[PATCH] test/T360-symbol-hiding: use readelf in place of nm

[PATCH] test/T360-symbol-hiding: use readelf in place of nm

[David Bremner]

It turns out that using nm -P isn't as portable as hoped. In particular on
architectures using ELF v1 (e.g. ppc64), the desired symbols end up in
the data section instead of text.

The test is currently only functional on ELF based architectures, so I
think it's legit to depend on readelf instead of nm.

The switch to readelf has the advantage that we can explicitely ask
for all of the symbols with global visibility, rather than grepping
for notmuch. That seems a more robust approach since it will catch any
strangely named global symbols.
---
 test/T360-symbol-hiding.sh | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/test/T360-symbol-hiding.sh b/test/T360-symbol-hiding.sh
index b34f1e54..4e5007da 100755
--- a/test/T360-symbol-hiding.sh
+++ b/test/T360-symbol-hiding.sh
@@ -26,7 +26,8 @@ test_begin_subtest 'checking output'
 test_expect_equal "$result" "$output"
 
 test_begin_subtest 'comparing existing to exported symbols'
-nm -P $NOTMUCH_BUILDDIR/lib/libnotmuch.so | awk '$2 == "T" && $1 ~ "^notmuch" {print $1}' | sort | uniq > ACTUAL
+readelf -Ws $NOTMUCH_BUILDDIR/lib/libnotmuch.so | \
+    awk '$4  == "FUNC" && $5 == "GLOBAL" && $7 != "UND" {print $8}' | sort | uniq > ACTUAL
 sed -n 's/^\(notmuch_[a-zA-Z0-9_]*\)[[:blank:]]*(.*/\1/p' $NOTMUCH_SRCDIR/lib/notmuch.h | sort | uniq > EXPORTED
 test_expect_equal_file EXPORTED ACTUAL
 
-- 
2.29.2

Re: [PATCH] test/T360-symbol-hiding: use readelf in place of nm

[Tomi Ollila]

On Fri, Dec 11 2020, David Bremner wrote:

> It turns out that using nm -P isn't as portable as hoped. In particular on
> architectures using ELF v1 (e.g. ppc64), the desired symbols end up in
> the data section instead of text.
>
> The test is currently only functional on ELF based architectures, so I
> think it's legit to depend on readelf instead of nm.
>
> The switch to readelf has the advantage that we can explicitely ask
> for all of the symbols with global visibility, rather than grepping
> for notmuch. That seems a more robust approach since it will catch any
> strangely named global symbols.

Looks good. I use this opportunity to mention that `sort | uniq` can be
replaced with `sort -u`  :D

(also) s/$4  == "FUNC"/$4 == "FUNC"/; i.e. remove one extra space

Tomi

> ---
>  test/T360-symbol-hiding.sh | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/test/T360-symbol-hiding.sh b/test/T360-symbol-hiding.sh
> index b34f1e54..4e5007da 100755
> --- a/test/T360-symbol-hiding.sh
> +++ b/test/T360-symbol-hiding.sh
> @@ -26,7 +26,8 @@ test_begin_subtest 'checking output'
>  test_expect_equal "$result" "$output"
>  
>  test_begin_subtest 'comparing existing to exported symbols'
> -nm -P $NOTMUCH_BUILDDIR/lib/libnotmuch.so | awk '$2 == "T" && $1 ~ "^notmuch" {print $1}' | sort | uniq > ACTUAL
> +readelf -Ws $NOTMUCH_BUILDDIR/lib/libnotmuch.so | \
> +    awk '$4  == "FUNC" && $5 == "GLOBAL" && $7 != "UND" {print $8}' | sort | uniq > ACTUAL
>  sed -n 's/^\(notmuch_[a-zA-Z0-9_]*\)[[:blank:]]*(.*/\1/p' $NOTMUCH_SRCDIR/lib/notmuch.h | sort | uniq > EXPORTED
>  test_expect_equal_file EXPORTED ACTUAL
>  
> -- 
> 2.29.2

Re: [PATCH] test/T360-symbol-hiding: use readelf in place of nm

[David Bremner]

Tomi Ollila <tomi.ollila@iki.fi> writes:

> On Fri, Dec 11 2020, David Bremner wrote:
>
>> It turns out that using nm -P isn't as portable as hoped. In particular on
>> architectures using ELF v1 (e.g. ppc64), the desired symbols end up in
>> the data section instead of text.
>>
>> The test is currently only functional on ELF based architectures, so I
>> think it's legit to depend on readelf instead of nm.
>>
>> The switch to readelf has the advantage that we can explicitely ask
>> for all of the symbols with global visibility, rather than grepping
>> for notmuch. That seems a more robust approach since it will catch any
>> strangely named global symbols.
>
> Looks good. I use this opportunity to mention that `sort | uniq` can be
> replaced with `sort -u`  :D

I actually did that, but then thought, there must be some reason we do
it this way, maybe portability? Having just checked, I see it's in POSIX
1003.1-2017 (and apparently earlier versions). So I guess it's most likely
safe.

Applied with s/sort|uniq/sort -u/ and one space deleted.