* utilfs.c (vfs_split_url): don't assign *pass if pass is NULL
* fish.c (archive_open, archive_same): memory allocated by
vfs_split_url () must be released after use, don't strdup()
values allocated by vfs_split_url () and NULL passed to
vfs_split_url () when password value is unused
* ftpfs.c (archive_open, archive_same): ditto plus same for
ftpfs_get_proxy_host_and_port ()
* syntax.c: Wery dirty fix to allow compilation of mceditor on ncurses
based systems. Hope, this will make FreeBSD midc maintainer to enable
build-in editor by default. Still, syntax highlighting is disabled.
Need to port...
* widget.c ([forward|backward]_word): Changed logic of moving - now,
it skips only spaces+punct OR alnum. Reason to have it deleting lines
like "lynx http://www.gnome.org", where "lynx http:" part was erased
in one operation. That's nasty, IMHO. Made this a separate commit, so,
if it seriously breaks compatability it's easy to revert it. Still hope,
nobody'll do that.
* ftpfs.c (netrc_next): Turned strange "const char * const keywords"
into more alike "const char const * keywords". Still, think, it's
too strict.
* smbfs.c: Add #define BOOL_DEFINED before inclusion of samba headers.
Libncurses(or slang?) defines BOOL and that can cause problems. (In fact,
I can't find references to that const, but sure, I hade reason to do that:)
* util-alone.h: Declare load_anon_passwd() to avoid warnings.
* file.c, widget.[ch]: Fixed long existing problems with filenames
and input widgets, then char is greater than 128. They were processed
incorrectly. As usual, char != unsigned char problem.
2000-03-03 Björn Eriksson <mdeans@algonet.se>
* Similar patches came from Andrew V. Samoilov <sav@bcs.zp.ua>.
* mad.[ch]: Some functions wrongly by-passed mad_alloc*() (causing
problems when later g_free()ed.
* Added mad_strndup().
* Repeated code collected in mad_fatal_error().
* The gcc-extension/C99-standard variable argument macros (adding
__FILE__, __LINE__ before ', ...') would be nice.
* Alloc_idx_hint and code added, the mad code was way too slow for my
poor computer. A hash-table would be nice.
* (mad_strconcat, mad_strdup_vprintf): g_malloc and g_new are
undefined at the top of mad.c, so mad_alloc must be used.
* vfs.c (mc_def_getlocalcopy): free() replaced by g_free() to prevent
crushes with mad, thanks to Timur I. Bakeyev <timur@bat.ru>
* fish.c (linear_start): name wasn't freed after usage
* cpio.c (cpio_ungetlocalcopy): return type changed to int,
function return 0 now
* direntry.c (vfs_s_close, vfs_s_resolve_symlink): memory, allocated
by vfs_s_fullpath () is freed after use
* ftpfs.c (linear_start): ditto;
* gaction.c (gmc_open_filename): Usae open.flags to determine
whether we need a terminal.
(gmc_open_filename): Fix the logic to make it work properly with
.desktop and .kdelnk files.
Patch by Andrew V. Samoilov <sav@bcs.zp.ua>
* vfs.c (vfs_parse_ls_lga): don't hard code st_ino and st_dev to 0's,
these variables must be assigned in vfs_s_new_inode () or extfs
layer, so ftpfs subdirectories now can be copied.
Patch by Andrew V. Samoilov <sav@bcs.zp.ua>
* vfs.[ch] (vfs_canon, mc_open, vfs_file_is_local,
mc_(un)?getlocalcopy): added const qualifier
(mc_def_getlocalcopy): tempnam() returns malloc()ed buffer,
so, free() it, not g_free();
(vfs_canon): prepend filename by cwd when tilde '~' cannot be expanded;
* vfs/ftpfs.c: Definition of PORT changed in the middle of ftpfs.c
with unexpected side effects (only when HSC_PROXY was
defined). Introduce HSC_PROXY_PORT and FTP_COMMAND_PORT and use
where appropriate.
(dir_load): Do not hard code the directory cache timeout to 10
seconds. Instead use the setting from Options/Virtual VFS ...
* vfs/fish.c (file_store): honour changed semantics for return value
(on error -1 is expected).
* vfs/ftpfs.c (file_store): dito
* gtkedit/editcmd.c (edit_replace_prompt): fix incorrect
initialization of array of structure (gcc version 2.96 20000131
chokes on it).
* lib/mc.hint: C-y and not C-u can be used to insert text in input
lines (e. g. text which has been deleted with M-d).
* configure.in: Fix the extfs-configure-fragment-in-final-output bug
* src/main.c (parse_control_file): add missing right parentheses
* src/boxes.c: fix incorrect initialization of array of structure (gcc
version 2.96 20000131 chokes on it).
in particular, leaking syntax.c that was chomping memory.
Also, should note the bzip2 and gzip filters and better support
for Windows CR importing - windows guys should take a look at
this if you are still out there. dunno if it works yet.