Stories
Slash Boxes
Comments

SoylentNews is people

posted by Fnord666 on Sunday August 13 2017, @12:08PM   Printer-friendly
from the firefox-loses-yet-more-users dept.

Firefox 57, which is slated for release on November 14, will "only run WebExtensions", according to Mozilla.

This is expected to break compatibility with many existing Firefox extensions, and in many cases there aren't WebExtensions-compatible alternatives available for these extensions.

During some recent discussion at Slashdot, it became clear that some users have nearly all of their extensions classified as "legacy", and susceptible to breakage.

Members of the SoylentNews community, if you use Firefox, how many of your extensions are set to no longer work with Firefox in the near future?

If Firefox 57 breaks compatibility with your existing extensions, will this finally be enough for you to discard Firefox and find an alternative browser to use?

Will this extension breakage, and subsequent loss of users, effectively end the viability of Firefox as a modern web browser?


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0) by Anonymous Coward on Sunday August 13 2017, @01:10PM (7 children)

    by Anonymous Coward on Sunday August 13 2017, @01:10PM (#553234)

    There was a sound server called ESD. It was so bad Alan Cox had to call it out, everyone hated it. It had 500-2000ms latency at the best of times. Eventually it gained renown as the worst idea anyone ever had for audio, and died. Just kidding! It was renamed to Pulse Audio and nearly every major vendor who ships on Linux supports it and nothing else. Microsoft, Mozilla, commercial VOIP software, etc.

    I have no love for systemd, but it was the very best idea that shitstain ever had.

  • (Score: 3, Insightful) by coolgopher on Sunday August 13 2017, @01:55PM (6 children)

    by coolgopher (1157) on Sunday August 13 2017, @01:55PM (#553246)

    Ah, I think you're slightly off the mark here. ESD was the Enlightenment Sound Daemon, and part of the Enlightenment desktop. Probably the neatest, coolest, flashiest-yet-still-highly-usable desktop of its time. Except the hardware to run it properly wasn't to be invented for another 5-10 years - everything had high latency :( I really tried to like it...

    PulseAudio on the other hand, well, it also had (has?) the same crap latency. And, in true Lennart style, it sits on top of everything like a fat frog and slurps everything through it. So if you're unlucky to have a program which outputs via ESD, that would end up getting both the ESD latency *and* the Pulse latency.

    PulseAudio is however not derived from ESD, to the best of my knowledge. Wikipedia seems to back me up on this: https://en.wikipedia.org/wiki/Enlightened_Sound_Daemon [wikipedia.org]

    And just to make it clear, I'm neither an ESD nor Pulse expert - neither of those daemons have been allowed on my systems for a very (very) long time. ALSA, possibly with the OSS compat layer, is generally sufficient for my needs. For heavier stuff, there's JACK.

    • (Score: 0) by Anonymous Coward on Sunday August 13 2017, @03:02PM (5 children)

      by Anonymous Coward on Sunday August 13 2017, @03:02PM (#553271)

      Wikipedia doesn't say anything about the relationship, other than ESD being outmoded by it in gnome. It's hard to find links on that ancient POS though. The point is they do share lineage, ESD was a piece of shit in its day, everyone knew it, and it would be just about as bad today—seriously, you don't need much to mix audio, the amount of latency it had was not because the cpu was slow—I've been doing it since the 286. Besides, aRts had similar features (and then some) and good latency under normal usage... and it's not exactly well-loved either. Pulse is bad for the exact same reasons as ESD: dipshit basically put every feature into the main code path. That's why he wanted to start fucking with the Linux kernel, separation of concerns is not in his vocabulary.

      • (Score: 5, Interesting) by TheRaven on Sunday August 13 2017, @04:01PM (3 children)

        by TheRaven (270) on Sunday August 13 2017, @04:01PM (#553291) Journal

        Linux deserves a lot of the blame. The Linux OSS implementation allowed multiple channels, but only with hardware that did sound mixing. This completely missed the point of a kernel: to provide abstractions for userspace. Only one application could open /dev/dsp and so you needed something in userspace to do the mixing. Unfortunately, GNOME and KDE, in their typical fashion, implemented incompatible ones.

        The next version of OSS, OSS 4, did support in-kernel sound mixing, but was proprietary. The reaction of the Linux kernel developers was to replace it with ALSA, which had a lot of extra complexity, was harder to program for, and didn't provide useful OSS compatibility (still no sound mixing).

        In contrast, FreeBSD took the last BSD-licensed version of OSS and update it to support the newer interfaces. They added in-kernel sound mixing, so multiple apps could open /dev/dsp and get their own independent channel. No need for any userspace mixing, no need to rewrite existing applications.

        --
        sudo mod me up
        • (Score: 0) by Anonymous Coward on Sunday August 13 2017, @07:28PM (2 children)

          by Anonymous Coward on Sunday August 13 2017, @07:28PM (#553337)

          I don't think BSD's approach made the difference. If BSD was as popular on the desktop as Linux it would attracted the same shitstains and end up with the same sound server problems. You've still gotta have PulseAudio if you want Skype, etc.

          It's hardly fair to blame Linux when half the time the sound server isn't serving the client or the hardware, it's just sitting there in the middle, taking up space. People want to write these deep stacks. Look at Android, with ~120ms latency typical not that long ago, on what is essentially a proprietary stack, the devs just don't care. When they realized this was a problem for musicians, they started targeting 50ms. (Musicians want about 9ms.) As far as latency in sound stacks go, developers by and large don't care, or even when they do, are so out of touch they miss the point.

          • (Score: 2) by kaszz on Monday August 14 2017, @03:09AM

            by kaszz (4211) on Monday August 14 2017, @03:09AM (#553447) Journal

            Skype is a spy tool for En-Ez-Ay that delivers a copy of your face every 5 minutes to their storage. So not much reason to keep it around.

          • (Score: 1, Informative) by Anonymous Coward on Monday August 14 2017, @11:35AM

            by Anonymous Coward on Monday August 14 2017, @11:35AM (#553597)

            Musicians want about 9ms

            Most places say 20ms is the limit before it becomes audible, but that's end to end. E.g. if you are using a soft synth app, that's 20 ms from the keyboard, through the midi router, through the USB driver into the midi stack, into the app, generating the sound, back out through the sound stack to the line out.

            So yeah, 9ms for just the audio stack is probably on the high side.

      • (Score: 2) by coolgopher on Monday August 14 2017, @03:05AM

        by coolgopher (1157) on Monday August 14 2017, @03:05AM (#553443)

        Oh I sure aren't disagreeing on the idiocy of the fat-toad-monolith approach here! Pulse did replicate it all on its own though, with a dedicated plugin/interface to allow it be an ESD client. Because, you know, peeps all love latency, so let's have some latency with that latency, yo.

        Man what a trip down memory lane here - I'd completely forgotten about artsd!