[Problems with the ARM build on a WD My Book World NAS]

It was fun reading about this today, but I am no closer to a solution per se.

I am wondering if the problem stems from MBWE using uClibc instead of glibc. If this turns out true, perhaps we can get a build with static links for the shared objects missing from uClibc (our MBWE version of libc.so.6).

On the other hand, I have other software that uses iNotify (a symbol apparently missing in my libc.so.6). Have they static linked the object code? Or is there another explanation?

[Problems with the ARM build on a WD My Book World NAS]

RE the relocation error: I ran "LD_DEBUG=bindings ./btsync 2>&1" and all the bindings "seem" to match. I notice various references to the error from many different applications, suggesting to me that libc.so.6 is NOT the problem--though this is not my area of John Hodgman--err, expertise.

Or maybe I don't understand LD_DEBUG.

I tried uncommenting the "share_folders" in the config file. First run, the the threads stay alive and the sync.log reports "MY_SECRET_1" is not a valid secret. I use the command-line to generate a secret and use this in the config file.

Now, the threads die and stderr reports "./btsync: relocation error: ./btsync: symbol inotify_init, version GLIBC_2.4 not defined in file libc.so.6 with link time reference"

So, I run "LD_DEBUG=bindings ./btsync | grep inotify". Nothing. This is why I'm thinking LD_DEBUG=bindings only shows SUCCESSFULLY bound symbols....

[Problems with the ARM build on a WD My Book World NAS]

Yes, this is the same behavior for me & WD My Book World ed, v2 White Lights.

Does someone know how to unpack the webgui.zip from MBWE? I get "has more than one entry--rest ignored".

RE the relocation error: I ran "LD_DEBUG=bindings ./btsync 2>&1" and all the bindings "seem" to match. I notice various references to the error from many different applications, suggesting to me that libc.so.6 is NOT the problem--though this is not my area of John Hodgman--err, expertise.

Very disappointing debug output.

Are these empty *.dmp files of any value in working this problem?

Is the problem simply that MBWE uses BusyBox instead of a "proper" linux??