zsh completions for commands that have multiple Vec arguments require
special care.
We can have two Vec args separated with a value terminator.
We can also have two Vec args with no value terminators specified
where the final arg uses 'raw' and thus requires '--' to be used.
The 2nd of these scenarios requires special handling to avoid
emitting a duplicate '*:arguments' completion entry.
Currently, the zsh completions generate an error in this scenario:
$ my-app <TAB>
_arguments:...: doubled rest argument definition:
*::second -- second set of of multi-length arguments:
We already use the '-S' option when calling _arguments.
This option makes it so that completion stops after '--' is encountered.
This means that the handling for trailing 'raw' arguments does not need
to specified.
Special-case multi-valued arguments so that we can skip emitting
the final multi-valued argument if a previous multi-valued argument
has already been emitted.
Closes#3022
Signed-off-by: David Aguilar <davvid@gmail.com>
Emit the user-defined value terminator into the zsh completion pattern
to avoid doubled rest-arguments definitions:
$ my-app <TAB>
_arguments:comparguments:325: doubled rest argument definition:
*::second -- second set of of multi-length arguments:
https://github.com/clap-rs/clap/issues/3266#issuecomment-1007901407
noted that including the value terminator is one step towards a
robust solution for handling multiple multi-valued arguments.
This change does not yet handle automatically detecting when a value
terminator is needed, but it does add tests to ensure that
user-specified value terminators are used on zsh.
Related-to: #3022
Signed-off-by: David Aguilar <davvid@gmail.com>
This is an intermediate solution for #4408. As there were no agreeed
upon goals, I went with what I felt read well and that I saw commonly
used on non-clap commands.
- "information" isn't really a necessary word.
- I originally favored `Print this help` but realied that doesn't read
correctly in completions.
- Besides being shorter, the reason for the flipped short/long hint is
it gives people the context they need for scanning, emphasizing
"summary" and "more".
Fixes#4409
With the previous fixes for #4273 and #4280 in place, it's now easy to
add support for subcommand aliases, which this commit does. This
addresses #4265 for Bash.
This continues the work started with the fix for #4273. There was
another bug caused by using the subcommand names without considering
their position in the argument list. If the user enters `git diff log
<TAB>`, we build up a string that identifies the subcommand. We ended
up making the string `git__diff__log` in this case because we appended
`__log` without considering the current state. Since `git__diff__log`
does not correspond to an actual command, we wouldn't provide any
suggestions.
This commit restructures the code so we walk subcommands and
subsubcommands in `bash.rs`. While walking those, we build up a list
containing triples of the parent `$cmd` name (e.g. `git__diff`), the
current command's name (e.g. `log`), and the `$cmd` for the current
command. We then build the shell script's case arms based on that
information.
We could instead have fixed#4280 by using the second element in the
pair returned from `utils::all_subcommands()` (a stringified list of
the subcommand path) instead of the first one. However, that would not
have helped us solve #4265.
Closes#4280
Early in the Bash-completion script, we build up a string that
identifies the command or subcommand. When we see the top-level
command's name (e.g. `git`) we set the command so far to that
value. We do that regardless of where in the argument list it
appears. For example, if the argument list is `git diff git`, we set
the current command to `git` when run into it the second time. We
therefore suggest arguments to the top-level command afterwards, which
is not correct.
This patch fixes that by also considering the string that identifies
the command so far, so we only set the overall command to `git` if the
command so far is the empty string.
This is actually just a step on the way to getting completion to work
for aliases of subcommands.
Closes#4273
There seems to be little reason to return early with an empty list
when there are no subcommands, instead of going through the loop 0
times and then returning the empty list.
The derive-based example has a `///` comment on one argument, which
ends up as a description for the argument in the generated completion
scripts. Let's switch to `//` so the two scripts produce the same
output (except for the binary name), so they're easy to compare.
The example binaries were renamed in 89c2b3bb0d, but the commands in
them were not, making the generated completion scripts not work
(because we use the command name as binary name in the examples).
