Seems like one tiredness wasn’t enough. Today I resumed labelling envelopes, and tomorrow I’ll be printing receipts, and the day after will be closing the envelops and afterward attaching the receipt to them. It’s a pita of a work, but I’m paid for it (not that well actually, but better than nothing).
Of course there’s also the boring stuff with XMMS. Of course the news that it will go way from Gentoo finally was misrepresented and misunderstood as me forcing users with all the rights on this world to use Audacious just for the sake of it. I know it’s only a vocal minority that is currently complaining, but I want to make it clear.
I was an XMMS user, and it took me quite a while before jumping on the Beep Media Player bandwagon, after that I used Audacious for a while and then moved to Amarok for good.. this latter move was also due to my move from a random listener to a more assiduous one (as I’ve started ripping my CDs and playing them from harddisk). This means I don’t have a personal grudge against XMMS (I used to use it too), and I don’t have anything against GTK itself this time (well, I don’t like GTK, but it’s not like I’m forcing everybody to use Qt… I think everybody has his taste). I’m not an Audacious guy either, although I do like it for random listening, and I’m glad to help from time to time (warning fixes, porting on FreeBSD, translations, …), I think that they are doing a good deal of work, so thanks a lot to Chainsaw, nenolod and the rest of the developers.
Why I was happy of getting rid of XMMS then? Because I think we should have phased it out last year already.. there were way more bugs open, okay, but I think Luis invested too much time on such a dead project, and although I’m not sure of this, I’d be ready to bet that this contributed on his stress. As Tony said, this is the XMMS curse, every one of the developers maintaining XMMS in the past ended up exhausted. This is quite easy to understand, although from one site the wide availability of XMMS plugins means that you can play almost everything and to almost every output system, and you can do quite interesting things with the misc plugins, the combinations are so many that you cannot test for them. Also, maintaining a project dead upstream is basically suicidal, as you end up maintaining every part of the code, not just the packaging and building.
This is not the first time XMMS faces this destiny. Last year I already announced the phase out plan, but then Luis took it over and so it was saved from the coming doom. Unfortunately as you can see, Luis only postponed the final cvs rm, as now we’re at the same point. This is why I wasn’tgoing to write once again a maintainer requrest on gentoo-dev and waiting to hear before going with the mask. Nobody in the sound herd is going to take care of it, and any other person wanting to do that should first be able to fix all the bugs for good and then make sure he has at least one fall back when he’s going to retire… as I doubt anyone is going to do that, XMMS is going away for good.
As I said, I won’t be putting it in an official overlay, because having an official overlay with unsupported software is just silly, a bad move in PR and a bad move in QA. I won’t stop anybody going to create an XMMS overlay, but be assured that we won’t support it in any way, and the way it currently works (faad and flac using the xmms useflag to build the plugin), if the libraries are overlaid, we’re not going to support you with them either.
After all, I think there are waaaay more users that will be happy with the reduced sync time than there will be pissed off by the removal of such an old piece of broken code like XMMS. If it worked for you it does not mean it works for everyone!
Uhm, I started this entry thinking of writing something about my plans for Gentoo/FreeBSD, but then I talked about XMMS up to now, so I’ll try to be short now. First of all, adding enterprise as buildbox for Prakesh helped, at least a bit, kdelibs took just 2 hours instead of 3, it’s still a lot, but at least a bit better than before. As farragut is still on 6.1, I need to find a way to be able to install a cross-compiler to 6.2 on a 6.1 system… right now it does not work that well, but I’m going to improve the situation soon. I’m also thinking of using an old 533 Pentium3 as an extra distcc box, although it won’t help that much, I’m sure.
Also, I now realise that the only way to make sure that Gentoo/FreeBSD can be considered a first class citizen of the Gentoo project, is to provide access for the developers.. a simple way to do this would be having a box online, with a lot of space on disk, a jail for generic usage for developers and one each developer if they are going to do something more complex projects. The problem is that right now Gentoo/FreeBSD can be pretty tricky and needing a manual tinkering, which means someone who understand G/FBSD should have direct access to the box.. I do have a friend who could house the box for me free of charge, but still remains the problem of the box itself.. needs to be at least a dual processor I suppose, and with big disks, being able to leave at least the core system on mirroring… which means it will take many months for me to be able to even think of buying it… I don’t even have the money to buy a new box for me ;)
But anyway the idea is not totally discarded yet.. I’d like to hear if someone can provide me with some estimate of how much I could end up spending, it might make my mind clearer.. after that maybe I might end up asking if anybody wants to contribute… but as I said, I feel that would be really only the last choice I have, unless people feel really motivated to.