[prev in list] [next in list] [prev in thread] [next in thread] 

List:       krb5-bugs
Subject:    [krbdev.mit.edu #8928] [Comment] [PATCH] Constructor/destructor configure check fails on certain com
From:       "Greg Hudson via RT" <rt-comment () krbdev ! mit ! edu>
Date:       2020-07-27 21:07:19
Message-ID: rt-4.4.4-24430-1595884039-1808.8928-8-0 () mit ! edu
[Download RAW message or body]


https://krbdev.mit.edu/rt/Ticket/Display.html?id=8928
This is a comment.  It is not sent to the Requestor(s):

Additional notes:

* The compiler line doesn't include -Werror in the options, but the error
message does, which is weird.
* Using godbolt.org, recent versions of the test program don't seem to give an
error on the test program by default.
* The presence of -fno-common in the compiler options suggests macOS
(aclocal.m4 line 589).
* If I explicitly configure with CFLAGS=-Werror, lots of tests go wrong.
TRY_WARN_CC_FLAG_1 errors out entirely (fixable by changing "1;" to ";" in the
AC_TRY_COMPILE program arguments), Tcl isn't found, functions like vsprintf and
strdup aren't found, etc..
* A search of the autoconf bug list suggests that making configure work with
CFLAGS=-Werror is not a design goal.

It's conceivable that Apple has modified clang to error out on
implicit-function-declaration by default, and that the diagnostic includes
Werror due to the nature of the local modification. If so, there may be two
other tests giving incorrect results. My (Linux) config.log reports an implicit
declaration warning for Tcl_CreateInterp in the tcl tests and for printf in the
in6addr_any test. AC_CHECK_FUNC doesn't seem to generate implicit declaration
warnings all the time because it sticks a "char fnname ();" in the body, but is
does generate "conflicting types for built-in function" warnings when the
function being tested for has a compiler builtin.


_______________________________________________
krb5-bugs mailing list
krb5-bugs@mit.edu
https://mailman.mit.edu/mailman/listinfo/krb5-bugs
[prev in list] [next in list] [prev in thread] [next in thread]