On Thu, Jun 19, 2025 at 06:38:43PM +0200, Marc Chantreux wrote: > hello Eric, > > On Thu, Jun 19, 2025 at 10:43:39AM -0500, Eric Blake wrote: > > The Austin Group has been asked whether future POSIX should include > > the sponge(1) utility > > thanks for sharing this very good news. I didn't know about the Austin > Group either so you made my day.
Sorry for not answering sooner, > > do you know a way to file a request to them? I really would like to > suggest the behavior of most of the shells to be adopted as a POSIX > behavior: no need to write a counpound expression expression to declare > a function so > > ll() ls -l "$@" > > would be the same as > > ll() { ls -l "$@" } > > so maybe people would stop using alias to to such things and then > complain when they discover they don't understand what alias is like in > > assert() { > echo "I will $*" > "$@" > } > > which fails with aliases Proposals for changes to POSIX can be submitted at https://www.austingroupbugs.net/. For this particular request, it may be worth checking the existing POSIX grammar at https://pubs.opengroup.org/onlinepubs/9799919799/utilities/V3_chap02.html#tag_19_10 which shows compound_command : brace_group | subshell | for_clause | case_clause | if_clause | while_clause | until_clause ; ... function_body : compound_command /* Apply rule 9 */ | compound_command redirect_list /* Apply rule 9 */ ; which does not include your desired syntax, so your bug report should include what the updated grammar should look like. -- Eric Blake, Principal Software Engineer Red Hat, Inc. Virtualization: qemu.org | libguestfs.org