org/community/w3process/ (I'm nominally a member, but didn't participate much.) Mozilla has the opportunity to send comments as a W3C member through June 13. Comments can also be made on the public-w3process mailing list. The document will also be under discussion at the Advisory Committee meeting on June 9-10, which I am attending, so if there are comments you think Mozilla should make, it's best that I hear about them prior to that meeting. This is an incremental change to the process. It's making some rather substantive changes to the stages W3C specifications go through, and mostly leaving other areas of the process alone. I'd prefer not to drag other (unrevised) areas of the process document into this process cycle in the hopes that there will be further revisions of the process document to improve other areas of the process (and thus not go another 9 years without a revision to the process), although I concede I haven't talked to others about that idea. If there are things you think Mozilla should raise in its comments, please bring them up in this thread. -David -- 턞 L. David Baron http://dbaron.org/ 턂 턢 Mozilla https://www.mozilla.org/ 턂 Before I built a wall I'd ask to know What I was walling in or walling out, And to whom I was like to give offense. - Robert Frost, Mending Wall (1914) signature.asc Description: Digital signature ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Proposed W3C Charter: CSS Working Group
I'd really like to see the CSS WG spend some time on properties that allowed more control over scrolling and zooming. Also something that addresses the complexity involved in building long scrollable lists. Right now a lot of websites implement their own scrolling behavior in JS by listening to UI events, cancelling their default behavior, and then setting various scroll related values through the DOM. UIs that snap scrolling to certain elements or that implement edge effects can't be done any other way. However this doesn't work well at all as we are moving to a world where scrolling happens off the main thread. Likewise, being able to implement a UI where a user can do zooming currently requires using JS to catch touch/mousewheel events, cancelling them, and then using the DOM to implement a zooming UI. UIs that allow zooming of a part of a webpage, for example zooming in on a picture or an email while leaving surrounding UI intact, can't be implementing right now without main-thread roundtrips. The touch-action CSS property defined in pointer-events attempts to address some of this, but falls far short. The obvious first problem is that it's touch specific (which is sort of ironic given that it's a spec that attempts to remove the separation between touch and mouse input). Finally there's the issue of how to use the DOM and CSS to create a lazily populated list. Currently lazily populating the DOM from a database leads to a lot of churn in reflows of content that's already on the page. These issues are *the* top difficulties on the layout side in creating a performant UI on low end mobile devices. I don't see any deliverables listed in the charter that's attempting to address this. Should we ask that it's added, or is it covered by any of the existing ones? / Jonas On Thu, May 8, 2014 at 5:22 PM, L. David Baron dba...@dbaron.org wrote: The W3C is proposing a revised charter for: Cascading Style Sheets (CSS) Working Group http://lists.w3.org/Archives/Public/public-new-work/2014May/.html http://www.w3.org/Style/2013/css-charter.html deadline for comments: May 29 Mozilla has the opportunity to send comments or objections through May 29. Please reply to this thread if you think there's something we should say. As with charters for other long-lived groups, the details in the list of deliverables largely matter only in that they might constrain what the group is allowed to work on without a new charter; the dates tend to be largely fiction. My intent (unless I get other comments) is to support the charter, but point to comments that have been made on the working group mailing list since the charter was sent out for review (some of which are about documents omitted from the list) so that the W3C has a chance to incorporate those comments (which it probably wouldn't if no AC representative mentioned them). -David -- 턞 L. David Baron http://dbaron.org/ 턂 턢 Mozilla https://www.mozilla.org/ 턂 Before I built a wall I'd ask to know What I was walling in or walling out, And to whom I was like to give offense. - Robert Frost, Mending Wall (1914) ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Time to revive the require SSE2 discussion
I'm curious how much of that 1% is on old versions of Firefox and aren't updating anyway? / Jonas On Thu, May 8, 2014 at 5:42 PM, matthew.br...@gmail.com wrote: On Tuesday, January 3, 2012 4:37:53 PM UTC-8, Benoit Jacob wrote: 2012/1/3 Jeff Muizelaar jmuizel...@mozilla.com: On 2012-01-03, at 2:01 PM, Benoit Jacob wrote: 2012/1/2 Robert Kaiser ka...@kairo.at: Jean-Marc Desperrier schrieb: According to https://bugzilla.mozilla.org/show_bug.cgi?id=594160#c6 , the Raw Dump tab on crash-stats.mozilla.com shows the needed information, you need to sort out from the info on the second line CPU maker, family, model, and stepping information whether SSE2 is there or not (With a little search, I can find that info again, bug 593117 gives a formula that's correct for most of the cases). https://crash-analysis.mozilla.com/crash_analysis/ holds *-pub-crashdata.csv.gz files that have that info from all Firefox desktop/mobile crashes on a given day, you should be able to analyze that for this info - with a bias, of course, as it's only people having crashes that you see there. No idea if the less biased telemetry samples have that info as well. On yesterday's crash data, assuming that AuthenticAMD\ family\ [1-6][^0-9] is the proper way to identify these old AMD CPUs (I didn't check that very well), I get these results: The measurement I have used in the past was: CPUs have sse2 if: if vendor == AuthenticAMD and family = 15 if vendor == GenuineIntel and family = 15 or (family == 6 and (model == 9 or model 11)) if vendor == CentaurHauls and family = 6 and model = 10 Thanks. AMD and Intel CPUs amount to 296362 crashes: bjacob@cahouette:~$ egrep AuthenticAMD\|GenuineIntel 20120102-pub-crashdata.csv | wc -l 296362 Counting SSE2-capable CPUs: bjacob@cahouette:~$ egrep GenuineIntel\ family\ 1[5-9] 20120102-pub-crashdata.csv | wc -l 58490 bjacob@cahouette:~$ egrep GenuineIntel\ family\ [2-9][0-9] 20120102-pub-crashdata.csv | wc -l 0 bjacob@cahouette:~$ egrep GenuineIntel\ family\ 6\ model\ 9 20120102-pub-crashdata.csv | wc -l 792 bjacob@cahouette:~$ egrep GenuineIntel\ family\ 6\ model\ 1[2-9] 20120102-pub-crashdata.csv | wc -l 52473 bjacob@cahouette:~$ egrep GenuineIntel\ family\ 6\ model\ [2-9][0-9] 20120102-pub-crashdata.csv | wc -l 103655 bjacob@cahouette:~$ egrep AuthenticAMD\ family\ 1[5-9] 20120102-pub-crashdata.csv | wc -l 59463 bjacob@cahouette:~$ egrep AuthenticAMD\ family\ [2-9][0-9] 20120102-pub-crashdata.csv | wc -l 8120 Total SSE2 capable CPUs: 58490 + 792 + 52473 + 103655 + 59463 + 8120 = 282993 1 - 282993 / 296362 = 0.045 So the proportion of non-SSE2-capable CPUs among crash reports is 4.5 %. Just for the record, I coded this analysis up here: https://gist.github.com/matthew-brett/9cb5274f7451a3eb8fc0 SSE2 apparently now at about one percent: 20120102-pub-crashdata.csv.gz: 4.53 20120401-pub-crashdata.csv.gz: 4.24 20120701-pub-crashdata.csv.gz: 2.77 20121001-pub-crashdata.csv.gz: 2.83 20130101-pub-crashdata.csv.gz: 2.66 20130401-pub-crashdata.csv.gz: 2.59 20130701-pub-crashdata.csv.gz: 2.20 20131001-pub-crashdata.csv.gz: 1.92 20140101-pub-crashdata.csv.gz: 1.86 20140401-pub-crashdata.csv.gz: 1.12 Cheers, Matthew ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
Re: Enabling new HTTP cache on nightly (browser only, not automated tests) soon
On Thu, May 8, 2014 at 3:48 PM, Vladan Djeric vdje...@mozilla.com wrote: Why didn't TP5 report a regression in memory usage? Because TP5's memory measurements are meagre and usually fail to detect even obvious regressions. And this leak only occurred in unusual circumstances; AWSY is much better than TP5 at detecting regressions (though still far from perfect) and it missed this case too. Nick ___ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform