We have a problem with the current mc-x.y.z-preW versioning scheme for
both Redhat and Debian. The problem is that
(1) mc-1:4.7.0-1.fc12.x86_64
(2) mc-1:4.7.0.pre4.231.g8cfffc5-1.fc12.x86_64
(1) is considered to be older than (2)
Signed-off-by: Slava Zanko <slavazanko@gmail.com>
Signed-off-by: Andrew Borodin <aborodin@vmail.ru>
* 1916_nonvfs_fixup:
Rewritten vfs_canon_and_translate() function for using glib functions.
VFS: renamed USE_VFS to ENABLE_VFS
VFS: fixed help description for --disable-vfs
src/file.c: Fixed warning on unused parameter on non-vfs build
VFS: removed duplicate O_LINEAR declaration in src/vfsdummy.h
VFS: added non-vfs version of vfs_canon_and_translate()
VFS: added non-vfs version of vfs_translate_path_n()
VFS: added non-vfs version of vfs_get_encoding()
VFS: fixing symbol clash on mc_ungetlocalcopy() at non-vfs build
VFS: fixing symbol clash on mc_readlink() at non-vfs build
VFS: fixing symbol clash on mc_utime() at non-vfs build
VFS: fixing symbol clash on mc_write() at non-vfs build
VFS: fixing symbol clash on mc_read() at non-vfs build
VFS: fixing symbol clash on vfs_file_is_local() at non-vfs build
VFS: fixing symbol clash on vfs_current_is_local() at non-vfs build
VFS: fixing symbol clash on vfs_init() and vfs_shut() at non-vfs build
VFS: fixed lots of missing includes to vfs/vfs.h
Ticket #1916: non-vfs build fixups
Unused static function causes build to break on -Werror.
Fixing this by commenting it out.
(not removing it since it might get used some day)
Signed-off-by: Slava Zanko <slavazanko@gmail.com>
valgrind log snippet: (from here http://pavlinux.ru/vgmc.log)
> ==26750== HEAP SUMMARY:
> ==26750== in use at exit: 0 bytes in 0 blocks
> ==26750== ==26749== Invalid free() / delete / delete[]
> ==26749== at 0x4A06DD8: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==26749== by 0x48B5F9: edit_read_syntax_rules (syntax.c:766)
> ==26749== by 0x48CDA7: edit_read_syntax_file (syntax.c:1140)
> ==26749== by 0x48D06D: edit_load_syntax (syntax.c:1219)
> ==26749== by 0x4762F6: edit_init (edit.c:834)
> ==26749== by 0x4858FC: edit_file (editwidget.c:241)
> ==26749== by 0x44D017: do_edit_at_line (cmd.c:304)
> ==26749== by 0x44D098: do_edit (cmd.c:324)
> ==26749== by 0x44D10E: edit_cmd (cmd.c:331)
> ==26749== by 0x46C2E4: midnight_execute_cmd (main.c:1193)
> ==26749== by 0x46CD41: midnight_callback (main.c:1690)
> ==26749== by 0x4353F4: buttonbar_call (widget.c:2654)
> ==26749== Address 0x60f8e10 is 0 bytes inside a block of size 30 free'd
> ==26749== at 0x4A06DD8: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==26749== by 0x48B32C: open_include_file (syntax.c:705)
> ==26749== by 0x48B6E8: edit_read_syntax_rules (syntax.c:784)
> ==26749== by 0x48CDA7: edit_read_syntax_file (syntax.c:1140)
> ==26749== by 0x48D06D: edit_load_syntax (syntax.c:1219)
> ==26749== by 0x4762F6: edit_init (edit.c:834)
> ==26749== by 0x4858FC: edit_file (editwidget.c:241)
> ==26749== by 0x44D017: do_edit_at_line (cmd.c:304)
> ==26749== by 0x44D098: do_edit (cmd.c:324)
> ==26749== by 0x44D10E: edit_cmd (cmd.c:331)
> ==26749== by 0x46C2E4: midnight_execute_cmd (main.c:1193)
> ==26749== by 0x46CD41: midnight_callback (main.c:1690)
We see doublefree memory corruption here, introduced by spurious 'g_free(error_file_name)'
of changeset:0c17219b2ab5cb5fe2e73f8d7cc9c11c755a3ae4 (syntax.c file)
The rest of code seems to store real syntax file name there.
Making code the same as part above: don't free 'error_file_name'
Signed-off-by: Sergei Trofimovich <slyfox@inbox.ru>
* Bump epoch, because otherwise users wouldn't be able to upgrade
properly from older RPMs. The versioning scheme has to be revisited.
* Add changelog entry.
* Small corrections to the changelog.
During last major escape transition "foo" -> properly_escaped_foo we forgot
to add "'" to list of specialcased symbols. This leads to mc hangup when you
try to do something with files/dirs, containing that symbol.
Adding it!
Signed-off-by: Sergei Trofimovich <slyfox@inbox.ru>
Add the user name to the ssh command line only if it's not the same as
the local logged-in user. OpenSSH will take it as a default value
anyway, but otherwise we'll override ~/.ssh/config settings which might
be not what the user wants.
* Factored out vfs_get_local_username()
* Moved default-user detection to fish subsystem (URL_USE_ANONYMOUS)
* Slight modifications to the comments and wording
