Message ID | 1541757573-30178-1-git-send-email-lstipakov@gmail.com |
---|---|
State | Accepted |
Headers | show |
Series | [Openvpn-devel] Remove extra token after #endif | expand |
Lev Stipakov <lstipakov@gmail.com> on Fri, 2018/11/09 11:59: > Commit ee80ce3d6f2ebc59068338757311e0488ae620fc wrapped > code in #ifdef/#endif and added extra token after #endif, > which produces compiler warning. > > This removes unneeded extra token. > > Signed-off-by: Lev Stipakov <lstipakov@gmail.com> > --- > src/openvpn/init.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/src/openvpn/init.c b/src/openvpn/init.c > index 586e4ca..1cdef31 100644 > --- a/src/openvpn/init.c > +++ b/src/openvpn/init.c > @@ -624,7 +624,7 @@ save_ncp_options(struct context *c) > c->c1.ciphername = c->options.ciphername; > c->c1.authname = c->options.authname; > c->c1.keysize = c->options.keysize; > -#endif ENABLE_CRYPTO > +#endif Perhaps this should go into comment: #endif /* ENABLE_CRYPTO */
Acked-by: Gert Doering <gert@greenie.muc.de> Thanks. Regarding Christian's comment about "#endif" vs. "#endif /* ENABLE_CRYPTO /*" - well, we have both in our tree, so neither is "correct" or "wrong". For such a short #ifdef/#endif span, I'd see it as optional because the corresponding #ifdef is easily spotted. For something spanning more code, possibly nested #ifdef, a comment after the #endif would definitely be better, though. Your patch has been applied to the release/2.4 branch. commit ca962f3837ab05b0e8a4382960391f289497eaa2 Author: Lev Stipakov Date: Fri Nov 9 11:59:33 2018 +0200 Remove extra token after #endif Signed-off-by: Lev Stipakov <lstipakov@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de> Message-Id: <1541757573-30178-1-git-send-email-lstipakov@gmail.com> URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg17883.html Signed-off-by: Gert Doering <gert@greenie.muc.de> -- kind regards, Gert Doering
On 21/11/2018 10:34, Gert Doering wrote: > > Regarding Christian's comment about "#endif" vs. "#endif /* ENABLE_CRYPTO /*" > - well, we have both in our tree, so neither is "correct" or "wrong". > > For such a short #ifdef/#endif span, I'd see it as optional because the > corresponding #ifdef is easily spotted. For something spanning more > code, possibly nested #ifdef, a comment after the #endif would definitely > be better, though. Just FTR; I share this point of view. We've not been too picky about these #endif remarks in the past but something we saw more of after the big reformatting commit right before the v2.4.0 release. We used these uncrustify settings for the great reformatting patch: ------------------- # Annotate #else and #endif statements mod_add_long_ifdef_endif_comment=20 mod_add_long_ifdef_else_comment=5 -------------------
diff --git a/src/openvpn/init.c b/src/openvpn/init.c index 586e4ca..1cdef31 100644 --- a/src/openvpn/init.c +++ b/src/openvpn/init.c @@ -624,7 +624,7 @@ save_ncp_options(struct context *c) c->c1.ciphername = c->options.ciphername; c->c1.authname = c->options.authname; c->c1.keysize = c->options.keysize; -#endif ENABLE_CRYPTO +#endif } /* Restores NCP-negotiable options to original values */ @@ -635,7 +635,7 @@ restore_ncp_options(struct context *c) c->options.ciphername = c->c1.ciphername; c->options.authname = c->c1.authname; c->options.keysize = c->c1.keysize; -#endif ENABLE_CRYPTO +#endif } void
Commit ee80ce3d6f2ebc59068338757311e0488ae620fc wrapped code in #ifdef/#endif and added extra token after #endif, which produces compiler warning. This removes unneeded extra token. Signed-off-by: Lev Stipakov <lstipakov@gmail.com> --- src/openvpn/init.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)