foundation-module: Chapter 1: Nowebify.
This commit is contained in:
parent
4b2c8bd3ff
commit
b3ba94b5a6
2 changed files with 117 additions and 88 deletions
|
|
@ -2,7 +2,7 @@
|
|||
|
||||
Starting up and shutting down.
|
||||
|
||||
@h Introduction.
|
||||
@ \section{Introduction.}
|
||||
The Foundation module supplies some of the conveniences of more modern
|
||||
programming languages to ANSI C. It offers the usual stuff of standard
|
||||
libraries everywhere: memory management, collection classes, filename
|
||||
|
|
@ -16,27 +16,27 @@ language's compiler and outlying tools, including Inweb itself. If you
|
|||
need to write a command-line utility in ANSI C with no dependencies on
|
||||
other tools or libraries to speak of, you could do worse.
|
||||
|
||||
To use |foundation|, the Contents section of a web should include:
|
||||
= (text)
|
||||
To use [[foundation]], the Contents section of a web should include:
|
||||
|
||||
Import: foundation
|
||||
=
|
||||
|
||||
before beginning the chapter rundown. There are then a few conventions
|
||||
which must be followed. The |main| routine for the client should, as one
|
||||
of its very first acts, call |Foundation::start()|, and should similarly, just
|
||||
before it exits, call |Foundation::end()|. Any other module used should be
|
||||
which must be followed. The [[main]] routine for the client should, as one
|
||||
of its very first acts, call [[Foundation::start()]], and should similarly, just
|
||||
before it exits, call [[Foundation::end()]]. Any other module used should be
|
||||
started after Foundation starts, and ended before Foundation ends.
|
||||
|
||||
In addition, the client's source code needs to define a few symbols to indicate
|
||||
what it needs in the way of memory allocation. For an example, see the code
|
||||
for Inweb itself.
|
||||
|
||||
@h Basic definitions.
|
||||
@ \section{Basic definitions.}
|
||||
These are all from the ANSI C standard library (or the pthread POSIX standard),
|
||||
which means that Inweb will tangle them up to the top of the C source code.
|
||||
Because pthread is not normally available on Windows, a special header is
|
||||
supplied instead for that case.
|
||||
|
||||
= (very early code)
|
||||
<<*>>=
|
||||
#include <ctype.h>
|
||||
#include <stdarg.h>
|
||||
#include <stdint.h>
|
||||
|
|
@ -47,65 +47,68 @@ supplied instead for that case.
|
|||
#include <math.h>
|
||||
#include <wchar.h>
|
||||
|
||||
@ =
|
||||
text_stream *DL = NULL; /* Current destination of debugging text: kept |NULL| until opened */
|
||||
<<*>>=
|
||||
text_stream *DL = NULL; /* Current destination of debugging text: kept [[NULL]] until opened */
|
||||
|
||||
@ We'll use three truth states, the third of which can also mean "unknown".
|
||||
|
||||
@d TRUE 1
|
||||
@d FALSE 0
|
||||
@d NOT_APPLICABLE 2
|
||||
<<*>>=
|
||||
#define TRUE 1
|
||||
#define FALSE 0
|
||||
#define NOT_APPLICABLE 2
|
||||
|
||||
@ And we recognise two different encodings for narrow (i.e., |char *|) C strings.
|
||||
@ And we recognise two different encodings for narrow (i.e., [[char *]]) C strings.
|
||||
|
||||
@d UTF8_ENC 1 /* Write as UTF-8 without BOM */
|
||||
@d ISO_ENC 2 /* Write as ISO Latin-1 (i.e., no conversion needed) */
|
||||
<<*>>=
|
||||
#define UTF8_ENC 1 /* Write as UTF-8 without BOM */
|
||||
#define ISO_ENC 2 /* Write as ISO Latin-1 (i.e., no conversion needed) */
|
||||
|
||||
@ It is assumed that our host filing system can manage at least 30-character
|
||||
filenames, that space is legal as a character in a filename, and that trailing
|
||||
extensions can be longer than 3 characters (in particular, that |.html| is
|
||||
allowed). There are no clear rules but on Windows |MAX_PATH| can be as low as
|
||||
extensions can be longer than 3 characters (in particular, that [[.html]] is
|
||||
allowed). There are no clear rules but on Windows [[MAX_PATH]] can be as low as
|
||||
260, and on Mac OS X the equivalent limit is 1024; both systems can house
|
||||
files buried more deeply, but in both cases the user interface to the
|
||||
operating system fails to recognise them. Some Linux implementations raise the
|
||||
equivalent |PATH_MAX| limit as high as 4096. This seems a reasonable
|
||||
equivalent [[PATH_MAX]] limit as high as 4096. This seems a reasonable
|
||||
compromise in practice:
|
||||
|
||||
@d MAX_FILENAME_LENGTH 1025
|
||||
<<*>>=
|
||||
#define MAX_FILENAME_LENGTH 1025
|
||||
|
||||
@ Very occasionally we'll store a pointer as data:
|
||||
|
||||
=
|
||||
<<*>>=
|
||||
typedef long int pointer_sized_int;
|
||||
|
||||
@h The beginning and the end.
|
||||
@ \section{The beginning and the end.}
|
||||
As noted above, the client needs to call these when starting up and when
|
||||
shutting down.
|
||||
|
||||
The Inweb notation |[[textliterals]]| inserts declarations of I-literals,
|
||||
that is, literal |text_stream *| values written as |I"strings"|. It should
|
||||
The Inweb notation [[[[textliterals]]]] inserts declarations of I-literals,
|
||||
that is, literal [[text_stream *]] values written as [[I"strings"]]. It should
|
||||
never be used anywhere but here.
|
||||
|
||||
=
|
||||
<<*>>=
|
||||
void Foundation::start(int argc, char **argv) {
|
||||
CommandLine::set_locale(argc, argv);
|
||||
Platform::configure_terminal();
|
||||
Memory::start();
|
||||
@<Register the default stream writers@>;
|
||||
<<Register the default stream writers>>;
|
||||
[[textliterals]];
|
||||
Time::begin();
|
||||
Pathnames::start();
|
||||
@<Register the default debugging log aspects@>;
|
||||
@<Register the default debugging log writers@>;
|
||||
@<Register the default command line switches@>;
|
||||
<<Register the default debugging log aspects>>;
|
||||
<<Register the default debugging log writers>>;
|
||||
<<Register the default command line switches>>;
|
||||
}
|
||||
|
||||
@ After calling |Foundation::start()|, the client can register further stream
|
||||
@ After calling [[Foundation::start()]], the client can register further stream
|
||||
writing routines, following these models: they define the meaning of escape
|
||||
characters in |WRITE|, our version of formatted printing. |%f|, for example,
|
||||
prints a filename by calling |Filenames::writer|.
|
||||
characters in [[WRITE]], our version of formatted printing. [[%f]], for example,
|
||||
prints a filename by calling [[Filenames::writer]].
|
||||
|
||||
@<Register the default stream writers@> =
|
||||
<<Register the default stream writers>>=
|
||||
Writers::register_writer('f', &Filenames::writer);
|
||||
Writers::register_writer('p', &Pathnames::writer);
|
||||
Writers::register_writer('v', &VersionNumbers::writer);
|
||||
|
|
@ -116,38 +119,39 @@ switched on or off. Each aspect represents an activity of the program about
|
|||
which a narrative is printed, or not printed, to the debugging log file.
|
||||
The following are always provided, but are all off by default.
|
||||
|
||||
@<Register the default debugging log aspects@> =
|
||||
<<Register the default debugging log aspects>>=
|
||||
Log::declare_aspect(DEBUGGING_LOG_INCLUSIONS_DA, L"debugging log inclusions", FALSE, FALSE);
|
||||
Log::declare_aspect(SHELL_USAGE_DA, L"shell usage", FALSE, FALSE);
|
||||
Log::declare_aspect(MEMORY_USAGE_DA, L"memory usage", FALSE, FALSE);
|
||||
Log::declare_aspect(TEXT_FILES_DA, L"text files", FALSE, FALSE);
|
||||
|
||||
@ Debugging log writers are similar to stream writers, but implement the |$|
|
||||
escapes only available to the debugging log. For example, |$S| calls the
|
||||
|Streams::log| function to print a textual representation of the current
|
||||
@ Debugging log writers are similar to stream writers, but implement the [[$]]
|
||||
escapes only available to the debugging log. For example, [[$S]] calls the
|
||||
[[Streams::log]] function to print a textual representation of the current
|
||||
state of a stream.
|
||||
|
||||
@<Register the default debugging log writers@> =
|
||||
<<Register the default debugging log writers>>=
|
||||
Writers::register_logger('a', &Tries::log_avinue);
|
||||
Writers::register_logger('S', &Streams::log);
|
||||
|
||||
@ We provide an optional service for parsing the command line. By default,
|
||||
the |-log A| switch makes that aspect active, though it's hyphenated, so
|
||||
for example |-log memory-usage| or |-log no-memory-usage|. |-fixtime| is
|
||||
the [[-log A]] switch makes that aspect active, though it's hyphenated, so
|
||||
for example [[-log memory-usage]] or [[-log no-memory-usage]]. [[-fixtime]] is
|
||||
used to ease automated testing: we don't want to reject the output from
|
||||
some tool just because it contains today's date and not the date when the
|
||||
test was set up. |-crash| tells the tool to crash on a fatal error, rather
|
||||
test was set up. [[-crash]] tells the tool to crash on a fatal error, rather
|
||||
than to exit cleanly, to make it easier to diagnose in a debugger.
|
||||
|
||||
@e LOG_CLSW from 0
|
||||
@e VERSION_CLSW
|
||||
@e CRASH_CLSW
|
||||
@e HELP_CLSW
|
||||
@e FIXTIME_CLSW
|
||||
@e AT_CLSW
|
||||
@e LOCALE_CLSW
|
||||
<<*>>=
|
||||
enum LOG_CLSW from 0
|
||||
enum VERSION_CLSW
|
||||
enum CRASH_CLSW
|
||||
enum HELP_CLSW
|
||||
enum FIXTIME_CLSW
|
||||
enum AT_CLSW
|
||||
enum LOCALE_CLSW
|
||||
|
||||
@<Register the default command line switches@> =
|
||||
<<Register the default command line switches>>=
|
||||
CommandLine::begin_group(FOUNDATION_CLSG, NULL);
|
||||
CommandLine::declare_switch(LOG_CLSW, L"log", 2,
|
||||
L"write the debugging log to include diagnostics on X");
|
||||
|
|
@ -166,11 +170,11 @@ than to exit cleanly, to make it easier to diagnose in a debugger.
|
|||
CommandLine::end_group();
|
||||
|
||||
@ Once the following has been called, it is not safe to use any of the
|
||||
|foundation| facilities. It should be called on any normal exit, but not on
|
||||
[[foundation]] facilities. It should be called on any normal exit, but not on
|
||||
an early termination due to a fatal error, as this may lead to thread
|
||||
safety problems.
|
||||
|
||||
=
|
||||
<<*>>=
|
||||
void Foundation::end(void) {
|
||||
if (Log::aspect_switched_on(MEMORY_USAGE_DA)) Memory::log_statistics();
|
||||
Log::close();
|
||||
|
|
|
|||
|
|
@ -13,7 +13,7 @@ purposes we need here, it's simplest to divide all operating systems into
|
|||
two groups: the POSIX group, and Windows.
|
||||
|
||||
This Foundation module therefore comes with two variant versions of the
|
||||
|Platform::| section of code. The one you're reading compiles on a POSIX
|
||||
[[Platform::]] section of code. The one you're reading compiles on a POSIX
|
||||
operating system, and the other one on Windows.
|
||||
|
||||
@ Some basics that apply to all POSIX-supporting systems.
|
||||
|
|
@ -38,6 +38,8 @@ operating system, and the other one on Windows.
|
|||
#define SHELL_QUOTE_CHARACTER '\''
|
||||
#define INFORM_FOLDER_RELATIVE_TO_HOME "Library"
|
||||
|
||||
#endif
|
||||
|
||||
@ \section{Generic Unix.}
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_UNIX
|
||||
|
|
@ -47,12 +49,13 @@ Adam Thornton, and for Ubuntu, Fedora, Debian and so forth, by Philip
|
|||
Chimento) and also for Solaris variants: they can probably be used for any
|
||||
Unix-based system.
|
||||
|
||||
<<*>>=
|
||||
#define PLATFORM_STRING "unix"
|
||||
#define INFORM_FOLDER_RELATIVE_TO_HOME ""
|
||||
|
||||
<<*>>=
|
||||
#include <strings.h>
|
||||
|
||||
#endif
|
||||
|
||||
@ \section{Linux.}
|
||||
These settings are used both for the Linux versions (both command-line, by
|
||||
Adam Thornton, and for Ubuntu, Fedora, Debian and so forth, by Philip
|
||||
|
|
@ -63,6 +66,7 @@ Unix-based system.
|
|||
#ifdef PLATFORM_LINUX
|
||||
#define PLATFORM_STRING "linux"
|
||||
#define INFORM_FOLDER_RELATIVE_TO_HOME ""
|
||||
#endif
|
||||
|
||||
<<*>>=
|
||||
#include <strings.h>
|
||||
|
|
@ -75,6 +79,7 @@ These settings are used for Nathan Summers's Android versions.
|
|||
#define PLATFORM_STRING "android"
|
||||
#define SUPPRESS_MAIN
|
||||
#define INFORM_FOLDER_RELATIVE_TO_HOME ""
|
||||
#endif
|
||||
|
||||
<<*>>=
|
||||
#include <strings.h>
|
||||
|
|
@ -83,7 +88,7 @@ These settings are used for Nathan Summers's Android versions.
|
|||
When using a Unix-like system such as Cygwin on Windows, it's inevitable that
|
||||
paths will sometimes contain backslashes and sometimes forward slashes, meaning
|
||||
a folder (i.e. directory) divide in either case. So:
|
||||
(a) When writing such a divider, always write |FOLDER_SEPARATOR|, a backslash;
|
||||
(a) When writing such a divider, always write [[FOLDER_SEPARATOR]], a backslash;
|
||||
(b) When testing for such a divider, call the following.
|
||||
|
||||
<<*>>=
|
||||
|
|
@ -121,21 +126,22 @@ always be unavailable: that doesn't mean we can't run on those platforms,
|
|||
just that installation and use of Foundation-built tools is less convenient.)
|
||||
|
||||
<<*>>=
|
||||
ifdef PLATFORM_LINUX
|
||||
#ifdef PLATFORM_LINUX
|
||||
void Platform::where_am_i(wchar_t *p, size_t length) {
|
||||
char buffer[PATH_MAX + 1];
|
||||
@<Follow the proc filesystem symlink to the real filesystem's file@>;
|
||||
@<Transcode buffer, which is locale-encoded, into the wide-char buffer@>;
|
||||
<<Follow the proc filesystem symlink to the real filesystem's file>>;
|
||||
<<Transcode buffer, which is locale-encoded, into the wide-char buffer>>;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ On Linux, |/proc/self/exe| is a symlink to the current process's executable.
|
||||
@ On Linux, [[/proc/self/exe]] is a symlink to the current process's executable.
|
||||
Follow that link to find the path. Normally when reading a symlink, one uses
|
||||
|lstat()| to find the path length instead of guessing |PATH_MAX|, but the
|
||||
symlinks in |/proc| are special and don't provide a length to |lstat()|.
|
||||
|lstat()| to find the path length instead of guessing [[PATH_MAX]], but the
|
||||
symlinks in [[/proc]] are special and don't provide a length to [[lstat()]].
|
||||
|
||||
<<Follow the proc filesystem symlink to the real filesystem's file>>=
|
||||
ssize_t link_len = readlink("/proc/self/exe", buffer, PATH_MAX);
|
||||
if (link_len < 0) @<Fail@>; // unable to find
|
||||
if (link_len < 0) <<Fail>>; // unable to find
|
||||
buffer[link_len] = '\0';
|
||||
|
||||
@ Next, convert the obtained buffer (which is a string in the local filename
|
||||
|
|
@ -144,7 +150,7 @@ string.
|
|||
|
||||
<<Transcode buffer, which is locale-encoded, into the wide-char buffer>>=
|
||||
size_t convert_len = mbstowcs(p, buffer, length);
|
||||
if (convert_len == (size_t)-1) @<Fail@>; // wouldn't fit
|
||||
if (convert_len == (size_t)-1) <<Fail>>; // wouldn't fit
|
||||
|
||||
@ And now the Mac version:
|
||||
|
||||
|
|
@ -160,30 +166,35 @@ void Platform::where_am_i(wchar_t *p, size_t length) {
|
|||
uint32_t tempsize = pathsize;
|
||||
|
||||
/* Get "a path" to the executable */
|
||||
if (_NSGetExecutablePath(relative_path, &tempsize) != 0) @<Fail@>;
|
||||
if (_NSGetExecutablePath(relative_path, &tempsize) != 0) <<Fail>>;
|
||||
|
||||
/* Convert to canonical absolute path */
|
||||
if (realpath(relative_path, absolute_path) == NULL) @<Fail@>;
|
||||
if (realpath(relative_path, absolute_path) == NULL) <<Fail>>;
|
||||
|
||||
/* Next, convert the obtained buffer (which is a string in the local
|
||||
* filename encoding, possibly multibyte) to a wide-char string. */
|
||||
convert_len = mbstowcs(p, absolute_path, length);
|
||||
if (convert_len == (size_t)-1) @<Fail@>;
|
||||
if (convert_len == (size_t)-1) <<Fail>>;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ For Unix, there's nothing we can generically do. #ifdef PLATFORM_UNIX"
|
||||
@ For Unix, there's nothing we can generically do.
|
||||
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_UNIX
|
||||
void Platform::where_am_i(wchar_t *p, size_t length) {
|
||||
@<Fail@>;
|
||||
<<Fail>>;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ On Android, there's no real need for this. #ifdef PLATFORM_ANDROID"
|
||||
@ On Android, there's no real need for this.
|
||||
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_ANDROID
|
||||
void Platform::where_am_i(wchar_t *p, size_t length) {
|
||||
@<Fail@>;
|
||||
<<Fail>>;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ All of the above make use of:
|
||||
|
||||
|
|
@ -198,26 +209,27 @@ void Platform::where_am_i(wchar_t *p, size_t length) {
|
|||
int Platform::system(const char *cmd) {
|
||||
return system(cmd);
|
||||
}
|
||||
#endif
|
||||
|
||||
@
|
||||
In MacOS 10.5, a new implementation of the C standard library
|
||||
crippled performance of |system()| by placing it behind a global mutex, so
|
||||
crippled performance of [[system()]] by placing it behind a global mutex, so
|
||||
that it was impossible for two cores to be calling the function at the same
|
||||
time. The net effect of this is that the Inform test suite, executing in
|
||||
Intest, ran in 1/16th speed. This issue didn't come to light until 2019,
|
||||
however, because the build setting |-mmacosx-version-min=10.4| turned out
|
||||
to force use of the (perfectly good) pre-10.5 library, where |system()|
|
||||
however, because the build setting [[-mmacosx-version-min=10.4]] turned out
|
||||
to force use of the (perfectly good) pre-10.5 library, where [[system()]]
|
||||
continued to run in a multi-threaded way, just as it does on Linux and
|
||||
most all other Unixes. The old library was eventually withdrawn by Apple
|
||||
in 2018, and in any case would stop working at some point in 2019-20 due
|
||||
to the final removal of 32-bit binary support from MacOS.
|
||||
|
||||
It took several days to find a pthread-safe way to reimplement |system()|.
|
||||
The obvious way, using |fork()| and then running |execve()| on the child
|
||||
process -- essentially the standard way to implement |system()|, if you forget
|
||||
It took several days to find a pthread-safe way to reimplement [[system()]].
|
||||
The obvious way, using [[fork()]] and then running [[execve()]] on the child
|
||||
process -- essentially the standard way to implement [[system()]], if you forget
|
||||
about signal-handling -- led to obscure and unrepeatable memory corruption
|
||||
bugs in Intest, with the worker threads apparently writing on each other's
|
||||
memory space. Using |posix_spawn()| instead appears to work better.
|
||||
memory space. Using [[posix_spawn()]] instead appears to work better.
|
||||
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_MACOS
|
||||
|
|
@ -239,6 +251,7 @@ int Platform::system(const char *cmd) {
|
|||
}
|
||||
return -1;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ \section{ Directory handling.}
|
||||
|
||||
|
|
@ -276,8 +289,8 @@ void Platform::closedir(void *D) {
|
|||
closedir(dirp);
|
||||
}
|
||||
|
||||
@ \section{ Timestamp and file size.
|
||||
There are implementations of the C standard library where |time_t| has
|
||||
@ \section{Timestamp and file size.}
|
||||
There are implementations of the C standard library where [[time_t]] has
|
||||
super-weird behaviour, but on almost all POSIX systems, time 0 corresponds to
|
||||
midnight on 1 January 1970. All we really need is that the "never" value
|
||||
is one which is earlier than any possible timestamp on the files we'll
|
||||
|
|
@ -349,7 +362,7 @@ to survive, given the MacOS team's current hostility to scripting; we're
|
|||
actually running a one-line AppleScript here.
|
||||
|
||||
<<*>>=
|
||||
#ifdef-PLATFORM_MACOS
|
||||
#ifdef PLATFORM_MACOS
|
||||
void Platform::notification(text_stream *text, int happy) {
|
||||
char *sound_name = "Bell.aiff";
|
||||
if (happy == FALSE) sound_name = "Submarine.aiff";
|
||||
|
|
@ -359,6 +372,7 @@ void Platform::notification(text_stream *text, int happy) {
|
|||
Shell::run(TEMP);
|
||||
DISCARD_TEXT(TEMP)
|
||||
}
|
||||
#endif
|
||||
|
||||
@
|
||||
|
||||
|
|
@ -366,12 +380,13 @@ void Platform::notification(text_stream *text, int happy) {
|
|||
#ifndef PLATFORM_MACOS
|
||||
void Platform::notification(text_stream *text, int happy) {
|
||||
}
|
||||
#endif
|
||||
|
||||
@ \section{Terminal setup.}
|
||||
The idea of this function is that if anything needs to be done to enable the
|
||||
output of ANSI-standard coloured terminal output, then this function has the
|
||||
chance to do it; similarly, it may need to configure itself to receive console
|
||||
output with the correct locale (calling |Locales::get(CONSOLE_LOCALE)| to
|
||||
output with the correct locale (calling [[Locales::get(CONSOLE_LOCALE)]] to
|
||||
find this).
|
||||
|
||||
On POSIX platforms, so far as we know, nothing need be done.
|
||||
|
|
@ -424,23 +439,29 @@ MacOS does not support.
|
|||
<<*>>=
|
||||
#ifdef PLATFORM_LINUX
|
||||
#include <sys/sysinfo.h>
|
||||
#endif
|
||||
|
||||
@ #ifdef PLATFORM_LINUX"
|
||||
<<*>>=
|
||||
@
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_LINUX
|
||||
int Platform::get_core_count(void) {
|
||||
int N = get_nprocs();
|
||||
if (N < 1) return 1;
|
||||
return N;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ #ifdef PLATFORM_MACOS"
|
||||
@
|
||||
While MacOS lacks |sysinfo.h|, it does have |sysctl.h|:
|
||||
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_MACOS
|
||||
#include <sys/sysctl.h>
|
||||
#endif
|
||||
|
||||
@ #ifdef PLATFORM_MACOS"
|
||||
@
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_MACOS
|
||||
int Platform::get_core_count(void) {
|
||||
int N;
|
||||
size_t N_size = sizeof(int);
|
||||
|
|
@ -448,17 +469,21 @@ int Platform::get_core_count(void) {
|
|||
if (N < 1) return 1;
|
||||
return N;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ #ifdef PLATFORM_ANDROID"
|
||||
@
|
||||
For Android it seems prudent simply to ignore multithreading:
|
||||
|
||||
<<*>>=
|
||||
<<*>>=
|
||||
#ifdef PLATFORM_ANDROID
|
||||
int Platform::get_core_count(void) {
|
||||
return 1;
|
||||
}
|
||||
#endif
|
||||
|
||||
@ \section{Mutexes.}
|
||||
|
||||
<<*>>=
|
||||
#define CREATE_MUTEX(name)
|
||||
static pthread_mutex_t name = PTHREAD_MUTEX_INITIALIZER;
|
||||
#define LOCK_MUTEX(name) pthread_mutex_lock(&name);
|
||||
|
|
|
|||
Loading…
Reference in a new issue