view dynodump/README @ 4477:e34711681f30

Don't determine whether to call general device-type code at startup, rather decide in the device-specific code itself. lisp/ChangeLog addition: 2008-07-07 Aidan Kehoe <kehoea@parhasard.net> Patch to make it up to the device-specific code whether various Lisp functions should be called during device creation, not relying on the startup code to decide this. Also, rename initial-window-system to initial-device-type (which makes more sense in this scheme), always set it. * startup.el (command-line): Use initial-device-type, not initial-window-system; just call #'make-device, leave the special behaviour to be done the first time a console type is initialised to be decided on by the respective console code. * x-init.el (x-app-defaults-directory): Declare that it should be bound. (x-define-dead-key): Have the macro take a DEVICE argument. (x-initialize-compose): Have the function take a DEVICE argument, and use it when checking if various keysyms are available on the keyboard. (x-initialize-keyboard): Have the function take a DEVICE argument, allowing device-specific keyboard initialisation. (make-device-early-x-entry-point-called-p): New. (make-device-late-x-entry-point-called-p): New. Rename pre-x-win-initted, x-win-initted. (make-device-early-x-entry-point): Rename init-pre-x-win, take the call to make-x-device out (it should be called from the device-creation code, not vice-versa). (make-device-late-x-entry-point): Rename init-post-x-win, have it take a DEVICE argument, use that DEVICE argument when working out what device-specific things need doing. Don't use create-console-hook in core code. * x-win-xfree86.el (x-win-init-xfree86): Take a DEVICE argument; use it. * x-win-sun.el (x-win-init-sun): Take a DEVICE argument; use it. * mule/mule-x-init.el: Remove #'init-mule-x-win, an empty function. * tty-init.el (make-device-early-tty-entry-point-called-p): New. Rename pre-tty-win-initted. (make-device-early-tty-entry-point): New. Rename init-pre-tty-win. (make-frame-after-init-entry-point): New. Rename init-post-tty-win to better reflect when it's called. * gtk-init.el (gtk-early-lisp-options-file): New. Move this path to a documented variable. (gtk-command-switch-alist): Wrap the docstring to fewer than 79 columns. (make-device-early-gtk-entry-point-called-p): New. (make-device-late-gtk-entry-point-called-p): New. Renamed gtk-pre-win-initted, gtk-post-win-initted to these. (make-device-early-gtk-entry-point): New. (make-device-late-gtk-entry-point): New. Renamed init-pre-gtk-win, init-post-gtk-win to these. Have make-device-late-gtk-entry-point take a device argument, and use it; have make-device-early-gtk-entry-point load the GTK-specific startup code, instead of doing that in C. (init-gtk-win): Deleted, functionality moved to the GTK device creation code. (gtk-define-dead-key): Have it take a DEVICE argument; use this argument. (gtk-initialize-compose): Ditto. * coding.el (set-terminal-coding-system): Correct the docstring; the function isn't broken. src/ChangeLog addition: 2008-07-07 Aidan Kehoe <kehoea@parhasard.net> Patch to make it up to the device-specific code whether various Lisp functions should be called during device creation, not relying on the startup code to decide this. Also, rename initial-window-system to initial-device-type (which makes more sense in this scheme), always set it. * redisplay.c (Vinitial_device_type): New. (Vinitial_window_system): Removed. Rename initial-window-system to initial-device type, making it a stream if we're noninteractive. Update its docstring. * device-x.c (Qmake_device_early_x_entry_point, Qmake_device_late_x_entry_point): New. Rename Qinit_pre_x_win, Qinit_post_x_win. (x_init_device): Call #'make-device-early-x-entry-point earlier, now we rely on it to find the application class and the app-defaults directory. (x_finish_init_device): Call #'make-device-late-x-entry-point with the created device. (Vx_app_defaults_directory): Always make this available, to simplify code in x-init.el. * device-tty.c (Qmake_device_early_tty_entry_point): New. Rename Qinit_pre_tty_win, rename Qinit_post_tty_win and move to frame-tty.c as Qmake_frame_after_init_entry_point. (tty_init_device): Call #'make-device-early-tty-entry-point before doing anything. * frame-tty.c (Qmake_frame_after_init_entry_point): New. * frame-tty.c (tty_after_init_frame): Have it call the better-named #'make-frame-after-init-entry-point function instead of #'init-post-tty-win (since it's called after frame, not device, creation). * device-msw.c (Qmake_device_early_mswindows_entry_point, Qmake_device_late_mswindows_entry_point): New. Rename Qinit_pre_mswindows_win, Qinit_post_mswindows_win. (mswindows_init_device): Call #'make-device-early-mswindows-entry-point here, instead of having its predecessor call us. (mswindows_finish_init_device): Call #'make-device-early-mswindows-entry-point, for symmetry with the other device types (though it's an empty function). * device-gtk.c (Qmake_device_early_gtk_entry_point, Qmake_device_late_gtk_entry_point): New. Rename Qinit_pre_gtk_win, Qinit_post_gtk_win. (gtk_init_device): Call #'make-device-early-gtk-entry-point; don't load ~/.xemacs/gtk-options.el ourselves, leave that to lisp. (gtk_finish_init_device): Call #'make-device-late-gtk-entry-point with the created device as an argument.
author Aidan Kehoe <kehoea@parhasard.net>
date Wed, 09 Jul 2008 20:46:22 +0200
parents 376386a54a3c
children
line wrap: on
line source

		DYNODUMP
		--------

Dynodump, not to be confused with DinoTurd (as seen in Jurassic Park),
is a shared object that provides one function:

	int dynodump(char *new_file);

dynodump(), called from a running program will write a new executable
in new_file a la unexec() in GNU Emacs.  The difference lies in the
relocations.

dynodump() will create an image with any relocations (which were
performed by the run-time dynamic linker) undone.  This allows the new
image to be run in a different environment.  There is, however, one
potentially major caveat.  If a symbol reference gets updated during
the running of the calling program, its updated value will be lost.
An example (with additional blank lines for legibility):

	$ cat lib.c
	char _foo[] = "hello";
	char _bar[] = "world";

	$ cc -G -o lib.so lib.c

	$ cat prog.c
	extern char _foo, _bar;
 
	int beenhere = 0;
	char * foo = &_foo;
	char * bar = &_bar;
 
	int
	main(void)
	{
	    (void) printf("%d: foo = %x\n", beenhere, foo);
	    (void) printf("%d: bar = %x, ", beenhere, bar);
 
	    if (!beenhere) {
	        beenhere = 1;
	        bar++;
	        dynodump("newfile");
	    }
	    (void) printf("%x\n", bar);
	}

	$ cc -o prog prog.c -R. lib.so dynodump.so

	$ ./prog
	0: foo = ef7503cc
	0: bar = ef7503d2, ef7503d3

	$ ./newfile
	1: foo = ef7503cc
	1: bar = ef7503d2, ef7503d2

Notice that in the run of newfile, bar points at "world" instead of
the perhaps expected "orld".

Dynodump supports sparc, intel, and power pc architectures.

Dynodump is buildable with GNU make and gcc.  If it works for you
with these tools, let me know.

unexec()
--------

Also supplied is an unexsol2.c which belongs in the XEmacs src
directory with the other unex*.c files.  The src/s/sol2.h should be
have the following added to it:

	#ifdef UNEXEC
	#undef UNEXEC
	#endif
	#define UNEXEC unexsol2.o

This unexec() will attempt to dlopen() the dynodump.so to find the
dynodump function.  The default is "../dynodump/dynodump.so" which
should be appropriate for the typical XEmacs build (unless you used
configure to set up separate build and source trees).  You may change
it by setting the DYNODUMP environment variable to the full path of
the dynodump shared object.

Other notes:

If you're using the 4.0 C compiler from Sun, you should arrange to
shut off the incremental linker.  This is done with the -xildoff
option to the compiler.  The common.mk and testsuite/Makefile files
have commentary and ready made lines for uncommenting for this
purpose.

If you're interested in playing with the UltraSPARC specific options
to the aforementioned compiler, then the same makefiles have some
commentary and flags settings to point you in that direction.

Questions:

dynodump() was developed by Rod.Evans@Eng.Sun.COM and
Georg.Nikodym@Canada.Sun.COM.  If you have questions, feel free to ask
them but be aware that Rod, "don't know jack about emacs."