I don't know if this is the right place to aks this.

[half-moon]

I've been told that Pale Moon the pale moon source code contains alot of the code from the mozilla source code. If that is true, then is it possible to compile seamonkey from the Pale Moon source code? What about Xulrunner from the Pale Moon source code? I'm just curious, but maybe I'd have some ideas if it is possible.

[New Tobin Paradigm]

XULRunner no problem. SeaMonkey is a bit of a wrinkle. Our version of comm-central that matches our codebase has been modified to compile with said codebase and right now that version of comm only builds with what has been termed 25LTS the sec updated branch used for Android and FossaMail. Additionally I stripped out all code and support for the suite.

It will not yeild a fully functional build with 26.x. Due to this situation and our divergent path stock comm-central will not build with our backend either.

We hope to reunify on Tycho and resolve this fully so that we are one codebase again. While this will allow the next milestone for FossaMail to build the suite is another matter.

I do have the skills to make this possible but I don't see the gains unless the council wanted to work with us which seems more and more like an idealistic fantasy on my part. Also, it would take time and resources to expend on such a project and would have to wait until Tycho has been completed and has had a first release.

[half-moon]

Ok, thanks for the reply. If I wanted to compile Xulrunner and release it, would I have to change the branding? If so, how would I change it? Also, how would I compile it? Is a dual-core system with 4gb (only 2.45 GB usable) of RAM sufficient?

I kind of wanted to compile Xulrunner since Mozilla is abandoning it and maybe I could compile for people who wanted to run chatzilla and other Xul apps. The main reason why, is because I'm bored and want to contribute something.

[New Tobin Paradigm]

We do have plans to do this our selves but they have been on the back burner. You do know you can launch any xulrunner application from the Pale Moon binary directly right?

After all what is Pale Moon but a private executable xulrunner application with a default application package.

The problem is it is literally easier for someone like my self to do this properly than to watch over someone else. You still have to build it properly and contributed builds of stuff without actual maintenance and development isn't helpful.

Are you a developer? Do you want to "own" and maintain the code for xulrunner? Are you going to be there to evaluate potential impacts of development against the stub?

Merely build and release just isn't good enough anymore and this really won't be contributing anything of substance to the project codebase as a whole.

You also don't have enough system resources to make very good if even successful builds. The linking process alone would eat the free ram up and then you system would start to page memory to disk which best case would be mind nummingly slow and at worse fail or produce unstable binaries.

[half-moon]

Are you a developer?

No.

Do you want to "own" and maintain the code for xulrunner?

I guess I couldn't even if I wanted since I'm not a developer.

Are you going to be there to evaluate potential impacts of development against the stub?

What do you mean by stub?

Merely build and release just isn't good enough anymore and this really won't be contributing anything of substance to the project codebase as a whole.

But would it really NEED to be maintained? Are you sure that build and release model isn't good enough?

You also don't have enough system resources to make very good if even successful builds. The linking process alone would eat the free ram up and then you system would start to page memory to disk which best case would be mind nummingly slow and at worse fail or produce unstable binaries.

Really?

[Moonchild]

Is a dual-core system with 4gb (only 2.45 GB usable) of RAM sufficient?

Matt A Tobin wrote:You also don't have enough system resources to make very good if even successful builds. The linking process alone would eat the free ram up and then you system would start to page memory to disk which best case would be mind nummingly slow and at worse fail or produce unstable binaries.

Really?

Yes, really. You need 4GB usable ram for guaranteed successful builds. Most of this is caused by Mozilla's aim to stuff everything into a single dll (xul.dll) that we build on, which makes optimized building a very memory-hungry process for a code base our size. It's been a fight against hitting available address space on x86 for quite the while as a result.
Running out of free RAM and having the OS start paging memory out to swap in the linking phase is something the linker can't really deal with, and it regularly borks on compiling properly as a result, creating unstable binaries.

[ketmar]

GNU/Linux: it is possible to build v26 with 2GB of RAM, no probs. i guess that v27 too. just throw out kde, gnome and other memory-hungry crap, so you'll have all that RAM available. and don't do LTO. also, ld.gold may fail or produce unstable builds, so it's better to switch back to old ld.

p.s. that was without swap at all.

p.p.s. actually, i often do private builds on my 32-bit OS with 4GB of RAM. i'm doing that on background, and have Pale Moon running in the same time, and alot of other apps.

[New Tobin Paradigm]

Even on linux the precaution stands. Just because it is possible does not mean it is a good idea. Also, this totally discounts Windows.

XULRunner needs a maintainer not merely a builder.

[ketmar]

Even on linux the precaution stands.

sure. i just said that it is possible if one will be careful. ;-)