Many OS X users are anxiously awaiting a binary release of 2.5 for their platform. I thought I should update everyone on the situation there.
There is now a totally alpha-test-build available to verify that we’re back to at least reasonable functionality. Details in the IRC channel as normal.
For better or for worse, I upgraded my version of the GTK GUI toolkit on OS X to keep up with the many subtle (and not so subtle) improvements to the way that GTK works on OS X. After several days of wrangling with a major blocking bug (GTK could no longer load PNG image files used for icons and more), I got Ardour to build and run. However, it is substantially less stable than the 2.4.1 release, and most of the reasons appear to have to do with changes inside GTK itself.
We have a difficult situation with GTK in that few of the core GTK development team use GTK on OS X and there is a tendency for them to introduce new features/improvements/changes without verifying that things work. This is really as it should be, given that we are using an unreleased development version of the toolkit. However, I am having to give some serious thought to how we (the Ardour project) can use a version of GTK that is both stable and yet still benefits from new work done on the OS X specifics. I have spent too much time resolving issues with build systems and untested GTK code when I’d rather be making Ardour itself a better program.
It may be that I can get things working reasonably well within a few days - I will keep you all posted one way or another.
I know it’s too late know and probably not a constructive criticism but… are you sure GTK+ was the right choice for a crossplatform project? By using Qt4 or wxWidgets you probably would not have had to think about toolkit portability issues.
I don’t think Ardour was ever intended as a crossplatform project, but was made solely for linux. As it happened along the way OSX came around and with its unix backend presented itself as a viable alternative.
If I’m not mistaken it was only recently you could run Ardour nativly on OSX, without X in the background.
Its easy to speculate in hindsight. I would make 3 comments. First, its true that there are some reasons to believe that Qt might have turned into a better choice, but this ignores the fact that in 2000, Qt was still very un-C++ like and was at that time absolutely not acceptable. Second, Qt has its own problems with the transition to Cocoa - their current native implementation on OS X is based on Carbon, which Apple has recently officially deprecated. If you read the Qt blog about the state of their Cocoa version, they have almost the same set of basic problems that GTK does. Notice that my issues with GTK portability are not related to Ardour code … they relate to the fact that someone has to get GTK working correctly on OS X (the native version, that is), and right now nobody is going to deal with the parts that are broken except for me. My guess is that at present Qt is in a similar state. Its unclear if any more people would be helping with the equivalent issues in the Qt/Cocoa implementation than with the GTK one.
Finally, in terms of porting: forget it. Ardour’s code is dominated by its GUI, and I have no plans to port 75k+ lines of GTK code to Qt. Porting from GTK1 to GTK2 took way too long; porting to Qt would be a nightmare. Not because the core stuff wouldn’t work, but because there would be some new things that became possible and some other things that became very hard (just like the GTK1->GTK2 transition). We would get bogged down in that, I believe.
As Paul states… this problem is as it should be. Development for GTK on Linux is going to be brisk and needs to be. Still… too bad it is so difficult to keep it somewhat stable in OS X. Not the GTK projects fault really… more a lack of man power with know-how on the OS X side. Paul - As I’ve said to you before, I find what you have done with GTK-OSX in the past impressive! I hope I can help out more in the future.
Awesome! This is so exciting!
I’m giving another Ardour workshop in Toronto in late august, hopefully 2.5 will be finished by then. Either way I’m very happy. Great work!
~bennyp
Not caring about OSX one bit. I hope the work on OSX is not affecting the pace of ardour’s development for linux. I am really amazed by what efforts you devs are producing to get ardour to work on alternative OS’s. This is of course great for users of these OS’s, especially if they see ardour as more interesting than commercial / proprietary softwares. However, linux cannot say it enjoys software DAW alternatives that are as good or better than ardour. So I would see linux as a priority since it all started from linux and for linux. Anyway, I don’t want to bother OSX users here, I just thought I would express my opinion about ardour’s development.
I’m with linkx in thanking you for the update. I tried to compile the source following the instructions, but having never done anything like that before I got completely lost and decided that I’m better off writing the music and waiting for the compiled app. I look forward to it.
Hmmm…no that wasn’t even a little bit of sarcasm…not
one bit.
I do care about development on OSX…more then a little
bit.
I really do plan on donating a significant amount sooner
rather then later.
However that kind of attitude makes me rethink that
idea.
If that kind of perspective is pervasive amongst users
of Ardour…well I hope it’s not like that among the
developers.
hey! why do you take it like this ? You’re free to express your opinion as I am free to do so. If now the ardour devs planned their work according to one anymous user’s voice, they wouldn’t get that far. Take my post for what it is : my non requested opinion on the huge work effort to make ardour work on OSX. If you use OSX, that’s great, it’s your choice and the fact that ardour is ported to it is great news to you. To me, it is not because it implies a lot of work just to keep up with ardour’s development state on linux, i.e. the linux port does not benefit from the time invested in OSX. But who cares about what I think ? But you seem to care since you are reconsidering your initial move to donate to ardour’s devs, and that I find weird.
I don’t speak for the Ardour project, but I am somebody who has started to support it financially (in a small way). If you are thinking about contributing, either code or money, you should talk with the developers about what you have to offer and how that matches their needs.
My commitment to the project is based on what the project has done, what they are doing, and their commitment to creating a world-class free/open source DAW. Random and divergent opinions from the user community is just a part of the process, and one that underscores its authenticity.
I encourage any and all to join me in stepping up their levels of financial contribution (as I did last month) so that we can all see more from the Ardour developers.
actually, in general you are wrong about the relationship between the linux and the OS X versions. the linux version does benefit from the OS X port, if only in terms of us discovering more bugs and fixing them. but it also contributes in a less countable way, by providing insights and feedback into design issues, integration issues, plugin support and more - things that don’t happen or can’t happen as easily when you work only on a single platform.
that said, it would be nice not to have to deal with so many low level GTK issues. on the other hand, there is no evidence that any other cross-platform toolkit (including Qt) would have made the task of getting Ardour running on top of Cocoa any easier. and the work i’ve done is for the benefit for all GTK apps that might want to port to OS X, so its very very far from wasted.
finally, keep in mind that the financial support from SAE (who are the primary instigators of the OS X native version) is substantially more than is received from all subscribers and donors in more or less every month. i actually do not spend most of my time on the OS X version (or haven’t since about April), but given that the people who pay me to work on Ardour are voting with their dollars and euros, it is something i need to pay attention to. of course, if every user of ardour paid a dollar per month to use it, i wouldn’t be reaching out to SAE and others to help fund my odd lifestyle and this software, but they do not.
the most important point is, though, that development on either platform benefits the other, and i fully intend to see them both continue to move forward. whether the financial picture is viable remains to be seen. at present, i make perhaps half of what i need to be able to justify continuing to work on ardour to my family.
You find it weird…wow…
An absolute air of indifference from a user?
Yeah…if that attitude carried over to the developers,
yes I’d be concerned.
I expect no more or less of a commitent from one
platform to another.
Having donated/purchased software in the past and
then subsequently having that support withdrawn after
throwing in a significant amount of money to something…
I’m just a bit more careful were my money goes.
…and who said anything about one user…(other then you
that is)
speeding up or slowing down development on OSX,
or ANY other platform.
I just don’t feel like throwing money down a
bottomless well and getting nothing back.
I do that enough with commercial software.
That’s why I’m looking at Ardour as a serious
alternative to what Im using now.
OK, what I find weird is this : that my opinion about the OSX impact on linux (which turned out to be unfounded according to Paul’s statements) could be shared by or carried over, to use your words, to the developers. This is weird indeed because real life shows the contrary, that the developers do care a lot about the OSX port of ardour. My “attitude” is mine and is not correlated to the devs planning decisions, fortunately. And after Paul’s reply, I happily admit that my view on how the OSX effort could plague or burden the work on linux was not only unfounded but plain wrong. It remains though that I feel indifferent to the work on OSX, why should I care about it when I am not using OSX ? That’s a perfectly valid user’s standpoint. But once again, it has NOTHING TO DO with the developer’s opinions and should not have anything to do with it. This mix of genres is what I find weird from you. But enough about all that crap. It’s great to hear your are donating or considering it seriously. Don’t get offended by my rubbish opinion, OK ? :lol:
haszari, its a good question. The X11 version is unaffected by these issues. It hasn’t been built mostly because of time constraints on the person/people who normally takes care of it. I’ll try to push that forward a bit, because by all accounts so far, 2.5 really is a lot better than 2.4.
