HNHacker News
TopNewBestAskShowJobs

nnethercote

3,691 karma · joined July 5, 2011

submissionscomments
nnethercote··on A “negative emissions” plant has opened in Iceland
https://en.wikipedia.org/wiki/Biochar
nnethercote··on It’s time to give Firefox another chance
How is Mozilla siding with Microsoft?
nnethercote··on It’s time to give Firefox another chance
Mozilla acquire Pocket a while back. That might address your concerns, depending on what those concerns are.
nnethercote··on It’s time to give Firefox another chance
Firefox now uses multiple processes too. By default Firefox is limited to one UI process and four web content processes. Chrome will use roughly one content process per tab (though that's an oversimplification). Firefox's choice to use fewer web content processes result in less memory usage, but if a content process crashes you'll have to reload 1/4 of your tabs.

Both browsers use threads extensively within each process as well.

nnethercote··on It’s time to give Firefox another chance
Here's a blog post from this week about memory usage: http://www.erahm.org/2017/09/25/firefox-memory-usage-in-the-...

That's a comparison on real websites rather than browser benchmarks.

nnethercote··on It’s time to give Firefox another chance
Mozilla acquired Pocket a while back. So "shares with them" really means "shares with itself" :)
nnethercote··on WebRender newsletter
Strictly speaking, Stylo and WebRender fall under the Quantum project umbrella, as "Quantum CSS" and "Quantum Render", respectively. https://wiki.mozilla.org/Quantum has details. (The names are definitely confusing!)

Also, Quantum Flow is not so small! It's actually now the part of Quantum that's getting the highest priority and most resources. It doesn't involved any single component, but is rather a focused effort to relentlessly profile Firefox on real-world workloads and eliminate as many noticeable slowdowns (jank, stutters, pauses, spinning beachballs, etc.) as possible. It's been hugely effective, with noticeable speed-ups in each release from Firefox 54 to 57. Ehsan Akhgari's blog has more details: https://ehsanakhgari.org/blog

nnethercote··on 64-bit Firefox is the new default on 64-bit Windows
I think the minimum sizes were chosen because of a combination of (a) there was a much greater variety of hardware available then, and (b) people hadn't learned that fixing these sizes is better.
nnethercote··on 64-bit Firefox is the new default on 64-bit Windows
`HEAP` is the name of a large typed array variable that represents the original program's heap.
nnethercote··on Rust: Not So Great For Codec Implementing
BMP isn't so trivial, largely because it's been extended multiple times while never having a proper spec.

http://searchfox.org/mozilla-central/rev/bbc1c59e460a27b2092... has some of the gory details.

nnethercote··on The New Firefox and Ridiculous Numbers of Tabs
Firefox ships ~100 localizations. (See https://www.mozilla.org/en-US/firefox/nightly/all/ for the list.) I'm pretty sure this will be handled appropriately :)
nnethercote··on How we made compiler warnings fatal in Firefox
From the blog post:

"Set things up so that fatal warnings are off by default, but enabled on continuous integration (CI). This means the primary coverage is via CI. You don’t want fatal warnings on by default because it causes problems for developers who use non-standard compilers (e.g. pre-release versions with new warning classes). Developers using the same compilers as CI can turn it on locally if they want without problem."

nnethercote··on How we made compiler warnings fatal in Firefox
Firefox has a custom build system, and it's implemented within that. So it's not something that will translate to another project, alas.
nnethercote··on How we made compiler warnings fatal in Firefox
To be fair, the process of making them fatal was a lot more recent. Probably only took 3 or 4 years!
nnethercote··on Firefox 54: E10S-Multi, WebExtension APIs, CSS Clip-Path
Go to about:config and change dom.ipc.processCount to a high number. You'll get the behaviour you want.

Four is a reasonable default, especially given that multi-process is so new. It's possible that the default number may increase in the future, but that will depend heavily on memory usage.

nnethercote··on Goodbye PNaCl, Hello WebAssembly
Plenty of people were left scratching their heads when the new Google Earth was introduced, precisely because it was clear that PNaCl's days were numbered.

I think the likely explanation is that Google is a large company with numerous departments and those departments don't always pull in the same direction. I imagine some people within Google weren't happy with the Google Earth announcement.

nnethercote··on Browse Against the Machine
Adding multi-process support is the single biggest change ever made to Firefox. It was enormous. That's why it took so long.
nnethercote··on How Australia Bungled Its $36B High-Speed Internet Rollout
Er, what? Labor came up with the idea of the NBN and started it. The Coalition then stopped building it properly and switched it to the watered-down version when they gained power. The Coalition deserves the blame here.
nnethercote··on Fewer mallocs in curl
There is a little-known Valgrind tool called "DHAT" (short for "Dynamic Heap Analysis Tool") that's designed to help find exactly these sorts of excessive allocations.

Here's an old blog post describing it, by DHAT's author: https://blog.mozilla.org/jseward/2010/12/05/fun-n-games-with...

Here's another blog post in which I describe how I used it to speed up the Rust compiler significantly: https://blog.mozilla.org/nnethercote/2016/10/14/how-to-speed...

And here is the user manual: http://valgrind.org/docs/manual/dh-manual.html

nnethercote··on Simplifying Firefox Release Channels
Mozilla is the only company that builds a browser because it genuinely wants to, rather than as a market defense tactic.
nnethercote··on Simplifying Firefox Release Channels
I run Nightly on all my machines, and in my experience its stability is good. YMMV, of course.
nnethercote··on Firefox 52 released
This is a nice way to refresh your profile without losing all your history and other stuff: https://support.mozilla.org/t5/Procedures-to-diagnose-and-fi...
nnethercote··on Firefox 52 released
> It feels like Mozilla is keeping Flash around just because it's more popular with browser gamers and video watchers.

That's exactly right. The amount of Flash content dwarfs the amount of other plugin content.

Banning other NPAPI plugins doesn't avoid the Flash security problems but it does avoid the other security problems. It's a step forward, but it's not the final step.

nnethercote··on Mozilla Acquires Pocket
Pocket will be open sourced.
nnethercote··on Rust creator Graydon Hoare is now at Apple working on Swift
> under no uncertain circumstances are they allowed to submit code

Really? Why is that?

nnethercote··on Rust vs. Go
C:

  #include <stdio.h>
  int main(void) {
    printf("hello world\n");
    return 0;
  }
Rust:

  fn main() {
    println!("hello world");
  }
nnethercote··on Transitioning Firefox's rendering engine from Gecko to Servo
> I have animations that run at a smooth 60fps in Chrome and Safari effortlessly, quite often producing 1-2 FPS in Firefox.

Can you give an example? I'm sure Mozilla's animation people would love to hear about this.

nnethercote··on Rust is mostly safety
"I trust your binaries but I don't trust your shell script for installing those binaries."
nnethercote··on C++17 – better than you might think
Move semantics are so useful and powerful, but so complex. I've read multiple tutorials on them (all of them very long) and I still keep forgetting minor but critical details.

(You could probably replace "move semantics" in that sentence with multiple other C++ features and still have a valid statement.)

nnethercote··on Rust is mostly safety
Rust code is also going into Firefox, slowly for now, but it will speed up over 2017.
← PreviousPage 2 of 17Next →