Servo Browser Engine Making Embedded App Progress With Tauri

[kpedersen]

I think it is a good idea. People apparently like web technology for layout so it makes sense for them.

Weirdly, beyond all the hype we don't have too many standalone / embeddable web technologies suitable for this. I think only webview/electron are the ones in current use.

I feel Microsoft missed a trick with their easily embedded .NET WebBrowser component for creating desktop layouts. They never really took that route.

I would personally like to see an embeddable NetSurf renderer instead. Yes, it is more limited in terms of HTML features than Chrome's backend but it would be so much more light and portable.

[vextium]

Servo Browser when :3

[Pajn]

Please help me to understand Servo. Is this engine a kind of lib that can be used in your own project so you don't have to reinvent the wheel (a browser engine) if you need some web browsing functionality? So would it be possible to write a new 'lynx/elinks' based on Servo? Or any program that has all the features of Firefox without the GUI?

I'm asking this, because last summer I was looking for a solution to 'dump' a whole fully rendered webpage from the memory into a html file from Firefox (or any modern browser) but there wasn't such an option. The best would be to dump it to PDF which was useless in my case. ('File > Save as...' ended up without the needed parts of the original HTML page, but because it was on my monitor, it was obviously somewhere in the RAM so I expected the browser that it has a way to dump it, like passing some command line parameters with the URL or something, but no).

You seem to be quite confused in what you want, curl or wget will get you all rendered html into an html file. What's on your screen depends and a lot of more things than the html.

[rmfx]

This is good news. With more performant front-end frameworks and more lightweight alternatives to Electron, we can see a bright future for universal web/desktop apps.

Hum, so far, Servo doesn't look like the super light web engine that's multiple times leaner than Blink or WebKit, the binaries they ship are already around 250MB :S

[User29]

You seem to be quite confused in what you want, curl or wget will get you all rendered html into an html file. What's on your screen depends and a lot of more things than the html.

No, I'm not, thank you.

There are pages with content which are dynamically generated and not even curl or wget can dump the whole content because those tools are not able to render the page properly. Firefox and others are exporting some skeleton only.

No, I can't give you examples because it's an internal project.

[Pajn]

No, I'm not, thank you.

There are pages with content which are dynamically generated and not even curl or wget can dump the whole content because those tools are not able to render the page properly. Firefox and others are exporting some skeleton only.

No, I can't give you examples because it's an internal project.

You don't render to html, if you want something rendered you need a format like pdf, or a screenshot. Your question is like wondering why gtk apps can't be saved as an .xml file since glade is a thing.

[Quackdoc]

Servo Browser when :3

the minibrowser while basic (doesn't support tabs) does exist. It probably wouldn't be too hard to expand it to the scope of gnome web, which I personally really hope we wind up seeing.

Hum, so far, Servo doesn't look like the super light web engine that's multiple times leaner than Blink or WebKit, the binaries they ship are already around 250MB :S

there is quite little optimization being done with servo. but a lot of that size is media stuff. Lets download and extract servo-latest for windows and take a look at the resulting folder. You will notice that servo.exe is only around 88MiB and the rest of the 1Mib+ is media related stuff from gstreamer, and some other smaller basic stuff like angle bins and glib stuff. If you want to, you can build without gstreamer which will massively cut down on the package size. Im not sure if you can cut angle stuff but maybe you can?

dua | tail -n 17
   1.04 MiB libvorbis-0.dll
   1.13 MiB libvorbisenc-2.dll
   1.23 MiB gstreamer-1.0-0.dll
   1.34 MiB glib-2.0-0.dll
   1.36 MiB avfilter-8.dll
   1.71 MiB gio-2.0-0.dll
   2.04 MiB libpng16-16.dll
   2.28 MiB libjpeg-8.dll
   2.63 MiB gstvpx.dll
   3.28 MiB libcrypto-1_1-x64.dll
   3.90 MiB avutil-57.dll
   6.18 MiB libGLESv2.dll
   9.46 MiB resources
  32.35 MiB avformat-59.dll
  88.05 MiB servo.exe
  92.73 MiB avcodec-59.dll
 266.44 MiB total​​

[Amaranth]

Please help me to understand Servo. Is this engine a kind of lib that can be used in your own project so you don't have to reinvent the wheel (a browser engine) if you need some web browsing functionality? So would it be possible to write a new 'lynx/elinks' based on Servo? Or any program that has all the features of Firefox without the GUI?

I'm asking this, because last summer I was looking for a solution to 'dump' a whole fully rendered webpage from the memory into a html file from Firefox (or any modern browser) but there wasn't such an option. The best would be to dump it to PDF which was useless in my case. ('File > Save as...' ended up without the needed parts of the original HTML page, but because it was on my monitor, it was obviously somewhere in the RAM so I expected the browser that it has a way to dump it, like passing some command line parameters with the URL or something, but no).

Isn't this what extensions like SingleFile do?

[pizzapill]

There are pages with content which are dynamically generated and not even curl or wget can dump the whole content because those tools are not able to render the page properly. Firefox and others are exporting some skeleton only.

I had a project that needed to scrape URLs with JS content (especially dynamic links). I had to set up a Selenium Cluster in K8s for this, which is a enormous rabbit hole. Look into WebDriver/Selenium if you want to do this automatically and headless.

If you want to do it from your desktop Browser I would just write a simple Firefox/Chrome extension, pretty sure you can just grab rendered DOM with it.

[flakmirror]

No, I'm not, thank you.

There are pages with content which are dynamically generated and not even curl or wget can dump the whole content because those tools are not able to render the page properly. Firefox and others are exporting some skeleton only.

No, I can't give you examples because it's an internal project.

There are extensions that do one-click "save entire page"-screenshots, or you could go full Selenium if you need automated. There's also Firebug (for Firefox at least) that can view generated HTML after scripts have processed the page.