Package: ucf Version: all
Hello, I see the sanity check for the conffold/conffnew setting as a bug. Consider the following situation: Setting in /etc/ucf.conf: conf_force_conffold=YES In the postinst script set: export UCF_FORCE_CONFFNEW=YES This will always lead to an abortion of the script with the message: Error: Only one of force_conffold and force_conffnew should be set Reason: It is not possible to overwrite the ucf.conf setting conf_force_conffold or conf_force_conffnew with an empty value. Doing export UCF_FORCE_CONFFOLD="" would not prevent the error, because if it is empty/unset, the value from ucf.conf is used. According to the manual and what I see in the code, the intention is: The priority of the command line is highest, second is the environment and last the settings in ucf.conf. Effects: The only package I saw (besides my own config packages) using the environment is "grub-pc" Many thanks for reading this and thinking about it. I really appreciate the work of Debian and the Debian package maintainers! Greetings ... Marco -- UZH/ETH Zurich, Marco Schmidt, Functional Genomics Center Zurich Y32 H06, Winterthurerstr. 190, 8057 Zurich Tel +41 44 635 39 02, Fax +41 44 635 39 22, https://www.fgcz.ch/ - IT Coordinator -
