5 ms·

I'm making browser games and there are some still some performance issues that make my game unplayable in Firefox, e.g. https://bugzilla.mozilla.org/show_bug.cg

by Kiro 4y ago

I'm making browser games and there are some still some performance issues that make my game unplayable in Firefox, e.g. https://bugzilla.mozilla.org/show_bug.cgi?id=925025 https://bugzilla.mozilla.org/show_bug.cgi?id=925025 I really don't want to have a "works best in Chrome" banner but until Firefox catches up I don't have much choice (don't get me started on Safari, I've already given up).

ghoomketu 4y ago
Same for me, my app used the html5 webspeech api and the text to speech and speech to text on Ff was nothing compared to chrome. So had to put that dreaded banner myself. I hope things have changed now and ff is just as good as chrome now with the recent progress in ML and so many good open source projects for tts and dictation.
the_gipsy 4y ago
I assure you, if you had developed against firefox first, you would have found "performance bugs" on chrome. Source: I developed a web game.
- koonsolo 4y ago
  For me, it seems with the latest update of firefox, some textures are not rendered properly. All works fine in the other browsers. Before that, some other update gave issues with some certificate. Looking at all the issues this browser causes, and the very low usage, I'm not spending time tracking down what is going wrong. Safari on iOS is such another problem child, but used high enough to justify the time to investigate and fix.
  - the_gipsy 4y ago
    The low usage is a fair point, but "all the issues this browser causes" simply isn't true. You would say the same about chrome if you developed for Firefox first.
    - koonsolo 4y ago
      Having updates breaking stuff is still a different beast.
      - rascul 4y ago
        Was that back when Mozilla removed the WoSign and StartCom for doing shady stuff? https://blog.mozilla.org/security/2016/10/24/distrusting-new-wosign-and-startcom-certificates/ https://blog.mozilla.org/security/2016/10/24/distrusting-new...
        koonsolo 4y ago
        No it was early this year so my guess was it was this one (happened at a user): https://support.mozilla.org/hsb/questions/1361315 https://support.mozilla.org/hsb/questions/1361315
      - the_gipsy 4y ago
        I thought Chrome leans strong on the "move fast, break things" side.
  - hansel_der 4y ago
    >Looking at all the issues this browser causes, and the very low usage, I'm not spending time tracking down what is going wrong. been hearing this for about a decade now...
    - koonsolo 4y ago
      Firefox usage in 2012: 24%. Now: 7.8% on desktop, almost non-existent on mobile, which brings it to a total of about 4%. Your point of "been hearing this" is what?
      - hansel_der 4y ago
        that it's been like this for a while and it changes/-ed nothing.
- jlokier 4y ago
  I agree, and will add that with ordinary HTML+CSS (not a game), I've occasionally seen Chrome/WebKit rendering bugs which I didn't see when testing during development with Firefox. Testing in Firefox is a pretty good way to approximate standards compliance, and then you have to deal with bugs when running on Chrome, if you do it that way around.
koonsolo 4y ago
Same here. Plus, I saw today that firefox doesn't support the file API while Chrome, Safari and Edge do. If you look at the usage numbers, it just makes no sense to support this browser anymore.
- koonsolo 4y ago
  Wow, stating facts gets you downvoted. Nice one! Here are the facts: - multiple game devs have issues with Firefox - Firefox doesn't support File System Access API https://developer.mozilla.org/en-US/docs/Web/API/File_System_Access_API https://developer.mozilla.org/en-US/docs/Web/API/File_System..., other browsers do. - Usage of Firefox is very low Live with reality people.
  - crtasm 4y ago
    > If you look at the usage numbers, it just makes no sense to support this browser anymore. It seems the reality is that people disagree with your opinion.
    - koonsolo 4y ago
      Probably the 4% of people still using Firefox, highly represented here on HN. Still doesn't change my assessment it's not worth it. Edit: especially since according to experience, when Firefox doesn't work for users, they just switch to another browser.
      - rascul 4y ago
        If it doesn't work in Firefox then I don't go to that site anymore. There's probably 2 other people in the world like me. Maybe 3.
        koonsolo 4y ago
        The thing is, as a solo developer, you just have to be ruthless with cost/benefit. Either I give the same experience to maybe 2% of the users, or an overall improvement to 98%. Is it ideal? No. But a lot of things are about tradeoffs and deciding where you spend your resources. Ok, people here don't like it. They can make their own tradeoffs. But in my opinion my choice is not that crazy.
butz 4y ago
How about adding "performance" mode and disabling advanced effects for worse performing systems? It could be done automatically, or by adding an option. So instead of "works best in Chrome", you would add short instructions how to set it. I think some players even prefer less resources demanding version of your game, especially on phones, where battery life is important.
- Kiro 4y ago
  It's a good idea but unfortunately the effect is instrumental to certain parts of the gameplay. However, as per the sibling comment it looks like it may actually have been fixed as of end of last year: https://news.ycombinator.com/item?id=32129916 https://news.ycombinator.com/item?id=32129916 I do have some other problems only happening on Firefox that I don't know the cause of (so probably something I can fix) but I certainly would like to remove that banner. Mozilla fixing the filter issues is a huge leap toward that and gives me motivation to look into the other problems.
solarkraft 4y ago
Ah, one of those eternal issues. I've seen a fair share of them, often well over a decade old.
muizelaar 4y ago
I've closed that bug as the problem as the originally reported problem is fixed by hardware accelerating filters as part of WebRender. Do you have more details on the performance problem that you're seeing?
- Kiro 4y ago
  Looks like WebRender has been enabled by default as of Firefox 92 (September 7, 2021) so that's great news! I think most of my performance problems were due to filters so I'm eager give it a try. I saw your comment on Bugzilla and will do that if I'm still experiencing issues after testing it. Thank you!

by Kiro 4y ago

ghoomketu 4y ago

Same for me, my app used the html5 webspeech api and the text to speech and speech to text on Ff was nothing compared to chrome. So had to put that dreaded banner myself. I hope things have changed now and ff is just as good as chrome now with the recent progress in ML and so many good open source projects for tts and dictation.

the_gipsy 4y ago

I assure you, if you had developed against firefox first, you would have found "performance bugs" on chrome. Source: I developed a web game.

koonsolo 4y ago
For me, it seems with the latest update of firefox, some textures are not rendered properly. All works fine in the other browsers. Before that, some other update gave issues with some certificate. Looking at all the issues this browser causes, and the very low usage, I'm not spending time tracking down what is going wrong. Safari on iOS is such another problem child, but used high enough to justify the time to investigate and fix.
- the_gipsy 4y ago
  The low usage is a fair point, but "all the issues this browser causes" simply isn't true. You would say the same about chrome if you developed for Firefox first.
  - koonsolo 4y ago
    Having updates breaking stuff is still a different beast.
    - rascul 4y ago
      Was that back when Mozilla removed the WoSign and StartCom for doing shady stuff? https://blog.mozilla.org/security/2016/10/24/distrusting-new-wosign-and-startcom-certificates/ https://blog.mozilla.org/security/2016/10/24/distrusting-new...
      - koonsolo 4y ago
        No it was early this year so my guess was it was this one (happened at a user): https://support.mozilla.org/hsb/questions/1361315 https://support.mozilla.org/hsb/questions/1361315
    - the_gipsy 4y ago
      I thought Chrome leans strong on the "move fast, break things" side.
- hansel_der 4y ago
  >Looking at all the issues this browser causes, and the very low usage, I'm not spending time tracking down what is going wrong. been hearing this for about a decade now...
  - koonsolo 4y ago
    Firefox usage in 2012: 24%. Now: 7.8% on desktop, almost non-existent on mobile, which brings it to a total of about 4%. Your point of "been hearing this" is what?
    - hansel_der 4y ago
      that it's been like this for a while and it changes/-ed nothing.
jlokier 4y ago
I agree, and will add that with ordinary HTML+CSS (not a game), I've occasionally seen Chrome/WebKit rendering bugs which I didn't see when testing during development with Firefox. Testing in Firefox is a pretty good way to approximate standards compliance, and then you have to deal with bugs when running on Chrome, if you do it that way around.

koonsolo 4y ago

Same here. Plus, I saw today that firefox doesn't support the file API while Chrome, Safari and Edge do. If you look at the usage numbers, it just makes no sense to support this browser anymore.

koonsolo 4y ago
Wow, stating facts gets you downvoted. Nice one! Here are the facts: - multiple game devs have issues with Firefox - Firefox doesn't support File System Access API https://developer.mozilla.org/en-US/docs/Web/API/File_System_Access_API https://developer.mozilla.org/en-US/docs/Web/API/File_System..., other browsers do. - Usage of Firefox is very low Live with reality people.
- crtasm 4y ago
  > If you look at the usage numbers, it just makes no sense to support this browser anymore. It seems the reality is that people disagree with your opinion.
  - koonsolo 4y ago
    Probably the 4% of people still using Firefox, highly represented here on HN. Still doesn't change my assessment it's not worth it. Edit: especially since according to experience, when Firefox doesn't work for users, they just switch to another browser.
    - rascul 4y ago
      If it doesn't work in Firefox then I don't go to that site anymore. There's probably 2 other people in the world like me. Maybe 3.
      - koonsolo 4y ago
        The thing is, as a solo developer, you just have to be ruthless with cost/benefit. Either I give the same experience to maybe 2% of the users, or an overall improvement to 98%. Is it ideal? No. But a lot of things are about tradeoffs and deciding where you spend your resources. Ok, people here don't like it. They can make their own tradeoffs. But in my opinion my choice is not that crazy.

butz 4y ago

How about adding "performance" mode and disabling advanced effects for worse performing systems? It could be done automatically, or by adding an option. So instead of "works best in Chrome", you would add short instructions how to set it. I think some players even prefer less resources demanding version of your game, especially on phones, where battery life is important.

Kiro 4y ago
It's a good idea but unfortunately the effect is instrumental to certain parts of the gameplay. However, as per the sibling comment it looks like it may actually have been fixed as of end of last year: https://news.ycombinator.com/item?id=32129916 https://news.ycombinator.com/item?id=32129916 I do have some other problems only happening on Firefox that I don't know the cause of (so probably something I can fix) but I certainly would like to remove that banner. Mozilla fixing the filter issues is a huge leap toward that and gives me motivation to look into the other problems.

solarkraft 4y ago

Ah, one of those eternal issues. I've seen a fair share of them, often well over a decade old.

muizelaar 4y ago

I've closed that bug as the problem as the originally reported problem is fixed by hardware accelerating filters as part of WebRender. Do you have more details on the performance problem that you're seeing?

Kiro 4y ago
Looks like WebRender has been enabled by default as of Firefox 92 (September 7, 2021) so that's great news! I think most of my performance problems were due to filters so I'm eager give it a try. I saw your comment on Bugzilla and will do that if I'm still experiencing issues after testing it. Thank you!