The new type CompletionFunc could lead to a regression.
This commit make the new `CompletionFunc` type a type alias instead,
in case projects using Cobra have created their own similar type.
This commit also adds a test to ensure that the completion function
remains backwards-compatible.
Signed-off-by: ccoVeille <3875889+ccoVeille@users.noreply.github.com>
When a program has no sub-commands, its root command can accept
arguments. If we add the default "completion" command to such programs
they will now have a sub-command and will no longer accept arguments.
What we do instead for this special case, is only add the "completion"
command if it is being called, or if it is being completed itself.
We want to have the "completion" command for such programs because it
will allow the completion of flags and of arguments (if provided
by the program).
Signed-off-by: Marc Khouzam <marc.khouzam@gmail.com>
The code has also been refactored to use a type alias for completion and a completion helper
Using a type alias is a non-breaking change and it makes the code more readable and easier to understand.
Signed-off-by: ccoVeille <3875889+ccoVeille@users.noreply.github.com>
Co-authored-by: Marc Khouzam <marc.khouzam@gmail.com>
Follow-up to #1956.
This commit allows a program to reset any of the tree templates to their
default behaviour, as it was possible to do before the change of #1956.
Signed-off-by: Marc Khouzam <marc.khouzam@gmail.com>
In the bash shell we used to print ActiveHelp messages on every
tab-press. In the example below, notice the "Command help" line which is
ActiveHelp:
bash-5.1$ tanzu context u[tab]
Command help: Configure and manage contexts for the Tanzu CLI
bash-5.1$ tanzu context u[tab]
Command help: Configure and manage contexts for the Tanzu CLI
bash-5.1$ tanzu context u
unset (Unset the active context so that it is not used by default.)
use (Set the context to be used by default)
bash-5.1$ tanzu context u
Above, on the first [tab] press, only the ActiveHelp is printed.
On the second [tab] press, the ActiveHelp is printed again, followed
by a re-print of the command-line, followed by the completions choices.
The separation between ActiveHelp and completion choices makes the
ActiveHelp harder to see. Furthermore, I find the double printing of the
ActiveHelp string to look bad.
Note that for zsh, the UX is different and that ActiveHelp messages are
printed at the same time as the completion choices.
This commit aligns the UX for ActiveHelp in bash with the one for zsh:
if there are other completions to be shown, the ActiveHelp messages are
printed at the same time.
New behaviour:
1- ActiveHelp is no longer printed on the first [tab] press. This is
better aligned with bash's standard approach.
2- ActiveHelp is printed on the second [tab] press, above the completion
choices, with a `--` delimiter.
3- If there are no completion choices, the `--` delimiter is omitted.
This behaviour is the same as what is done for zsh (except that for zsh
the first [tab] press immediately shows completion choices).
Below is the above example, but using this commit.
Notice the more concise and easier to read completion output:
bash-5.1$ tanzu context u[tab][tab]
Command help: Configure and manage contexts for the Tanzu CLI
--
unset (Unset the active context so that it is not used by default.)
use (Set the context to be used by default)
bash-5.1$ tanzu context u
Signed-off-by: Marc Khouzam <marc.khouzam@gmail.com>
* Restructure code to let linker perform deadcode elimination step
Cobra, in its default configuration, will execute a template to generate
help, usage and version outputs. Text/template execution calls MethodByName
and MethodByName disables dead code elimination in the Go linker, therefore
all programs that make use of cobra will be linked with dead code
elimination disabled, even if they end up replacing the default usage, help
and version formatters with a custom function and no actual text/template
evaluations are ever made at runtime.
Dead code elimination in the linker helps reduce disk space and memory
utilization of programs. For example, for the simple example program used by
TestDeadcodeElimination 40% of the final executable size is dead code. For a
more realistic example, 12% of the size of Delve's executable is deadcode.
This PR changes Cobra so that, in its default configuration, it does not
automatically inhibit deadcode elimination by:
1. changing Cobra's default behavior to emit output for usage and help using
simple Go functions instead of template execution
2. quarantining all calls to template execution into SetUsageTemplate,
SetHelpTemplate and SetVersionTemplate so that the linker can statically
determine if they are reachable
Co-authored-by: Marc Khouzam <marc.khouzam@gmail.com>
The completion code attempts to detect whether a flag can be specified
more than once, and therefore should provide completion even if already
set.
Currently, this code depends on conventions used in the pflag package,
which uses an "Array" or "Slice" suffix or for some types a "stringTo"
prefix.
Cobra allows custom value types to be used, which may not use the same
convention for naming, and therefore currently aren't detected to allow
multiple values.
The pflag module defines a [SliceValue] interface, which is implemented
by the Slice and Array value types it provides (unfortunately, it's not
currently implemented by the "stringTo" values).
This patch adds a reduced interface based on the [SliceValue] interface
mentioned above to allow detecting Value-types that accept multiple values.
Custom types can implement this interface to make completion work for
those values.
I deliberately used a reduced interface to keep the requirements for this
detection as low as possible, without enforcing the other methods defined
in the interface (Append, Replace) which may not apply to all custom types.
Future improvements can likely still be made, considering either implementing
the SliceValue interface for the "stringTo" values or defining a separate
"MapValue" interface for those types.
Possibly providing the reduced interface as part of the pflag module and
to export it.
[SliceValue]: d5e0c0615a/flag.go (L193-L203)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Signed-off-by: Marc Khouzam <marc.khouzam@gmail.com>
Co-authored-by: Marc Khouzam <marc.khouzam@gmail.com>
When running tests in verbose mode (or other options), tests involving
Cobra may fail if the test does not explicitly set Command.args to an
empty slice; in this case, Cobra defaults to using `os.Args`, which
will contain arguments passed to the test (such as `-v` (verbose)).
Commits e5762054c9 and 1ef0913976
implemented a workaround for this when running (unit) tests for Cobra
itself, but this check is specifig to Cobra (checking for `cobra.test`),
and don't work on Windows (which will have a `.exe` extension),
This patch implements a more universal check, so that users of Cobra
as a module also benefit from this workaround.
go1.21 and up provides a `testing.Testing()` utility ([1]); as the Cobra
module still supports Go1.16 and up, an alternative implementation was
added for older versions, based on golang.org/x/mod/lazyregexp [2].
Before this patch:
go test -c -o foo.test
./foo.test -test.run TestNoArgs
--- FAIL: TestNoArgs (0.00s)
args_test.go:37: Unexpected output: Error: unknown command "TestNoArgs" for "c"
Usage:
c [flags]
Flags:
-h, --help help for c
args_test.go:40: Unexpected error: unknown command "TestNoArgs" for "c"
FAIL
After this patch:
go test -c -o foo.test
./foo.test -test.run TestNoArgs
PASS
[1]: https://pkg.go.dev/testing#Testing
[2]: https://cs.opensource.google/go/x/mod/+/refs/tags/v0.19.0:internal/lazyregexp/lazyre.go;l=66-78
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Signed-off-by: Marc Khouzam <marc.khouzam@gmail.com>
Co-authored-by: Marc Khouzam <marc.khouzam@gmail.com>
Since cpuguy83/go-md2man 2.0.5 no paraTag is written after "SEE ALSO".
With go-md2man 2.0.4:
.SH SEE ALSO
.PP
\fBroot-bbb(1)\fP, \fBroot-ccc(1)\fP
With go-md2man 2.0.5:
.SH SEE ALSO
\fBroot-bbb(1)\fP, \fBroot-ccc(1)\fP
See: https://github.com/cpuguy83/go-md2man/pull/122
Signed-off-by: Mikel Olasagasti Uranga <mikel@olasagasti.info>
