Re: [PATCH 1/3] myfirstcontrib: add 'psuh' to command-list.txt

> On Fri, Oct 25, 2019 at 06:08:57PM -0700, Jonathan Nieder wrote:
> > > Users can discover commands and their brief usage by running 'git help
> > > git' or 'git help -a'; both of these pages list all available commands
> > > based on the contents of 'command-list.txt'. That means adding a new
> > > command there is an important part of the new command process, and
> > > therefore belongs in the new command tutorial.
> >
> > Makes sense.
> >
> > Not about this patch: is there a way to detect this automatically?
> > E.g. if a command in git.c::commands doesn't appear in
> > command-list.txt, could we make Git fail "make test"?
>
> We almost detect this already:
>
>   $ sed -i -e '/^git-bisect/d' command-list.txt
>   $ make check-docs
>   make -C Documentation lint-docs
>   make[1]: Entering directory '/home/szeder/src/git/Documentation'
>       GEN cmd-list.made
>       GEN doc.dep
>   make[2]: Entering directory '/home/szeder/src/git'
>   make[2]: 'GIT-VERSION-FILE' is up to date.
>   make[2]: Leaving directory '/home/szeder/src/git'
>   make[2]: Entering directory '/home/szeder/src/git'
>   make[2]: 'GIT-VERSION-FILE' is up to date.
>   make[2]: Leaving directory '/home/szeder/src/git'
>       LINT lint-docs
>   make[1]: Leaving directory '/home/szeder/src/git/Documentation'
>   no link: git-bisect
>   $ echo $?
>   0
>
> See that last "no link: git-bisect" line?  That's what happened to
> catch my eyes when Derrick forgot to add his new 'sparse-checkout'
> builtin to 'command-list.txt'.  I still haven't looked up what that
> 'no link' is supposed to mean, but if it were an error, then we would
> have surely detected the missing entry in 'command-list.txt' in out CI
> builds.

>
> Another possibility would be to auto-generate that long list of
> 'cmd_foo()' function declaration in 'builtin.h' from
> 'command-list.txt', by adding a 'builtin' attribute to mark builtin
> commands.
>
>