Gnash (Flash for Haiku)

Is development on this project continuing? I haven’t been able to find any recent news.

As fair as I understand it the linux Gnash development is still ongoing. They are just slow and carefull. I am not sure if there is an interest to port Gnash to Haiku, but that would be interesting to see.

You mean like these news bites?

(that last one actually shows gnash running on Haiku)

Yes. It was on the IsComputerOn site that I first read about Gnash for BeOS.

How’s the gnash port doing? I’d happily contribute to a gnash-for-Haiku bounty if that would help.

A webkit-based browser with gnash-support would probably fill the biggest feature gap for Haiku today.

Me too. And I agree about the lack of flash support. It is one of the main reasons that people I encourage to use BeOS end up abandoning it.

In case those of you missed it. There won’t be any Gnash for BeOS or Haiku.

I always wondered why somone didn’t come up with an open source alternative to Flash. This way people wouldn’t get stuck using Flash. I’m guessing that since Windows, Linux & MacOS X were supported that most people were happy enough with letting Flash be.

Looks like TubePositive is the way instead to watch Flash based videos.

Haiku is an OS of the future. BeOS 5 and GCC2 are the past. Look to the future, otherwise this is simply going to get more and more common. I mean, boost is a very common dependency in open source software.

I won’t pretend to know all the issues surrounding needing to use GCC2, but I assume it has to do with drivers and old software. Well the former is getting obseleted bit by bit by open source drivers and the latter - well, software that can’t be ported to GCC4 is only going to bitrot.

Perhaps the Haiku team should consider a clean break?

There were various discussions about getting GCC4 to Haiku. From what i understand, they need to compile under 2.9x to keep compatibility with BeOS.
Several developers have been asking to switch to GCC4 in order to port their applications (NoGravity,SkyORB, WebKit, etc…) . I’am sure the break-up will happen, but isn’t possible to have a double compatibily for R1.x ?

well the whole point of wanting haiku to work with old software is that you have a constant standard to test haiku with. otherwise you have a changing alpha Haiku with changing alpha Haiku software and very hard to make a proper comparison if something doesn’t work right. Haiku inc needs to prove that it can deliver and support a final release of an OS in order to take the next step.

At least that is my understanding for why they want binary compatibility. I don’t think that anyone thinks that any abandoned closed source BeOS is really that amazing in 2008.

Haiku is a volunteer-based community project, so both delivery of any release of Haiku and its support will be community-based. If anybody needs to prove anything with regards to delivering and supporting Haiku, it is the community itself and not Haiku Inc…

There is no magic wand: the success of Haiku will heavily depend on the degree to which community members are willing to contribute their time and skills to the project. Open source at its best.

The reasoning for using GCC 2.95.3 is:

  1. Make Haiku as compatible to BeOS R5 as possible. This is made easier by using same GCC as BeOS & running “older” BeOS programs.
  2. Allow closed source BeOS applications to run on Haiku. BeOS programs, done in C++ with gcc2.95, will not work on a gcc4 Haiku unless they are recompiled. If all software was open source then this wouldn’t be an issue.

The idea is to get out a 2.95 Haiku R1. At which point it’ll attract more developers and they can start working on new software or replacements to BeOS programs. Then with R2, switch to gcc 4.x & these programmers can recompile their stuff.

Going to gcc4 would be better because it’d make it easier to write the programs ( support newer features ), compile them and even in regards to porting - add newer software. Gcc2.95 is too outdated and even gcc3.4x is getting outdated too. Who knows, by the time Haiku R1 is ready to be released they may decide to do only a gcc4 version.

Or they could do two R1s, one 2.95 & one 4.x. List the 2.95 as BeOS compatible & the 4.x as BeOS Incompatible. Or they may be able to do a hybrid system, one that is gcc4 compiled but able to run gcc2.95 programs. Time will tell.

there still must be a top down decision: “this is what Haiku r1 will look like and that’s it.” Many projects are in perpetual beta because they can’t commit to anything.

When the time comes, the development team, not Haiku Inc., will make that decision. And since the development team is comprised by the volunteer developers, it is going to be more of a consensus-based decision rather than a top-down one.

Uhm, Jorge, wasn’t it decided that there will be an option to have old binaries and new binaries on the same system, the old binaries linked to libs compiled with the old gcc and the new ones linked to the same libs compiled with gcc 4.x?

I was under the assumption that the top Haiku developers = Haiku Inc. For example says Axel is the president.

I know it has been discussed, but I am not sure whether a final decision has been made. Anyway, I was making more a general statement with regards to the “Haiku inc needs to prove” comment, not responding specifically to the GCC compatibility thing. See below.

Haiku Inc. is currently composed of developers, although that may not necessarily be the case in the future. The point is that it is not the role of Haiku Inc. to be directly involved in the decision making related to development, distros, etc. As stated in this post…

…it is the development team that makes all development-related decisions, including the creation of one or more eventual Haiku distributions.

The “Haiku Inc. will have to prove” comment seemed to transpire a common misunderstanding that Haiku Inc. controls every aspect of the project and is in a position to make top down decisions and tell the volunteers what to do and when. I just wanted to point out that that is simply not the case. :slight_smile:

I thought Haiku Inc. = Jorge Mare. :slight_smile:

And that top down decisions were made by Jorge. :slight_smile:

I really don’t know what I must have been thinking.

Ok, let’s get serious here. Gnash was being ported by Michael Lotz. This was his side project. It was so he could give Flash to Haiku & BeOS. I’m pretty certain it wasn’t a decision by the developers of Haiku to port over Gnash. Something Michael was working on individually and decided was not worth his time pursuing any further. Maybe some other programmers might have interest in picking up where he left off and continuing the port? It depends on if they think Flash support is important enough to have.

Currently I believe the consensus is to make a gcc 2.95 system. If a hybrid is doable with gcc 4 system/libs & 2.95 libs then I’m sure they’ll look at doing this instead ( because it makes more sense to have gcc4 system to get newer software ported over ). Haiku still has a few years before R1 so what is today might change in the future. There was a discussion on a hybrid system but I don’t think they reached a definite decision - like Jorge said. Maybe we’re wrong? Hybrid system would be the best solution overall. Next, it’d be a toss up between 2.95 or 4.x because on the one hand you get to use older BeOS software but harder to port newer software over and on the other hand you get the opposites. These are both win-lose situations, compared to a hybrid system which is a win-win.

You are taking what I said out of context. Please take a look at what I am quoting in my first reply to arielb, and you will see that I was specifically referring to his “Haiku Inc. needs to prove that…” comment, and not anything else.

I do plead guilty to being off-topic, and for that you can trout-slap me ten times. :slight_smile:

For what it is worth I will push hard for a hybrid system and do what I can to make that a reality. Otherwise my WebKit port won’t have anywhere to run unless I make my own Haiku distro, which I would rather not have to do. Plus there is plenty of other useful software which requires GCC4, such as libtorrent which I would also like to port to Haiku.

Regarding Gnash, it is really a shame that the code is such a mess, and I trust Michael’s word if he says it may not be worth porting anymore. It is such a shame that there isn’t a nice open source Flash replacement that could gain some popularity, at least for video. The HTML5 stuff offers some hope, but given how much the development of the web is limited by IE, we may not see HTML5 widespread for a while. Though the IE team seems to finally be getting their stuff together, so maybe it isn’t totally hopeless.

Of course Microsoft has their Silverlight framework so it probably is not in their interest to promote something “open” like HTML5 when they could be pushing their proprietary stuff. Which again seems to me that we need an open source alternative, and by that I don’t mean Moonlight (the Mono team’s implementation of Silverlight.) It is a tough situation.