hi,
(cc'ing the address that closed this bug to make sure they get this
feedback)
On Fri, 2007-03-16 at 19:16 +0000, VMware Build Team wrote:
> Why should the player setup succeed when in reality something is broken
> and the software won't work. At least if the installation fails the
> user will see there is an error. Otherwise they'll believe everything
> is great and they'll try to power on a VM and things will fail.
in this case there was no user to see it, it was an automated
installation environment. the automated installation environment (FAI)
uses a custom kernel which is not the same as the kernel used after the
system boots. the custom kernel does not have vmware modules available,
and it shouldn't need them.
> The init script is not the most robust piece of code ever written,
> that's for sure. But its problems should be fixed instead of blindly
> ignoring them. Not having the right kernel modules installed is not a
> valid reason, these should be installed automatically on behalf of the
> user.
instead of ignoring the bug report for 5 months only to set it to
rejected, you could have asked if i have a suggestion for how to work
around this problem while still meeting your desired behaviour.
for example, you could have look for an environment variable
VMWARE_PLEASE_IGNORE_SERVICE_START_FAILURES or similar, and if
it's non-empty ignore the failure status of the init script.
personally, i think that a failure of a system service to start should
not ever, ever break the package management system, which is what
currently happens in this situation. it's much easier for a user to
discover "oh, i need to apt-get install this-kernel and these-modules"
on a working system, as opposed to having a system where the package
management system is in an inconsistant state.
but in any event, whatever.
sean'
--
vmware-player uninstallable running default kernel
https://launchpad.net/bugs/69509
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
