I ran into a problem on Solaris 10 gcc when building bleeding-edge grep
and installed the attached into Gnulib to fix it.
diff --git a/ChangeLog b/ChangeLog
index d474155df7..61107c67fa 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,16 @@
+2023-01-22 Paul Eggert
+
+ sigsegv, vma-iter: port t
On 1/21/23 18:10, Bruno Haible wrote:
I don't understand. Is the #warning that I added misleading?
No, all I meant was that if I see the warning I do the same thing that I
do if I see the error from the missing #include, so there's not much
point in having a transition period with the warning
Paul Eggert wrote:
> I ... wouldn't object to a very short transition period
Short transition periods tend to cause unnecessary hassle to some users.
I therefore favour 1 year or 2 years as a transition period.
> since the current #warning is something I would treat
> like a diagnostic that getp
On 1/21/23 02:56, Bruno Haible wrote:
For a transition period, "getprogname.h" can continue to exist, but will
emit a deprecation warning.
Thanks for doing all that. I've removed "#include " from
the programs I help maintain, and wouldn't object to a very short
transition period since the cur
The login_tty module does not provide the function's declaration in a
consistent way. The simple recipe
Include:
extern int login_tty (int);
breaks down as soon as Gnulib may want to override the login_tty function.
Which we'll want to do soon on Android.
What Gnulib should do is to provide
One of the changes, two days ago, was wrong: The line
gl_CHECK_FUNCS_ANDROID([login_tty], [[#include ]])
causes the declaration check to be performed with on *all* platforms.
But the function login_tty is declared in or on BSD
platforms. Thus this test fails, ac_cv_func_login_tty comes out a
> alignasof tests: Fix GCC warning (regression 2023-01-15).
> * tests/test-stdalign.c (alignas): Undefine before redefine.
The common file naming pattern is broken after the refactoring from 2023-01-15.
test-stdalign.c is no longer testing (since it does not even
include any more).
On platforms with GCC 5.5.0 or 11.3.0, I see this compilation warning:
../../gltests/test-stdalign.c:38:0: warning: "alignas" redefined
This patch fixes it.
2023-01-21 Bruno Haible
alignasof tests: Fix GCC warning (regression 2023-01-15).
* tests/test-stdalign.c (alignas): U
Why is the getprogname() function declared in "getprogname.h"?
All platforms that have this function — Mac OS X, FreeBSD, NetBSD,
OpenBSD >= 5.4, Solaris >= 11, Cygwin, Android API level >= 21 —
declare it in . It thus appears very unlikely that, if
glibc ever adds this function, they would choose
