The case for reviving Firefly

[Anonymous]

as this all seems to have ground to a halt, i’d like to suggest that we give this one final push.

i’m not a programmer by any means, but i’d like to be able to help anyway i can. ( hosting, website setup, GIT set-up, forum etc etc)

i guess the key points are;
code that will run on windows, linux and maybe osx.
a structured team where we don’t have to rely on any one person.
a real push to get the code working again.

so as a show of hands, who would be interested in doing this?

[mas]

I would doubt that the fork will run on a NSLU. The guy made the code really for himself and the portability is not (yet?) really high.

For instance I tried to compile the mt-daapd fork. First I had to install antlr and a few other dependencies. Some work but in the end straightforward to do. And as antlr seems to be a java tool there already it ends for the NSLUs as well I am afraid.

With my miniPC based on a VIA CPU and gentoo distribution I did not come much further though. After installing all dependencies I get now in the configure run

checking avl.h usability… no
checking avl.h presence… no
checking for avl.h… no
configure: error: avl.h not found

Doing a google search does not really reveal the package that belongs to. And a configure run should not bark out just with a missing header file error. There seems to be something either bugged or highly distribution dependant.

[mas]

And with that beeing said we have a problem. We do not have the manpower in term of coders to do a rewrite on multiple platforms. So we need to focus and throw some things out. After all the old version remains to stay and maybe later we find the manpower to support them as well.

Thus starting with forked-mt-daapd would be the logical thing. But then we need to get this code at least robust enough that it works on more than one linux distribution and is not a nightmare in dependencies.

I have no idea if the author of the fork would be willing to help a bit. I for myself am no coder but test it on my platform I can and will. Right now it is just not compilable on a gentoo/VIA system for unknown reasons.

[Anonymous]

Face it guys this is going nowhere.

I’m think of upgrading my NSLU2 with the new Iomega Iconnect. Trouble is it doesn’t support FLAC out of the box.

[Anonymous]

where do we go from here?
Andrew’s attempt at getting this up and running has ground to a halt. partly due to him not being contactable (no fault of his own i’m sure)

i am more than willing to set another website up, forum, GIT repositaries etc.

can we have a show of hands to see who would be interested.

Peter

[thorstenhirsch]

@Ace/Julien: are you going to release binary packages of forked-daap?

[w1ll14m]

@mas wrote:

I would doubt that the fork will run on a NSLU. The guy made the code really for himself and the portability is not (yet?) really high.

For instance I tried to compile the mt-daapd fork. First I had to install antlr and a few other dependencies. Some work but in the end straightforward to do. And as antlr seems to be a java tool there already it ends for the NSLUs as well I am afraid.

With my miniPC based on a VIA CPU and gentoo distribution I did not come much further though. After installing all dependencies I get now in the configure run

checking avl.h usability… no
checking avl.h presence… no
checking for avl.h… no
configure: error: avl.h not found

Doing a google search does not really reveal the package that belongs to. And a configure run should not bark out just with a missing header file error. There seems to be something either bugged or highly distribution dependant.

I had this one too, not sure how i fixed it though.
i got it to run, but it lacked some features.

I want the scanner to index files it does not support.
For custom backend goodness 😉
Had some strange crashes, so i reverted to mt-daapd

— Edit —
You need to install avl-0.3.5 (sometimes called libavl)

Although it uses antlr most of it is for automating compilation and some sort of template based programming (correct me if i’m wrong).
I think cross compiled binaries could work.

[w1ll14m]

–configure output–
checking for evhttp.h… yes
checking avl.h usability… yes
checking avl.h presence… yes
checking for avl.h… yes
checking for avl_alloc_tree in -lavl… yes
checking antlr3.h usability… yes
checking antlr3.h presence… yes

---

ldd output:

---

Asuka sbin # ldd forked-daapd
linux-gate.so.1 => (0xffffe000)
libavahi-client.so.3 => /usr/lib/libavahi-client.so.3 (0xb7fb4000)
libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0xb7f7c000)
libavahi-common.so.3 => /usr/lib/libavahi-common.so.3 (0xb7f6f000)
libdl.so.2 => /lib/libdl.so.2 (0xb7f6b000)
libsqlite3.so.0 => /usr/lib/libsqlite3.so.0 (0xb7f08000)
libavcodec.so.52 => /usr/lib/libavcodec.so.52 (0xb75b7000)
libavformat.so.52 => /usr/lib/libavformat.so.52 (0xb74fb000)
libswscale.so.0 => /usr/lib/libswscale.so.0 (0xb74ba000)
libavutil.so.50 => /usr/lib/libavutil.so.50 (0xb74a7000)
libconfuse.so.0 => /usr/lib/libconfuse.so.0 (0xb749b000)
libevent_core-1.4.so.2 => /usr/lib/libevent_core-1.4.so.2 (0xb7491000)
libnsl.so.1 => /lib/libnsl.so.1 (0xb7479000)
librt.so.1 => /lib/librt.so.1 (0xb7470000)
libresolv.so.2 => /lib/libresolv.so.2 (0xb745c000)
libavl.so.1 => /usr/lib/libavl.so.1 (0xb7459000)
libmxml.so.1 => /usr/lib/libmxml.so.1 (0xb744e000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7437000)
libantlr3c.so => /usr/lib/libantlr3c.so (0xb7424000)
libgcrypt.so.11 => /usr/lib/libgcrypt.so.11 (0xb73a6000)
libgpg-error.so.0 => /usr/lib/libgpg-error.so.0 (0xb73a1000)
libc.so.6 => /lib/libc.so.6 (0xb7264000)
/lib/ld-linux.so.2 (0xb7fd3000)
libz.so.1 => /lib/libz.so.1 (0xb7251000)
libm.so.6 => /lib/libm.so.6 (0xb722a000)
libdirac_encoder.so.0 => /usr/lib/libdirac_encoder.so.0 (0xb7195000)
libdirac_decoder.so.0 => /usr/lib/libdirac_decoder.so.0 (0xb7126000)
libfaac.so.0 => /usr/lib/libfaac.so.0 (0xb7114000)
libfaad.so.2 => /usr/lib/libfaad.so.2 (0xb70d6000)
libmp3lame.so.0 => /usr/lib/libmp3lame.so.0 (0xb705e000)
libopenjpeg.so.2 => /usr/lib/libopenjpeg.so.2 (0xb7041000)
libspeex.so.1 => /usr/lib/libspeex.so.1 (0xb702a000)
libtheora.so.0 => /usr/lib/libtheora.so.0 (0xb6ff9000)
libogg.so.0 => /usr/lib/libogg.so.0 (0xb6ff3000)
libvorbisenc.so.2 => /usr/lib/libvorbisenc.so.2 (0xb6ef7000)
libvorbis.so.0 => /usr/lib/libvorbis.so.0 (0xb6ec4000)
libx264.so.78 => /usr/lib/libx264.so.78 (0xb6e29000)
libxvidcore.so.4 => /usr/lib/libxvidcore.so.4 (0xb6d14000)
libbz2.so.1 => /lib/libbz2.so.1 (0xb6d03000)
libstdc++.so.6 => /usr/lib/gcc/i486-pc-linux-gnu/4.4.2/libstdc++.so.6 (0xb6c10000)
libgcc_s.so.1 => /usr/lib/gcc/i486-pc-linux-gnu/4.4.2/libgcc_s.so.1 (0xb6bf1000)

---

Not sure what is needed on a nslu like device?
If intrested, i can upload these binaries somewhere..

hehehe i have those 🙂

Running gentoo and as stated above you need to install avl/libavl (search for the correct version as said above) or else you get something totaly different.
If you can’t find it i can upload the source somewhere.

[Anonymous]

Do all the features work in your case william?

After compiling does one still need java?

Is glibc 2.9 mandatory?

[mas]

Ok, after about 2h of fixing the dependency issues I got it and what an amazingly long lib list:

-rwxr-xr-x 1 root root 1175053 16. Mar 16:05 forked-daapd
sprinter sbin # ldd -r forked-daapd
linux-gate.so.1 => (0xb7783000)
libavahi-client.so.3 => /usr/lib/libavahi-client.so.3 (0xb7767000)
libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0xb772d000)
libavahi-common.so.3 => /usr/lib/libavahi-common.so.3 (0xb7720000)
libdl.so.2 => /lib/libdl.so.2 (0xb771c000)
libsqlite3.so.0 => /usr/lib/libsqlite3.so.0 (0xb7653000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7639000)
libavcodec.so.52 => /usr/lib/libavcodec.so.52 (0xb6d7c000)
libavformat.so.52 => /usr/lib/libavformat.so.52 (0xb6cd0000)
libswscale.so.0 => /usr/lib/libswscale.so.0 (0xb6ca2000)
libavutil.so.50 => /usr/lib/libavutil.so.50 (0xb6c8f000)
libconfuse.so.0 => /usr/lib/libconfuse.so.0 (0xb6c83000)
libFLAC.so.8 => /usr/lib/libFLAC.so.8 (0xb6c49000)
libogg.so.0 => /usr/lib/libogg.so.0 (0xb6c42000)
libm.so.6 => /lib/libm.so.6 (0xb6c1c000)
libtag_c.so.0 => /usr/lib/libtag_c.so.0 (0xb6c17000)
libtag.so.1 => /usr/lib/libtag.so.1 (0xb6b7f000)
libevent_core-1.4.so.2 => /usr/lib/libevent_core-1.4.so.2 (0xb6b75000)
libnsl.so.1 => /lib/libnsl.so.1 (0xb6b5d000)
librt.so.1 => /lib/librt.so.1 (0xb6b54000)
libresolv.so.2 => /lib/libresolv.so.2 (0xb6b3f000)
libavl.so.1 => /usr/local/lib/libavl.so.1 (0xb6b3c000)
libantlr3c.so => /usr/lib/libantlr3c.so (0xb6b26000)
libgcrypt.so.11 => /usr/lib/libgcrypt.so.11 (0xb6ab0000)
libgpg-error.so.0 => /usr/lib/libgpg-error.so.0 (0xb6aaa000)
libc.so.6 => /lib/libc.so.6 (0xb6960000)
/lib/ld-linux.so.2 (0xb7784000)
libz.so.1 => /lib/libz.so.1 (0xb694c000)
libmp3lame.so.0 => /usr/lib/libmp3lame.so.0 (0xb68d7000)
libopenjpeg.so.2 => /usr/lib/libopenjpeg.so.2 (0xb68b9000)
libspeex.so.1 => /usr/lib/libspeex.so.1 (0xb68a1000)
libvorbisenc.so.2 => /usr/lib/libvorbisenc.so.2 (0xb67a7000)
libvorbis.so.0 => /usr/lib/libvorbis.so.0 (0xb677d000)
libx264.so.78 => /usr/lib/libx264.so.78 (0xb6707000)
libxvidcore.so.4 => /usr/lib/libxvidcore.so.4 (0xb661c000)
libbz2.so.1 => /lib/libbz2.so.1 (0xb6609000)
libstdc++.so.6 => /usr/lib/gcc/i686-pc-linux-gnu/4.3.4/libstdc++.so.6 (0xb651d000)
libgcc_s.so.1 => /usr/lib/gcc/i686-pc-linux-gnu/4.3.4/libgcc_s.so.1 (0xb650f000)

---

Was quite a nightmare to get this to compile on gentoo-linux. Except installing the antlr3 packages it was required to

1. Manually download the avl-0.3.5 source, compile it and install it manually as the makefile of this old package seems to be not compatible with a modern Make version really. Sorry, such old, outdated stuff should not be used any more. Theres likely a reason that no gentoo ebuilt exists for this. I bet it will give trouble in other distributions as well.

2. Manually downloading and installing the runtime/C stuff and headers as the existing antl3 package from gentoo didn’t seem to include the headers from the C runtime. Also nasty but I am blaming gentoo here. Seems theit antlr3 isn’t a real developer package.

I haven’t yet found time to really test the forked-mtdaapd executable, but in order to use it as a substitute or starting point for new development I think that it would be necessary to rework the code from forked-daapd to not use antlr and avl as I foresee otherwise problems also with other distributions.
And we would likely want something that at least runs on the Top10 linux distributions without too many troubles of portability.
It would also not be a bad thing to remove eventually some of the dependencies, as the list is mighty long.

Just compare to the old mt-daapd version, which is also half the size of the binary:

sprinter sbin # ldd -r /usr/sbin/mt-daapd
linux-gate.so.1 => (0xb7868000)
libavahi-common.so.3 => /usr/lib/libavahi-common.so.3 (0xb7850000)
libavahi-client.so.3 => /usr/lib/libavahi-client.so.3 (0xb783f000)
libdl.so.2 => /lib/libdl.so.2 (0xb783b000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7822000)
libid3tag.so.0 => /usr/lib/libid3tag.so.0 (0xb7810000)
libz.so.1 => /lib/libz.so.1 (0xb77fc000)
libsqlite3.so.0 => /usr/lib/libsqlite3.so.0 (0xb7732000)
libogg.so.0 => /usr/lib/libogg.so.0 (0xb772b000)
libvorbis.so.0 => /usr/lib/libvorbis.so.0 (0xb7701000)
libvorbisfile.so.3 => /usr/lib/libvorbisfile.so.3 (0xb76f8000)
libFLAC.so.8 => /usr/lib/libFLAC.so.8 (0xb76bf000)
libc.so.6 => /lib/libc.so.6 (0xb7575000)
libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0xb753a000)
/lib/ld-linux.so.2 (0xb7869000)
libm.so.6 => /lib/libm.so.6 (0xb7514000)

So in short, why is this fork – after stripping some functions out – using soo many libraries?

[Author]

Posts