I am automating the process of installation of QNX 6.3 with some beautiful scripts. I was wondering if I install, first some standard packages, for example:
freetype
libxml2
libtool
libiconv
workspace…
And some custom products.
etc.
And “after” that, the CoreOS 6.3.2A. And then reboot.
Do you think there could be some problem?
Before “what” is strictly necessary to install CoreOS 6.3.2A first?
(Maybe some newer version of libc… if not the coreos will overwrite it. But what else?)
What could happen if after the installation of coreOS 6.3.2 is complete (and the installer suggestion to reset the PC) I copy some new libc then I change some extra things in the “.build” file to customize my system, then I recompile my new boot image, then I copy the new image into /.boot… and AFTER ALL THAT, I restart the node?.. Could I have some strange problem 5 month later?
Ah, now I see what you are asking. You want to know if you are doing something in the wrong order.
With a new libc, the thing to do is check your boot image. In most cases (the default QNX boot image for example) libc.so is put into the image for faster access.
If that’s your case, you’ll have to make sure you put the libc.so you got from QNX in the right directory before you build a boot image. Otherwise even if you copy it to the drive, it won’t be used because the one in the boot image is being used first.
I actually had this happen to me a month ago. I got a new devb-eide from foundry27. I copied it to /sbin/devb-eide. I then couldn’t understand why at boot time I wasn’t using the new one. Then I realized that devb-eide was in my boot image and that one was being used, not the one in /sbin. Even worse was that I made a 2nd mistake and rebuilt my image not realizing that mkifs grabs devb-eide from another directory, not from /sbin. So that took a bit more time to track down.
So make sure to look carefully in /proc/boot to make sure the new libc is there in your new bootimage.