| [ < ] | [ > ] | [ << ] | [ Up ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |
This chapter describes how the runnable XEmacs executable is dumped with the preloaded Lisp libraries in it and how storage is allocated.
There is an entire separate document, the XEmacs Internals Manual, devoted to the internals of XEmacs from the perspective of the C programmer. It contains much more detailed information about the build process, the allocation and garbage-collection process, and other aspects related to the internals of XEmacs.
| B.1 Building XEmacs | How to preload Lisp libraries into XEmacs. | |
| B.2 Garbage Collection | Reclaiming space for Lisp objects no longer used. |
| [ < ] | [ > ] | [ << ] | [ Up ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |
This section explains the steps involved in building the XEmacs executable. You don't have to know this material to build and install XEmacs, since the makefiles do all these things automatically. This information is pertinent to XEmacs maintenance.
The XEmacs Internals Manual contains more information about this.
Compilation of the C source files in the `src' directory produces an executable file called `temacs'. It contains the XEmacs Lisp interpreter and I/O routines, but not the editing commands.
Before XEmacs is actually usable, a number of Lisp files need to be loaded. These define all the editing commands, plus most of the startup code and many very basic Lisp primitives. This is accomplished by loading the file `loadup.el', which in turn loads all of the other standardly-loaded Lisp files.
It takes a substantial time to load the standard Lisp files. Luckily, you don't have to do this each time you run XEmacs; `temacs' can dump out an executable program called `xemacs' that has these files preloaded. `xemacs' starts more quickly because it does not need to load the files. This is the XEmacs executable that is normally installed.
To create `xemacs', use the command `temacs -batch -l loadup dump'. The purpose of `-batch' here is to tell `temacs' to run in non-interactive, command-line mode. (`temacs' can only run in this fashion. Part of the code required to initialize frames and faces is in Lisp, and must be loaded before XEmacs is able to create any frames.) The argument `dump' tells `loadup.el' to dump a new executable named `xemacs'.
The dumping process is highly system-specific, and some operating systems don't support dumping. On those systems, you must start XEmacs with the `temacs -batch -l loadup run-temacs' command each time you use it. This takes a substantial time, but since you need to start Emacs once a day at most--or once a week if you never log out--the extra time is not too severe a problem. (In older versions of Emacs, you started Emacs from `temacs' using `temacs -l loadup'.)
You are free to start XEmacs directly from `temacs' if you want, even if there is already a dumped `xemacs'. Normally you wouldn't want to do that; but the Makefiles do this when you rebuild XEmacs using `make all-elc', which builds XEmacs and simultaneously compiles any out-of-date Lisp files. (You need `xemacs' in order to compile Lisp files. However, you also need the compiled Lisp files in order to dump out `xemacs'. If both of these are missing or corrupted, you are out of luck unless you're able to bootstrap `xemacs' from `temacs'. Note that `make all-elc' actually loads the alternative loadup file `loadup-el.el', which works like `loadup.el' but forces XEmacs to ignore any compiled Lisp files even if they exist.)
You can specify additional files to preload by writing a library named `site-load.el' that loads them. However, the advantage of preloading additional files decreases as machines get faster. On modern machines, it is often not advisable, especially if the Lisp code is on a file system local to the machine running XEmacs.
You can specify other Lisp expressions to execute just before dumping by putting them in a library named `site-init.el'. However, if they might alter the behavior that users expect from an ordinary unmodified XEmacs, it is better to put them in `default.el', so that users can override them if they wish. See section 57.1.1 Summary: Sequence of Actions at Start Up.
Before `loadup.el' dumps the new executable, it finds the
documentation strings for primitive and preloaded functions (and
variables) in the file where they are stored, by calling
Snarf-documentation (see section 34.2 Access to Documentation Strings). These
strings were moved out of the `xemacs' executable to make it
smaller. See section 34.1 Documentation Basics.
If you use this function in an XEmacs that was already dumped, you must
set command-line-processed to nil first for good results.
See section 57.1.4 Command Line Arguments.
When called interactively with a prefix argument, insert string at point.
Don't use this function in programs to choose actions according
to the system configuration; look at system-configuration instead.
(emacs-version)
=> "XEmacs 20.1 [Lucid] (i586-unknown-linux2.0.29)
of Mon Apr 7 1997 on altair.xemacs.org"
|
Called interactively, the function prints the same information in the echo area.
emacs-build-time "Mon Apr 7 20:28:52 1997"
=>
|
"20.1 XEmacs Lucid".
The following two variables did not exist before FSF GNU Emacs version 19.23 and XEmacs version 19.10, which reduces their usefulness at present, but we hope they will be convenient in the future.
| [ < ] | [ > ] | [ << ] | [ Up ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |
When a program creates a list or the user defines a new function (such as by loading a library), that data is placed in normal storage. If normal storage runs low, then XEmacs asks the operating system to allocate more memory in blocks of 2k bytes. Each block is used for one type of Lisp object, so symbols, cons cells, markers, etc., are segregated in distinct blocks in memory. (Vectors, long strings, buffers and certain other editing types, which are fairly large, are allocated in individual blocks, one per object, while small strings are packed into blocks of 8k bytes. [More correctly, a string is allocated in two sections: a fixed size chunk containing the length, list of extents, etc.; and a chunk containing the actual characters in the string. It is this latter chunk that is either allocated individually or packed into 8k blocks. The fixed size chunk is packed into 2k blocks, as for conses, markers, etc.])
It is quite common to use some storage for a while, then release it by (for example) killing a buffer or deleting the last pointer to an object. XEmacs provides a garbage collector to reclaim this abandoned storage. (This name is traditional, but "garbage recycler" might be a more intuitive metaphor for this facility.)
The garbage collector operates by finding and marking all Lisp objects that are still accessible to Lisp programs. To begin with, it assumes all the symbols, their values and associated function definitions, and any data presently on the stack, are accessible. Any objects that can be reached indirectly through other accessible objects are also accessible.
When marking is finished, all objects still unmarked are garbage. No matter what the Lisp program or the user does, it is impossible to refer to them, since there is no longer a way to reach them. Their space might as well be reused, since no one will miss them. The second ("sweep") phase of the garbage collector arranges to reuse them.
The sweep phase puts unused cons cells onto a free list for
future allocation; likewise for symbols, markers, extents, events,
floats, compiled-function objects, and the fixed-size portion of
strings. It compacts the accessible small string-chars chunks so they
occupy fewer 8k blocks; then it frees the other 8k blocks. Vectors,
buffers, windows, and other large objects are individually allocated and
freed using malloc and free.
Common Lisp note: unlike other Lisps, XEmacs Lisp does not call the garbage collector when the free list is empty. Instead, it simply requests the operating system to allocate more storage, and processing continues untilgc-cons-thresholdbytes have been used.This means that you can make sure that the garbage collector will not run during a certain portion of a Lisp program by calling the garbage collector explicitly just before it (provided that portion of the program does not use so much space as to force a second garbage collection).
gc-cons-threshold bytes of
Lisp data since the previous garbage collection.)
garbage-collect returns a list containing the following
information:
((used-conses . free-conses) (used-syms . free-syms) (used-markers . free-markers) used-string-chars used-vector-slots (plist)) => ((73362 . 8325) (13718 . 164) (5089 . 5098) 949121 118677 (conses-used 73362 conses-free 8329 cons-storage 658168 symbols-used 13718 symbols-free 164 symbol-storage 335216 bit-vectors-used 0 bit-vectors-total-length 0 bit-vector-storage 0 vectors-used 7882 vectors-total-length 118677 vector-storage 537764 compiled-functions-used 1336 compiled-functions-free 37 compiled-function-storage 44440 short-strings-used 28829 long-strings-used 2 strings-free 7722 short-strings-total-length 916657 short-string-storage 1179648 long-strings-total-length 32464 string-header-storage 441504 floats-used 3 floats-free 43 float-storage 2044 markers-used 5089 markers-free 5098 marker-storage 245280 events-used 103 events-free 835 event-storage 110656 extents-used 10519 extents-free 2718 extent-storage 372736 extent-auxiliarys-used 111 extent-auxiliarys-freed 3 extent-auxiliary-storage 4440 window-configurations-used 39 window-configurations-on-free-list 5 window-configurations-freed 10 window-configuration-storage 9492 popup-datas-used 3 popup-data-storage 72 toolbar-buttons-used 62 toolbar-button-storage 4960 toolbar-datas-used 12 toolbar-data-storage 240 symbol-value-buffer-locals-used 182 symbol-value-buffer-local-storage 5824 symbol-value-lisp-magics-used 22 symbol-value-lisp-magic-storage 1496 symbol-value-varaliases-used 43 symbol-value-varalias-storage 1032 opaque-lists-used 2 opaque-list-storage 48 color-instances-used 12 color-instance-storage 288 font-instances-used 5 font-instance-storage 180 opaques-used 11 opaque-storage 312 range-tables-used 1 range-table-storage 16 faces-used 34 face-storage 2584 glyphs-used 124 glyph-storage 4464 specifiers-used 775 specifier-storage 43869 weak-lists-used 786 weak-list-storage 18864 char-tables-used 40 char-table-storage 41920 buffers-used 25 buffer-storage 7000 extent-infos-used 457 extent-infos-freed 73 extent-info-storage 9140 keymaps-used 275 keymap-storage 12100 consoles-used 4 console-storage 384 command-builders-used 2 command-builder-storage 120 devices-used 2 device-storage 344 frames-used 3 frame-storage 624 image-instances-used 47 image-instance-storage 3008 windows-used 27 windows-freed 2 window-storage 9180 lcrecord-lists-used 15 lcrecord-list-storage 360 hash-tables-used 631 hash-table-storage 25240 streams-used 1 streams-on-free-list 3 streams-freed 12 stream-storage 91)) |
Here is a table explaining each element:
The initial threshold value is 500,000. If you specify a larger value, garbage collection will happen less often. This reduces the amount of time spent garbage collecting, but increases total memory use. You may want to do this when running a program that creates lots of Lisp data.
You can make collections more frequent by specifying a smaller value,
down to 10,000. A value less than 10,000 will remain in effect only
until the subsequent garbage collection, at which time
garbage-collect will set the threshold back to 10,000. (This does
not apply if XEmacs was configured with `--debug'. Therefore, be
careful when setting gc-cons-threshold in that case!)
gc-pointer-glyph specifies a value (i.e. a
pointer image instance) in the domain of the selected frame, the mouse
cursor will change instead of this message being printed.
gc-message. See section 50. Glyphs.
If XEmacs was configured with `--debug', you can set the following two variables to get direct information about all the allocation that is happening in a segment of Lisp code.
debug-allocation.
| [ << ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |