* 2926_screen_lib_flags:
Clarify usage of AM_CPPFLAGS.
PCRE_CFLAGS are CPPFLAGS, actually.
m4.include/mc-with-screen-slang.m4: clarify usage of CPPFLAGS, CFLAGS and LDFLAGS
Ticket #2926: fail to check ncurses library
...if --with-ncurses-inc and --with-ncurses-libs options are used.
m4.include/mc-with-screen-ncurses.m4: clarify usage of LDFLAGS in
ncurses library checking: add $ac_curses_lib_path.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
* 2924_preserve_dir_attributes:
(copy_dir_dir): get rid of extra string duplication.
(copy_dir_dir): refactoring: get rid of goto dont_mkdir.
Ticket #2924: attributes of existing directories are never preserved.
If superuser copies a directory tree (e.g. from a safe backup location
(with sane permission, owner, group, timestamp) over an existing tree,
any attributes are not preserved. All destination dirs retain all their
existing attributes instead of change attributes from source.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
In
tests/lib/mcconfig/Makefile.am
tests/lib/vfs/Makefile.am
tests/src/filemanager/Makefile.am
AM_LDFLAGS = -Wl,-z,muldefs
is used, but on cygwin the linker does not accept the "-z" option and
AM_LDFLAGS = -Wl,--allow-multiple-definition
is required instead.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
Set defines via CPPFLAGS variable not via CFLAGS one.
Use AM_CPPFLAGS and AM_CFLAGS variables instead of per-target ones.
Cherry-picked b37af47023.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
* 2906_copy_crash_on_solaris:
(get_fs_usage): avoid compile warning about mixed declarations and code.
Sync with gnulib ffe10f3d20c2eb77efd56b68f2c22b2a810c289d.
Sync with gnulib 2ab2617ee340ff35a9a4c713004fb302868d41b0
Sync with gnulib 0a6dafee2bb8f69a931345a1f67d835ee5062967
Sync with gnulib b1fac377605c0eef8844fc8d3818d360f37d6fa4:
Ticket #2906: crash on Solaris while trying to copy a file.
fsusage: port back to Solaris
* src/filemanager/mountlist.c (get_fs_usage): Fix busted logic causing
compile-time error (fsd not declared) on Solaris 10.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
* src/filemanager/mountlist.c [STAT_STATVFS && (__linux__ && (__GLIBC__||__UCLIBC__))]:
Undefine STAT_STATFS2_FRSIZE to exclude code not used in this case.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
Avoid needless check on GNU/Linux.
* m4.include/fsusage.m4 (gl_FILE_SYSTEM_USAGE): Omit STAT_STATFS3_OSF1 check
on GNU/Linux systems, since it can't possibly work.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
fs usage: fix block size returned on older Linux 2.6.
* src/filemanager/mountlist.c: Fall back to (struct statfs).f_frsize
which is available since Linux 2.6.
* m4.include/fsusage.m4 (STAT_STATFS2_FRSIZE): Always define
when the member is available so it can be used as a fallback.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
Sync with gnulib c25bdbae48977a527dff69150f59fb0746d31b51:
fs usage: check for GNU/Linux statvfs problem dynamically.
* src/filemanager/mountlist.c [STAT_STATVFS && __linux__ && (__GLIBC__||__UCLIBC__)]:
Define STAT_STATFS2_BSIZE too, since in this case the code now
checks dynamically whether statvfs is reliable, falling back on
Linux-style statfs otherwise.
(statvfs_works): New function, for dynamically testing statvfs.
(get_fs_usage) [STAT_STATVFS]: Use it.
* src/filemanager/filegui.c (filegui__check_attrs_on_fs): apply the same
statvfs_works function.
* m4.include/fsusage.m4 (gl_FILE_SYSTEM_USAGE): Remove static check for
statvfs on GNU/Linux hosts, since it's now done dynamically.
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
* 2877_stable_cleanup:
Update EN and RU man pages: add missing options.
fixed doxygen.cfg, excluded directory /tests/ from doxygen path's
Ticket #2861: additional error handling in edit_save_block().
(str_utf8_normalize): ret rid of extra memory allocation.
Use g_strndup() instead of g_strdup()
(get_absolute_name): missing finish NULL in vfs_path_append_vpath_new() call.
(edit_do_undo): cleanup of labeled statement usage.
mcdiffview: code cleanup and cosmetics.
Cleanup of screen library detection m4 code.
Refactoring of many conditions.
(edit_do_search): refactoring of search loop condition.
Avoid double checks of bookmarks in editor.
Ticket #2877: code cleanup before 4.8.1.6 release.
If edit->column_highlight is on and the disk is so full that the editor
is unable even to write the magic (r <= 0), then subsequent if (len)
check might fail, since the value of len is undefined.
The solution is to initialize len with a non-zero value, so that the
function properly returns an error value in all cases (adding an
explicit return 0; is also possible, but then one must take care of
closing file descriptors, which is less convenient).
Signed-off-by: Yury V. Zaytsev <yury@shurup.com>
