Mercurial > vim
view src/nbdebug.c @ 33581:403d57b06231 v9.0.2035
patch 9.0.2035: [security] use-after-free with wildmenu
Commit: https://github.com/vim/vim/commit/8f4fb007e4d472b09ff6bed9ffa485e0c3093699
Author: Yee Cheng Chin <ychin.git@gmail.com>
Date: Tue Oct 17 10:06:56 2023 +0200
patch 9.0.2035: [security] use-after-free with wildmenu
Problem: [security] use-after-free with wildmenu
Solution: properly clean up the wildmenu when exiting
Fix wildchar/wildmenu/pum memory corruption with special wildchar's
Currently, using `wildchar=<Esc>` or `wildchar=<C-\>` can lead to a
memory corruption if using wildmenu+pum, or wrong states if only using
wildmenu. This is due to the code only using one single place inside the
cmdline process loop to perform wild menu clean up (by checking
`end_wildmenu`) but there are other odd situations where the loop could
have exited and we need a post-loop clean up just to be sure. If the
clean up was not done you would have a stale popup menu referring to
invalid memory, or if not using popup menu, incorrect status line (if
`laststatus=0`).
For example, if you hit `<Esc>` two times when it's wildchar, there's a
hard-coded behavior to exit command-line as a failsafe for user, and if
you hit `<C-\><C-\><C-N>` it will also exit command-line, but the clean
up code would not have hit because of specialized `<C-\>` handling.
Fix Ctrl-E / Ctrl-Y to not cancel/accept wildmenu if they are also
used for 'wildchar'/'wildcharm'. Currently they don't behave properly,
and also have potentially memory unsafe behavior as the logic is
currently not accounting for this situation and try to do both.
(Previous patch that addressed this: #11677)
Also, correctly document Escape key behavior (double-hit it to escape)
in wildchar docs as it's previously undocumented.
In addition, block known invalid chars to be set in `wildchar` option,
such as Ctrl-C and `<CR>`. This is just to make it clear to the user
they shouldn't be set, and is not required for this bug fix.
closes: #13361
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
author | Christian Brabandt <cb@256bit.org> |
---|---|
date | Tue, 17 Oct 2023 10:15:08 +0200 |
parents | 50555279168b |
children |
line wrap: on
line source
/* vi:set ts=8 sw=8 noet: * * VIM - Vi IMproved by Bram Moolenaar * Visual Workshop integration by Gordon Prieur * * Do ":help uganda" in Vim to read copying and usage conditions. * Do ":help credits" in Vim to see a list of people who contributed. * See README.txt for an overview of the Vim source code. */ /* * NetBeans Debugging Tools. What are these tools and why are they important? * There are two main tools here. The first tool is a tool for delaying or * stopping gvim during startup. The second tool is a protocol log tool. * * The startup delay tool is called nbdebug_wait(). This is very important for * debugging startup problems because gvim will be started automatically from * netbeans and cannot be run directly from a debugger. The only way to debug * a gvim started by netbeans is by attaching a debugger to it. Without this * tool all startup code will have completed before you can get the pid and * attach. * * The second tool is a log tool. * * This code must have NBDEBUG defined for it to be compiled into vim/gvim. */ #ifdef NBDEBUG #include "vim.h" FILE *nb_debug = NULL; u_int nb_dlevel = 0; // nb_debug verbosity level void nbdb(char *, ...) ATTRIBUTE_FORMAT_PRINTF(1, 2); static int lookup(char *); #ifdef USE_NB_ERRORHANDLER static int errorHandler(Display *, XErrorEvent *); #endif /* * nbdebug_wait - This function can be used to delay or stop execution of vim. * It's normally used to delay startup while attaching a * debugger to a running process. Since NetBeans starts gvim * from a background process this is the only way to debug * startup problems. */ void nbdebug_wait( u_int wait_flags, // tells what to do char *wait_var, // wait environment variable u_int wait_secs) // how many seconds to wait { init_homedir(); // not inited yet #ifdef USE_WDDUMP WDDump(0, 0, 0); #endif // for debugging purposes only if (wait_flags & WT_ENV && wait_var && getenv(wait_var) != NULL) { sleep(atoi(getenv(wait_var))); } else if (wait_flags & WT_WAIT && lookup("~/.gvimwait")) { sleep(wait_secs > 0 && wait_secs < 120 ? wait_secs : 20); } else if (wait_flags & WT_STOP && lookup("~/.gvimstop")) { int w = 1; while (w) { ; } } } void nbdebug_log_init( char *log_var, // env var with log file char *level_var) // env var with nb_debug level { char *file; // possible nb_debug output file char *cp; // nb_dlevel pointer if (log_var && (file = getenv(log_var)) != NULL) { time_t now; nb_debug = fopen(file, "a"); time(&now); fprintf(nb_debug, "%s", get_ctime(now, TRUE)); if (level_var && (cp = getenv(level_var)) != NULL) { nb_dlevel = strtoul(cp, NULL, 0); } else { nb_dlevel = NB_TRACE; // default level } #ifdef USE_NB_ERRORHANDLER XSetErrorHandler(errorHandler); #endif } } void nbdbg(char *fmt, ...) { va_list ap; if (nb_debug != NULL && nb_dlevel & NB_TRACE) { va_start(ap, fmt); vfprintf(nb_debug, fmt, ap); va_end(ap); fflush(nb_debug); } } static int lookup(char *file) { char buf[BUFSIZ]; expand_env((char_u *) file, (char_u *) buf, BUFSIZ); return #ifndef FEAT_GUI_MSWIN (access(buf, F_OK) == 0); #else (access(buf, 0) == 0); #endif } #ifdef USE_NB_ERRORHANDLER static int errorHandler( Display *dpy, XErrorEvent *err) { char msg[256]; char buf[256]; XGetErrorText(dpy, err->error_code, msg, sizeof(msg)); nbdbg("\n\nNBDEBUG Vim: X Error of failed request: %s\n", msg); sprintf(buf, "%d", err->request_code); XGetErrorDatabaseText(dpy, "XRequest", buf, "Unknown", msg, sizeof(msg)); nbdbg("\tMajor opcode of failed request: %d (%s)\n", err->request_code, msg); if (err->request_code > 128) { nbdbg("\tMinor opcode of failed request: %d\n", err->minor_code); } return 0; } #endif #endif // NBDEBUG