Ryan Gordon Is Fed Up, FatELF Is Likely Dead

[kraftman]

My response to that is why hasn't that happened already if the work is "easily" done?

Convince everyone to switch to rpms or debs.

[deanjo]

However, I prefer to install everything using package manager. If some third party member doesn't want to give me a package I consider it's their fault. :>

• Ship one file that works across Linux and FreeBSD (without a platform compatibility layer on either of them).

Why me or many other Linux user would want this? (except those who use bsd too)? ;p

And how many apps do you think are going to need this? It would be mostly used for commercial games and apps. BTW you can also remove the unwanted arches and os as well.

Q: Can I strip out the bits of a FatELF file that I don't need? If I only run an x86 system, I'd like to delete the PowerPC (or whatever) pieces.
A: Yes. Run fatelf-extract newfilename fatelffilename i386

[deanjo]

Convince everyone to switch to rpms or debs.

The LSB has stated RPM's for years. How is that working out for ya? It's made alot of progress hasn't it. While your at it, convince them all to follow the same symlinks and directory hierarchy. Oh wait a minute, ya that standard has been around for decades too.

[kraftman]

And how many apps do you think are going to need this? It would be mostly used for commercial games and apps. BTW you can also remove the unwanted arches and os as well.

So it seems it's not worth to do some changes in kernel just for few commercial games and some apps.

[kraftman]

The LSB has stated RPM's for years. How is that working out for ya? It's made alot of progress hasn't it. While your at it, convince them all to follow the same symlinks and directory hierarchy. Oh wait a minute, ya that standard has been around for decades too.

I don't know how it's working, but some distros prefer using their own package formats. Those who don't follow LSB can maybe loose some commercial apps and that's probably all.

[deanjo]

So it seems it's not worth to do some changes in kernel just for few commercial games and some apps.

Then don't expect to see any great push from commercial developers to support linux to further push linux adoption. Believe it or not, those quality commercial apps are probably the biggest reason for people not adopting linux. Sorry but there are still many apps that linux still doesn't have a comparable equal to in the FOSS world.

[kraftman]

Then don't expect to see any great push from commercial developers to support linux to further push linux adoption. Believe it or not, those quality commercial apps are probably the biggest reason for people not adopting linux. Sorry but there are still many apps that linux still doesn't have a comparable equal to in the FOSS world.

It can be true if FatElf would change this, but there can be some other reasons which maybe stops some companies to bringing their apps to Linux.

[deanjo]

It can be true if FatElf would change this, but there can be some other reasons which maybe stops some companies to bringing their apps to Linux.

I'm not saying it is the only reason, but given the efforts taken by Ryan to address it, I would say it is a big one otherwise he probably wouldn't have even started it.

My company for example decided not to make a native linux port of our application for the exact reason of a "shotgun" layout between distro's. Instead we did a java port. Sure it's slower then native solutions but supportwise it is a hell of a lot easier. The Mac and Windows versions however do have native ports because of the relative stability (compatibility wise) in those OS's. If we did a native linux port chances are we would spend more times actually troubleshooting the distro rather then the application.

[energyman]

And how many apps do you think are going to need this? It would be mostly used for commercial games and apps. BTW you can also remove the unwanted arches and os as well.

and completly screw up packet managment. 'uh, I didn't install this. I can't remove this' says the pm and your partitions fill up with garbage. Even worse, good pm don't overwrite unknown files. So after 'stripping' you can't install stuff anymore.

And your 32bit compat lib example is in itself a complete and utter FAILURE.

You still have / need all that compat crap. But this time it is all bundled up in one single fragile bninary.

Great, just more wasted space and more crap.

Just face it, fatelf is an idiotic idea without benefits and a lot of downsides.

[deanjo]

and completly screw up packet managment. 'uh, I didn't install this. I can't remove this' says the pm and your partitions fill up with garbage. Even worse, good pm don't overwrite unknown files. So after 'stripping' you can't install stuff anymore.

And your 32bit compat lib example is in itself a complete and utter FAILURE.

You still have / need all that compat crap. But this time it is all bundled up in one single fragile bninary.

Great, just more wasted space and more crap.

Just face it, fatelf is an idiotic idea without benefits and a lot of downsides.

But yet real life application of the concept has proven to be very successful and pain free. Figure that one out.