As the original author of GtkMozEmbed I'd like to say that this is way too little, way to late. Mozilla has historically been very much against anyone trying to use the engine in anything other than the firefox app.
This is completely up to them of course, and it probably let them move firefox forward faster, due to not having to care about APIs. However, it also means that chrome completely took over the entire market with electron and similar things.
I've always wondered why Mozilla doesn't make it easier to embed Gecko everywhere it possibly can. If Electron was built with Gecko, it would go a huge way towards preventing Chrome from becoming the de-facto way of accessing the Web.
As you said, this is too little, too late. I'm a Firefox user, but Chrome has pretty much won the browser wars. The only thing preventing it from achieving complete market dominance is the existence of Mobile Safari.
Supporting embedding well is a lot of work, both to create the APIs and then deal with their constraints as the browser evolves. Mozilla has always been under-resourced compared to the competition (IE, then Chrome) and taking people off improving Firefox/Gecko to work on embedding never made the cut. I don't really regret those decisions.
What made it extra unappealing for Mozilla was that we knew the Firefox/Gecko architecture needed lots of work that would destabilize an embedding API. Not much point in supporting an API that assumes single-process while you're moving to multi-process, or that exposes XUL which you know is a dead end, or that isn't compatible with off-main-thread rendering or GPU rendering, etc. Things are much better now that a lot of that debt has been paid off.
There's nothing stopping a 3rd party from making an Gecko version of Firefox. Electron is not run by Google nor the Chromium team. It's run by github (which only recently became Microsoft)
That said, I'm not sure what the point is. I believe (but could be wrong) that Chromium has more features that Gecko doesn't? So if you're choosing one over the other I'm only guessing most people would choose Electron
I don't have any direct insight into why the Atom people chose Chromium but I can see a few reasons that are more likely than "features":
-- Chromium was faster. (For Electron apps you don't really care about Web compat, where Firefox was ahead of Chromium for a while, but you do care about performance.)
-- Chromium was already multiprocess and had some other architectural advantages where Gecko was still catching up. So, less upcoming architectural churn.
-- Chromium had Google's resources committed to it. (Google's shine has worn off a bit since then but it's still a powerful effect.)
> Mozilla has always been under-resourced compared to the competition (IE, then Chrome) and taking people off improving Firefox/Gecko to work on embedding never made the cut.
Yeah, sure, imitating your competition is the right way to handle that. /s
Moving away from the unsustainable XUL extensions model ("no stable API boundary, just feel free to poke the internals of Firefox any way you like") was another piece of architectural churn that had to happen.
> If Electron was built with Gecko, it would go a huge way towards preventing Chrome from becoming the de-facto way of accessing the Web.
People coding specifically to Chrome when writing Electron apps is way less harmful to the web than people doing the same with web pages. The latter case specifically hurts other browsers, because people who don't use Chrome can't use the site.
Additionally, each Electron app represents a (smallish group of) developers. They're a very small number compared to all the developers writing public-facing websites that (should!) work in all commonly used browsers.
It makes a difference though. The fact that Chromium is Electron’s backend is why MS picked Chromium to base Edge off. It’s likely they would have chosen something else if that wasn’t the case (or maybe continued with their own web tech).
> It’s likely they would have chosen something else if that wasn’t the case
Brave is run by ex-Mozilla folks and started with Gecko, but switched to Blink very early on. [1] Their reasons didn't include Electron, so Microsoft's may not have either.
> The fact that Chromium is Electron’s backend is why MS picked Chromium to base Edge off.
What evidence do you have for this?
AFAIK the main reason why Microsoft picked Chromium is that Web sites are pretty much guaranteed to work well in Chromium without Microsoft having to do any work.
FWIW that is apparently for "rendering Web content" inside a Windows application, and that's what the examples do; not clear to me that it's exposing Electron.
Electron is nice for a quick and dirty app. Sadly, people think it's ok to use it for making apps that people are going to use.
What I mean by this - I can recognize a non-Elector vs Electron app by the amount of RAM it consumes. Is your simple color picker/text editor/REST caller/mouse config/ToDo app taking 1GB+ of RAM? It's an Electron app.
For the record, I doubt FF based app would face much better.
And the CPU usage spike as you type into a simple textbox. Without even looking at RAM usage. Text input lag and UI slowness gives it away at the first interaction.
At this point, I could almost tolerate the wasted RAM and CPU if interaction latency wasn't that bad.
Mostly because it is a lot of work and tends to lead to new and exciting compromises.
This is for example why Chrome decided to fork WebKit.
(I do believe that people messed around with Electron drop-in replacements, but it's probably hard to sell if it doesn't bring anything new to the table.)
It is expensive, but also it is a competitive advantage (of sorts), as we can see with electron. At the end of the day its just a strategic decision Mozilla made to focus on the application itself.
Considering the potential performance issues of Electron apps, I am still hopeful that someone can come up with an Electron alternative with the right engine, but I don't know if that engine is Gecko or not.
> I've always wondered why Mozilla doesn't make it easier to embed Gecko everywhere it possibly can. If Electron was built with Gecko, it would go a huge way towards preventing Chrome from becoming the de-facto way of accessing the Web.
"Quite successful" is perhaps an exaggeration. Most of the projects using XULRunner were directly related to the Mozilla project (forks of Mozilla Navigator and Composer, Thunderbird, Chatzilla, etc). There are very few projects which independently chose to use XULRunner.
> As the original author of GtkMozEmbed I'd like to say that this is way too little, way to late. Mozilla has historically been very much against anyone trying to use the engine in anything other than the firefox app.
Your comment reminds me of a 15 or 20 year old remark from Feeddemon and Topstyle author (Nick Bradbury) about how he gave up baking firefox preview into his CSS editor.
I loved Topstyle. It was the first editor I invested time into (after Visual Basic for windows 3.11 in the 90's but I was a kid, not a professional or a student).
Yeah, the Sailfish OS project, which is an operating system for mobile devices. They use Gecko as engine for their default web browser, but are always lagging with engine updates as every time they have to painstakingly rebase all of their embedding patches on top of the new Gecko codebase.
Thank you so much for creating GtkMozEmbed. I spent a good part of 2004 working with it. The most fun I had was hacking it to run on Windows. I recall having to inject a callback function into the win32 event loop with assembly. It was like the coolest just out of college project ever for me.
> As the original author of GtkMozEmbed I'd like to say that this is way too little, way to late. Mozilla has historically been very much against anyone trying to use the engine in anything other than the firefox app.
This has been in the works since at least 2016. I believe conversations around improving the embedding story were happening before that, too.
I wonder if they changed the licensing to something you can actually use in a non FOSS product. Way way back (8 years ago?) we were deciding between webkit and gecko for our solution (we needed a web engine for server side GFX rendering) and the licence made it very simple for us to decide - WebKit was MIT, Gecko wasn't.
Wasn't WebKit LGPL (because of the KHTML heritage)? I was under the impression that, for commercial usage, Gecko (with MPL) was more friendly. Either can be made to work, of course.
> we needed a web engine for server side GFX rendering
> the licence made it very simple for us to decide - WebKit was MIT, Gecko wasn't.
...So what? If it runs on your server, it could be just about anything and it wouldn't matter so long as it's not AGPL. Or was the server meant to belong to the customer?
It can be used in proprietary software. You shouldn't, because that's unethical, but the MPL isn't a strong copyleft. And even with strong copylefts, there are only three that I can think of that apply to software that won't end up on user devices.
Preach. Not only is Gecko a nightmare to embed, Firefox is terrible for headless. It's why something like phantomjs even had to exist.
An embeddable Gecko would've created a wonderful ecosystem around it, produced headless browsers, a much wider range of browser alternatives, and had much more leverage when it comes to standard. They had... 8 years? 5 years? on Chomium and now have one hell of a task to correct that error.
Looking at the code of their new layout engine written in Rust, making it embeddable wasn't even remotely on their minds either. A real pity.
..for the 90s. There was very little about XUL that would survive today other than its layout model
Ignoring that XUL has a single implementation and for the most part a single application suite actually using it (not 100% true but very close), I'm not sure what there was to blow.
OTOH I remember some custom XUL apps of yesteryear, and they were always pretty nice to use. Does Firefox have any XUL left? I know there have been ideas to rip it out for a very long time.
I remember there was a whole Linux distro written in XUL. Can't remember its name for the life of me, but from memory it was vaguely based on RedHat and was targeting kiosks and internet appliances. I remember setting it up for my parents circa 2001 as it looked fool proof and tamper proof.
What's the issue with that? GTK, WPF Qt, Flutter ... all have a single major implementation.
> for the most part a single application suite actually using it
That is now, but it could've been a lot different.
I'm not that familiar with XUL (I'm just old enough to experience it being phased out) but it appears like a great idea to me: XML is not exactly bad for a UI, but XAML is Microsoft-specific (and not exactly great) and HTML is made for documents (leading to problems with worries about semantic tags, lack of layout features ...).
I also quite like how well Firefox integrates with all kinds of platforms. But who knows, maybe that's not in XUL, but specific to FF.
Remember XULRunner that was doing Electron stuff about 6-7 years before the first version came out. Mozilla has so many good ideas and such terrible execution.
The problem is that XUL was a) nonstandard and b) pretty badly designed and implemented. (For example: XUL trees relied on Javascript callbacks being called at paint time in horribly unsafe ways; XUL flexbox never integrated well with CSS block layout.) It was mostly hacked together in a hurry to get Netscape 6 out the door (and we know how that went). It was XML-centric and the Web ended up going in a different direction. All in all, it was clear pretty early on that XUL was going to be a dead end. Most of the good parts of XUL were standardized as CSS features (or in a few cases, HTML features).
Given more resources, Mozilla absolutely could have done a better job of transitioning faster from XUL to standards-friendly equivalents, and popularizing the latter in an Electron-like framework. But there was never a time when Mozilla had excess resources floating around, and other things took priority. (Also see my comment above about how architectural churn made it unattractive to support a stable embedding API.) As was pointed out in other comments, the benefits of regular Web market share tend to outweigh the benefits of embedding popularity.
Maybe they lacked resources and dev mind share ? Javascript of those days wasn't javascript of today ? “Apps” were almost just glorified shortcuts for some.
Yup, I built my little browser on Chromium due to ease of embedding though I would have loved to use FF.
I am investigating a way to use FF on system as is though by providing a custom profile, a user.css to remove everything but the browser window itself (i.e. no tabs, toolbar, etc), capturing it as a native window (e.g. via QWindow::fromWinId and QWidget::createWindowContainer), and communicating with it via marionette or other remote approach. So far it seems to be ok, but it's a bit of a hack.
Not sure what language you are using but CEF was good for Chrome under C#. Might be available under C++ too I think.
There was an equivalent Gecko one for Firefox but I could never get it working correctly so deferred to CEF under Visual Studio (think there's a nuget package for it, conveniently).
Also wxWidgets under C++ offers ability to host a local WebView instance or Trident with wxWebView.
Have you tried calling it with the same profile multiple times? I've been running into "Firefox is already running" a lot while trying to use it to integrate web apps with the OS ("site specific browser"). Also the WM-Class would never be set according to the command line parameters on the second call, but you probably don't care about that.
For my use case I don't want to call it w/ the same profile multiple times. Once I have the FF process, I'll work with tabs/windows via the remote protocol.
This is completely up to them of course, and it probably let them move firefox forward faster, due to not having to care about APIs. However, it also means that chrome completely took over the entire market with electron and similar things.