[PATCH] scripts: rust_is_available: check for empty libclang version

[PATCH] scripts: rust_is_available: check for empty libclang version

[Matthew Leach]

If bindgen can't find libclang, then bindgen will panic with the
following error:

thread 'main' panicked at 'Unable to find libclang: "couldn't find any
valid shared libraries matching: ['libclang.so', 'libclang-*.so',
'libclang.so.*', 'libclang-*.so.*'], set the `LIBCLANG_PATH` environment
variable to a path where one of these files can be found.

This is outputted to stderr, leaving stdout empty. The empty string is
then passed to get_canonical_version and the following is show to the
user:

$ make LLVM=1 rustavailable
./scripts/rust_is_available.sh: line 21: 100000 *  + 100 *  + : syntax error: operand expected (error token is "+ ")
make: *** [Makefile:1825: rustavailable] Error 1

Note: bindgen's bad exit code isn't caught by 'set -e' since it is ran
in a subshell.

Fix this by prnting out a more helpful error message if the output of
bindgen is empty.

Signed-off-by: Matthew Leach <dev@mattleach.net>
---
 scripts/rust_is_available.sh | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/scripts/rust_is_available.sh b/scripts/rust_is_available.sh
index aebbf1913970..e5478429841c 100755
--- a/scripts/rust_is_available.sh
+++ b/scripts/rust_is_available.sh
@@ -107,6 +107,15 @@ bindgen_libclang_version=$( \
 		| head -n 1 \
 )
 bindgen_libclang_min_version=$($min_tool_version llvm)
+if [ -z "$bindgen_libclang_version" ]; then
+	if [ "$1" = -v ]; then
+		echo >&2 "***"
+		echo >&2 "*** libclang could not be found by bindgen.  Ensure that LIBCLANG_PATH"
+		echo >&2 "*** is set correctly."
+		echo >&2 "***"
+	fi
+	exit 1;
+fi
 bindgen_libclang_cversion=$(get_canonical_version $bindgen_libclang_version)
 bindgen_libclang_min_cversion=$(get_canonical_version $bindgen_libclang_min_version)
 if [ "$bindgen_libclang_cversion" -lt "$bindgen_libclang_min_cversion" ]; then
-- 
2.39.2

Re: [PATCH] scripts: rust_is_available: check for empty libclang version

[Martin Rodriguez Reboredo]

On 5/7/23 05:41, Matthew Leach wrote:
> If bindgen can't find libclang, then bindgen will panic with the
> following error:
> 
> thread 'main' panicked at 'Unable to find libclang: "couldn't find any
> valid shared libraries matching: ['libclang.so', 'libclang-*.so',
> 'libclang.so.*', 'libclang-*.so.*'], set the `LIBCLANG_PATH` environment
> variable to a path where one of these files can be found.
> 
> This is outputted to stderr, leaving stdout empty. The empty string is
> then passed to get_canonical_version and the following is show to the
> user:
> 
> $ make LLVM=1 rustavailable
> ./scripts/rust_is_available.sh: line 21: 100000 *  + 100 *  + : syntax error: operand expected (error token is "+ ")
> make: *** [Makefile:1825: rustavailable] Error 1
> 
> Note: bindgen's bad exit code isn't caught by 'set -e' since it is ran
> in a subshell.
> 
> Fix this by prnting out a more helpful error message if the output of
> bindgen is empty.
> 
> Signed-off-by: Matthew Leach <dev@mattleach.net>
> ---
> [...]

Reviewed-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>

Re: [PATCH] scripts: rust_is_available: check for empty libclang version

[Miguel Ojeda]

On Sun, May 7, 2023 at 10:41 AM Matthew Leach <dev@mattleach.net> wrote:
>
> If bindgen can't find libclang, then bindgen will panic with the
> following error:
>
> thread 'main' panicked at 'Unable to find libclang: "couldn't find any
> valid shared libraries matching: ['libclang.so', 'libclang-*.so',
> 'libclang.so.*', 'libclang-*.so.*'], set the `LIBCLANG_PATH` environment
> variable to a path where one of these files can be found.
>
> This is outputted to stderr, leaving stdout empty. The empty string is
> then passed to get_canonical_version and the following is show to the
> user:
>
> $ make LLVM=1 rustavailable
> ./scripts/rust_is_available.sh: line 21: 100000 *  + 100 *  + : syntax error: operand expected (error token is "+ ")
> make: *** [Makefile:1825: rustavailable] Error 1
>
> Note: bindgen's bad exit code isn't caught by 'set -e' since it is ran
> in a subshell.
>
> Fix this by prnting out a more helpful error message if the output of
> bindgen is empty.
>
> Signed-off-by: Matthew Leach <dev@mattleach.net>

Thanks a lot for this patch! This was previously reported and should
be fixed in this patch series (I added you as a reporter in v2):

    https://lore.kernel.org/rust-for-linux/20230616001631.463536-1-ojeda@kernel.org/

If you could please test the series, to make sure it solves your
issue, it would be great!

Cheers,
Miguel