> On Feb 9, 2015, at 7:05 PM, Susan Hinrichs <shinr...@network-geographics.com> > wrote: > >>> >>> On Feb 9, 2015, at 4:12 PM, shinr...@apache.org wrote: >>> >>> >> [snip] >>> http://git-wip-us.apache.org/repos/asf/trafficserver/blob/2dbdd9c2/build/crypto.m4 >>> ---------------------------------------------------------------------- >>> diff --git a/build/crypto.m4 b/build/crypto.m4 >>> index 32276a2..2c83793 100644 >>> --- a/build/crypto.m4 >>> +++ b/build/crypto.m4 >>> @@ -89,7 +89,12 @@ AC_DEFUN([TS_CHECK_CRYPTO_SNI], [ >>> enable_tls_sni=yes >>> >>> TS_ADDTO(LIBS, [$OPENSSL_LIBS]) >>> - AC_CHECK_HEADERS(openssl/tls1.h openssl/ssl.h openssl/ts.h) >>> + AC_CHECK_HEADERS(openssl/ssl.h openssl/ts.h) >>> + AC_CHECK_HEADERS(openssl/tls1.h, [], [], >>> +[ #ifdef HEADER_SSL_H >>> +#include <openssl/tls1.h> >>> +#endif ]) >> Where does HEADER_SSL_H come from? The previous check for <openssl/tls1.h> >> works on my systems ... what configuration was this change for? > > Without this change I was seeing the following when compiling against openssl > 1.0.2 > > checking openssl/tls1.h presence... yes > configure: WARNING: openssl/tls1.h: present but cannot be compiled > configure: WARNING: openssl/tls1.h: check for missing prerequisite > headers? > configure: WARNING: openssl/tls1.h: see the Autoconf documentation > configure: WARNING: openssl/tls1.h: section "Present But Cannot Be > Compiled" > configure: WARNING: openssl/tls1.h: proceeding with the compiler's result
Ah, I see. Does HEADER_SSL_H come from openssl? The more conventional expression would be to use HAVE_OPENSSL_SSL_H as defined by autoconf ... J
