On 11/15/21 10:37, Sudhip Nashi wrote:
Turns out lseek is broken (or at least works differently) on macOS as well
(https://lists.gnu.org/archive/html/bug-gnulib/2018-09/msg00054.html). Funny
coincidence! I’ll take a better look later this week if I can and try to see
what the exact problem is.
Thanks, I think I see the problem now.
Eventually macOS will likely get fixed to work around this
lseek+SEEK_DATA incompatibility (as FreeBSD, Solaris, etc. all do things
the Linux way and that's what I think will appear in the next POSIX),
but in the meantime I attempted to work around the portability issue by
installing the attached patch into Gnulib, and by syncing coreutils to
the latest Gnulib.
I don't use macOS so have not tested this. Please give it a try, either
by building from bleeding-edge coreutils on Savannah, or by building
from the tarball temporarily here:
https://www.cs.ucla.edu/~eggert/coreutils-9.0.26-0f4d9.tar.gz
Thanks.
From 4db8db34112b86ddf8bac48f16b5acff732b5fa9 Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Mon, 15 Nov 2021 15:08:25 -0800
Subject: [PATCH] lseek: port around macOS SEEK_DATA glitch
Problem reported by Sudhip Nashi (Bug#51857).
* doc/posix-functions/lseek.texi (lseek): Mention macOS SEEK_DATA
issue.
* lib/lseek.c (rpl_lseek): Work around macOS portability glitch.
* m4/lseek.m4 (gl_FUNC_LSEEK): Replace lseek on Darwin.
* modules/lseek (Depends-on): Depend on msvc-nothrow
and fstat only if needed.
---
ChangeLog | 11 +++++++++++
doc/posix-functions/lseek.texi | 4 ++++
lib/lseek.c | 16 ++++++++++++++++
m4/lseek.m4 | 10 ++++++++--
modules/lseek | 4 ++--
5 files changed, 41 insertions(+), 4 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index f47071a72..71a226570 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,14 @@
+2021-11-15 Paul Eggert <egg...@cs.ucla.edu>
+
+ lseek: port around macOS SEEK_DATA glitch
+ Problem reported by Sudhip Nashi (Bug#51857).
+ * doc/posix-functions/lseek.texi (lseek): Mention macOS SEEK_DATA
+ issue.
+ * lib/lseek.c (rpl_lseek): Work around macOS portability glitch.
+ * m4/lseek.m4 (gl_FUNC_LSEEK): Replace lseek on Darwin.
+ * modules/lseek (Depends-on): Depend on msvc-nothrow
+ and fstat only if needed.
+
2021-11-11 Fabrice Fontaine <fontaine.fabr...@gmail.com> (tiny change)
sigsegv: fix builds on microblazeel, or1k
diff --git a/doc/posix-functions/lseek.texi b/doc/posix-functions/lseek.texi
index 4a9d55dcf..2f8e2b587 100644
--- a/doc/posix-functions/lseek.texi
+++ b/doc/posix-functions/lseek.texi
@@ -9,6 +9,10 @@ Gnulib module: lseek
Portability problems fixed by Gnulib:
@itemize
@item
+On some platforms, @code{lseek (fd, offset, SEEK_DATA)} returns a value
+greater than @code{offset} even when @code{offset} addresses data:
+macOS 12
+@item
This function is declared in a different header file (namely, @code{<io.h>})
on some platforms:
MSVC 14.
diff --git a/lib/lseek.c b/lib/lseek.c
index 0042546a8..7dcd6c9da 100644
--- a/lib/lseek.c
+++ b/lib/lseek.c
@@ -52,6 +52,22 @@ rpl_lseek (int fd, off_t offset, int whence)
errno = ESPIPE;
return -1;
}
+#elif defined __APPLE__ && defined __MACH__ && defined SEEK_DATA
+ if (whence == SEEK_DATA)
+ {
+ /* If OFFSET points to data, macOS lseek+SEEK_DATA returns the
+ start S of the first data region that begins *after* OFFSET,
+ where the region from OFFSET to S consists of possibly-empty
+ data followed by a possibly-empty hole. To work around this
+ portability glitch, check whether OFFSET is within data by
+ using lseek+SEEK_HOLE, and if so return to OFFSET by using
+ lseek+SEEK_SET. */
+ off_t next_hole = lseek (fd, offset, SEEK_HOLE);
+ if (next_hole < 0)
+ return next_hole;
+ if (next_hole != offset)
+ whence = SEEK_SET;
+ }
#else
/* BeOS lseek mistakenly succeeds on pipes... */
struct stat statbuf;
diff --git a/m4/lseek.m4 b/m4/lseek.m4
index 0af63780a..faab09b73 100644
--- a/m4/lseek.m4
+++ b/m4/lseek.m4
@@ -1,4 +1,4 @@
-# lseek.m4 serial 11
+# lseek.m4 serial 12
dnl Copyright (C) 2007, 2009-2021 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -59,7 +59,7 @@ AC_DEFUN([gl_FUNC_LSEEK],
;;
esac
])
- if test $gl_cv_func_lseek_pipe = no; then
+ if test "$gl_cv_func_lseek_pipe" = no; then
REPLACE_LSEEK=1
AC_DEFINE([LSEEK_PIPE_BROKEN], [1],
[Define to 1 if lseek does not detect pipes.])
@@ -69,4 +69,10 @@ AC_DEFUN([gl_FUNC_LSEEK],
if test $WINDOWS_64_BIT_OFF_T = 1; then
REPLACE_LSEEK=1
fi
+
+ dnl macOS SEEK_DATA is incompatible with other platforms.
+ case $host_os in
+ darwin*)
+ REPLACE_LSEEK=1;;
+ esac
])
diff --git a/modules/lseek b/modules/lseek
index ced443123..f60809319 100644
--- a/modules/lseek
+++ b/modules/lseek
@@ -9,8 +9,8 @@ Depends-on:
unistd
sys_types
largefile
-msvc-nothrow [test $REPLACE_LSEEK = 1]
-fstat [test $REPLACE_LSEEK = 1]
+msvc-nothrow [test $WINDOWS_64_BIT_OFF_T = 1]
+fstat [test "$gl_cv_func_lseek_pipe" = no]
configure.ac:
gl_FUNC_LSEEK
--
2.33.1
