Hi, should we really allow for the usage of alias in maintainer scripts?
Policy 10.4 says that /bin/sh scripts may assume an interpreter that implements the POSIX standard, plus some additional features listed in the policy. And the POSIX standards describes alias substitution in shell scripts [1], so it seems we currently do allow for it. However, /bin/sh may be linked to bash, and bash does alias substitution only when in interactive mode but not in batch mode [2]. That is, unless a special bash option is set. Hence, maintainer scripts relying on alias substitution will not work correctly when /bin/sh is bash. I think we should - state in the policy that /bin/sh maintainer scripts must not use the alias buildin. We might suggest to use a shell function instead. - file bug reports against packages that use alias in their maintainer scripts. Luckily, there currently seem to be only two such packages, namely logdata-anomaly-miner and exim4-config. Opinions? -Ralf. [1] http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_03_01 [2] https://www.gnu.org/software/bash/manual/html_node/Aliases.html
