[Buildroot] [PATCH 1/1] package/rtty: fix wolfssl build

[Peter Korsgaard]

>>>>> "Yann" == Yann E MORIN <yann.morin.1998 at free.fr> writes:

 > Fabrice, All,
 > On 2023-11-30 23:49 +0100, Fabrice Fontaine spake thusly:
 >> Fix the following wolfssl build failure raised at least since bump to
 >> version 7.4.0 in commit 6b5907bf65d27ed98532e9783f92f5575f38b3d2:
 >> 
 >> /home/autobuild/autobuild/instance-4/output-1/build/rtty-8.1.0/src/ssl/openssl.c: In function 'ssl_last_error_string':
 >> /home/autobuild/autobuild/instance-4/output-1/build/rtty-8.1.0/src/ssl/openssl.c:143:24:
 >> error: implicit declaration of function 'ERR_peek_error_line_data';
 >> did you mean 'wolfSSL_ERR_get_error_line_data'?
 >> [-Werror=implicit-function-declaration]
 >> 143 |         ssl_err_code = ERR_peek_error_line_data(&file, &line, &data, &flags);
 >> |                        ^~~~~~~~~~~~~~~~~~~~~~~~
 >> |                        wolfSSL_ERR_get_error_line_data
 >> 
 >> Fixes:
 >> - http://autobuild.buildroot.org/results/9db9f1dcc6760de4b78771bb79f109c4efd06c36
 >> - http://autobuild.buildroot.org/results/16422af9469de114e552124542508c3b18ea8f19
 >> 
 >> Signed-off-by: Fabrice Fontaine <fontaine.fabrice at gmail.com>
 >> ---
 >> package/rtty/Config.in | 2 ++
 >> package/rtty/rtty.mk   | 2 +-
 >> 2 files changed, 3 insertions(+), 1 deletion(-)
 >> 
 >> diff --git a/package/rtty/Config.in b/package/rtty/Config.in
 >> index 198d465713..434c5d6ae7 100644
 >> --- a/package/rtty/Config.in
 >> +++ b/package/rtty/Config.in
 >> @@ -2,6 +2,8 @@ config BR2_PACKAGE_RTTY
 >> bool "rtty"
 >> depends on BR2_USE_MMU # fork()
 >> select BR2_PACKAGE_LIBEV
 >> +	select BR2_PACKAGE_WOLFSSL_ALL if BR2_PACKAGE_WOLFSSL && \
 >> +		!BR2_STATIC_LIBS

 > That's not nice, as that forces wolfssl-all even in the case that it is
 > not used by rtty, i.e. when either mbedtls or openssl are enabled.

 > So, the full condition should be something like:

 >     select BR2_PACKAGE_WOLFSSL_ALL if BR2_PACKAGE_WOLFSSL && \
 >         !BR2_STATIC_LIBS && !BR2_PACKAGE_MBEDTLS && !BR2_PACKAGE_OPENSSL

 > But that becomes a bit too much.

 > I think the simplest solution is to just carry the .mk hunk of this
 > patch for now.

 > Eventually, I think it would be bettter to have a choice for the crypto
 > backend, like we have elsewhere, rather than rely on ordering of the
 > conditional blocks to defone what gets used, especially since that:

 >   - it does not match the ordering that upstream uses to find the
 >     backend (prefers openssl, then wolfssl, then mnedtls);

 >   - the ordering does not match the other places (e.g. libuhttpd [0] and
 >     libwebsockets have the same set of crypto backends, but each in yet
 >     a different order);

 >   - ordering is subject to being, well, reordered, and people would get
 >     confused when they update their buildroot to see their rtty change
 >     its crypto backend...

 > [0] libuhttpd suffers from the same issue in its Config.in, that this
 > patch would introduce: wolfssl-all os forced even if wolfssl will not be
 > used.

 > Applied to master, just the .mk hunk, thanks.

Committed to 2023.02.x and 2023.08.x, thanks.

-- 
Bye, Peter Korsgaard