Firefox 69.0 Released • Hacker News Board

Re: Firefox 69.0 Released

#311

Post by hispanic » Wed, Sep 04, 2019, 2:59 PM UTC

denzil_correa wrote: post #20

> macOS users on dual-graphics-card machines (like MacBook Pro) will switch back to the low-power GPU more aggressively, saving battery life. macOS user base can now consider FF as a viable browser alternative.

I wonder if this will help address the system freezes I've been getting for the past month on my early 2011 MacBook Pro. The fix I was going to pursue makes an attempt to disable usage of the "bad" GPU: https://forums.macrumors.com/threads/2011-%E2%80%A2-15-17-in...

Most of my lock-ups seem to happen when I try to access specific sites using Firefox, especially financial sites.

Re: Firefox 69.0 Released

#312

Post by la_barba » Wed, Sep 04, 2019, 4:00 PM UTC

Const-me wrote: post #248

la_barba wrote: post #233

Earlier quoted context omitted.

I guess I was thinking of this from a heterogeneous architecture standpoint (e.g. big.LITTLE) . You have differently-capable compute resources, and you need to pick the one thats best suited for the work-load. Rather than the app talking directly to the hardware, I suppose the OS should let the app pick its work-load type, similar to letting it choose a process scheduling priority.

Unlike ARM cores, GPU code is expensive to migrate between them. Two GPUs have different ISA, each GPU driver compiles platform-independent bytecode like DXBC or SPIR-V into proprietary instruction set. VRAM can contain many GB of data, when migrating, everything needs to be copied. The asymmetric ARM cores at least have same RAM, and very similar instruction set. These issues make live migration impractical. AFAIK,…

When switching from onboard -> dedicated, the resources (textures, shaders, etc) to migrate will be quite small. There is no question of migrating resources from dedicated -> onboard as that is never going to happen, unless .. again the resources are tiny. The context switch performance hit will depend on the other components on the board, but for a high end CPU/RAM/SSD combo, it wont be much. It will be equal to re-loading the all the browser tabs (for e.g. after a browser crashes).

Re: Firefox 69.0 Released

#313

Post by kup0 » Wed, Sep 04, 2019, 4:24 PM UTC

Hope this release somehow magically fixes an ongoing issue in Windows 10 I've had with Firefox.

The gist of it is that Firefox, and only Firefox, will cause my Windows 10 system to completely hard freeze (no mouse movement, no response period) for about 30-60 seconds at a time, multiple times daily. It's so frustrating that it makes FF unusable for me.

It's odd because it seems to be a rare issue that has to do with it not letting go of a GPU handle/process/thread or something, from what I've been able to deduce from others having this issue that have posted bugs in the tracker. I've tried everything under the sun to fix it, but no other program, period has this problem. I game, I use other browsers, I do all sorts of stuff on this PC with zero issues, but Firefox gives me these temporary hard-freezes. Ugh.

Re: Firefox 69.0 Released

#314

Post by canada_dry » Wed, Sep 04, 2019, 5:12 PM UTC

kup0 wrote: post #313

Hope this release somehow magically fixes an ongoing issue in Windows 10 I've had with Firefox. The gist of it is that Firefox, and only Firefox, will cause my Windows 10 system to completely hard freeze (no mouse movement, no response period) for about 30-60 seconds at a time, multiple times daily. It's so frustrating that it makes FF unusable for me. It's odd because it seems to be a rare issue that has to do with…

> only Firefox will cause my Windows 10 system to completely hard freeze

Whenever I experience this type of thing I can't help but speculate that issues like this are not serendipity at work.

I recollect reading an article years back from a Microsoft developer who spoke about how it was quite common for Windows code to take specific actions based on which app was running - ostensibly to 'improve the user experience'. Though, it's not hard to imagine MS using this for the opposite purpose. And, since their code is proprietary, who would know?

Re: Firefox 69.0 Released

#315

Post by gregknicholson » Wed, Sep 04, 2019, 5:36 PM UTC

thirdsun wrote: post #303

bradgessler wrote: post #175

Earlier quoted context omitted.

I created https://legiblenews.com/ because of this frustration ... and took it all the way to the level where loading a page is exactly 1 request. The other frustration I’ve had with news websites is they don’t link to source material, so it’s impossible to dig into a topic and accidentally learn something.

I like it. However given your goal of delivering sane and lighweight news articles you might want to extend that mission to the selection of your sources and consider excluding some. For the Boris Johnson story you link to The Independent, which returns a 14 MB article (while autplaying is disabled in Firefox) and basically has become a collection of worst practices when it comes to user experience and web developmen…

The data comes from [Wikipedia's Current Events portal][1], so the site's author isn't directly choosing those sources.

[1]: https://en.m.wikipedia.org/wiki/Portal:Current_events

Re: Firefox 69.0 Released

#316

Post by smt88 » Wed, Sep 04, 2019, 5:36 PM UTC

veidr wrote: post #271

smt88 wrote: post #126

Earlier quoted context omitted.

Last time I installed Brave, the new tab page had a link to a cryptocurrency news site (probably a paid placement). If you want the least-gross behavior, it's not going to come from a for-profit startup.

Maybe not, but it seems to be the least annoying right now . I've never seen it show me an ad (including paid placement type ads) so far. I guess if they junked up their new tab page with intrusive encouragement to start using Brave Rewards and Basic Attention Token (BAT) then that would sort of be the equivalent of what Firefox is doing...

It doesn't block ads very well for me on iOS, and it's really buggy and unstable. My Firefox experience is excellent on every device that supports it. It's unfortunate and frustrating that all iOS browsers are just Safari.

Re: Firefox 69.0 Released

#317

Post by grey_earthling » Wed, Sep 04, 2019, 5:41 PM UTC

bradgessler wrote: post #175

Carpetsmoker wrote: post #130

Earlier quoted context omitted.

> 5KB news article On what site do you find "5KB news articles"? Pretty much all news sites that I know of load a gigantic amount of useless and obnoxious JavaScript, CSS, images, etc. with or without video.

I created https://legiblenews.com/ because of this frustration ... and took it all the way to the level where loading a page is exactly 1 request. The other frustration I’ve had with news websites is they don’t link to source material, so it’s impossible to dig into a topic and accidentally learn something.

Nice!

This would be slightly more legible if there was a clearer distinction between the tags introducing each item and the text of the item itself.

Also, it would be more usable if the dates in the URLs were numeric and international-style (ISO-8601 / xkcd-1179).

But nice!

Re: Firefox 69.0 Released

#318

Post by gregknicholson » Wed, Sep 04, 2019, 5:58 PM UTC

DenseComet wrote: post #294

autoexec wrote: post #253

Earlier quoted context omitted.

> "without selling out their user's privacy" This is the key. I think sending any unsolicited traffic to 3rd party websites counts as violating my privacy. Pocket goes far beyond that. Pocket Recommendations are personalized based on your browsing behavior in Firefox. It doesn't matter that Mozilla and Pocket don't see your browsing history directly. While the choice of which links to push at you are made client side…

Whats really interesting to me is that this seems to be more of a huge marketing issue. Pocket IS Mozilla* and it seems like people would be more ok with the integration if they knew this. Now, obviously, Mozilla does need to improve the sponsored stories and not track the user without input, but at the same time, its not a 3rd party in the general stance. * https://blog.mozilla.org/blog/2017/02/27/mozilla-acquires-p…

> Pocket IS Mozilla* and it seems like people would be more ok with the integration if they knew this.

I'm less forgiving. The Pocket service is proprietary - this technology couldn't fall into the right hands - and it directly competes against open web standards like RSS/Atom. I honestly don't know how they justify it against the Mozilla Manifesto.

Re: Firefox 69.0 Released

#319

Post by gregknicholson » Wed, Sep 04, 2019, 6:17 PM UTC

TheRealPomax wrote: post #134

pseudalopex wrote: post #124

Earlier quoted context omitted.

It's been 2 years. They have no specific plans to release the server code.[1] They won't say what's been done or has to be done. [1] https://github.com/Pocket/extension-save-to-pocket/issues/75

On a software product note, "Pocket", the client-side product that does local-only interest analysis so it can get a full list of interesting articles from the server, and then locally filter out all the stuff you probably don't care about, is not "the server application that accepts volutary user submission analysis to determine what goes on the API response list with which categories". No one ever talked about open…

1. Is it possible to use the open source Pocket client built into Firefox with any other server, e.g. by specifying another domain name in about:config?

2. Is there an open source implementation of compatible server software, that you could use to run your own server?

The answers to these questions will tell you whether Pocket is open source.

Re: Firefox 69.0 Released

#320

Post by gregknicholson » Wed, Sep 04, 2019, 6:27 PM UTC

bzbarsky wrote: post #152

kevingadd wrote: post #56

Earlier quoted context omitted.

It has to have been showing up in profiles for them to bother, nobody would go out of their way to disable it otherwise. They added a pref to turn it back on and then disabled it instead of the easy thing (just removing it) which is some engineering effort plus testing effort to make sure the pref works right. Silly to do that for nothing. Antivirus software does add a measurable delay to file operations sometimes, a…

Not just that, but in an earlier release code was shipped to set the pref for users who already had the file, so that anyone who already had one of these files would not observe any behavior regressions. And yes, file I/O was the issue.

Thank you for doing this!