Re: [yocto] how to create a named_pipe in my image

2016-02-08 Thread Baumann, Michael
Thank you Ross. Switching to rpm package showed the same behavior. Is there another workaround than a post install script? Regards, Michael From: Burton, Ross [mailto:ross.bur...@intel.com] Sent: Monday, February 08, 2016 5:15 PM To: Baumann, Michael Cc: yocto@yoctoproject.org Subject: Re:

Re: [yocto] how to create a named_pipe in my image

2016-02-08 Thread Burton, Ross
On 8 February 2016 at 16:12, Baumann, Michael wrote: > Possibly it's because named pipes cannot be packed into an archive? If > yes, how can I circumvent this? > It would be interesting to see what happens if you change PACKAGE_CLASSES to use rpm instead of ipkg.

Re: [yocto] how to create a named_pipe in my image

2016-02-08 Thread Baumann, Michael
To make my question more precise: mkfifo in the do_install() section works. I can see the named pipe in build/tmp-glibc/work/*/*/*/ but my bitbake hangs in the do_package_write_ipk and do_package_qa tasks. Possibly it's because named pipes cannot be packed into an archive? If yes, how can I

[yocto] how to create a named_pipe in my image

2016-02-08 Thread Baumann, Michael
Hi, I try to create a named pipe (fifo) for my target filesystem. I tried it in my .bb file in the do_install() section with "mkfifo " Any ideas gratefully received. Regards Michael -- ___ yocto mailing list yocto@yoctoproject.org

Re: [yocto] how to create a named_pipe in my image

2016-02-08 Thread Burton, Ross
On 8 February 2016 at 16:39, Baumann, Michael wrote: > Switching to rpm package showed the same behavior. > > Is there another workaround than a post install script? > > So I had a quick look at this. The fifo certainly ends up in the resulting ipk at least, but as

Re: [yocto] how to create a named_pipe in my image

2016-02-08 Thread Burton, Ross
On 8 February 2016 at 17:36, Burton, Ross wrote: > So I had a quick look at this. The fifo certainly ends up in the > resulting ipk at least, but as you say package_qa hangs. It's probably > encountering a non-file and breaking somewhere. A really nasty workaround >