cosmopolitan libc

a64l

Converts base64 to 32-bit integer, the posix way.

a_ensure_

abort

Terminates program abnormally.

This function first tries to trigger your SIGABRT handler. If the signal handler returns, then signal(SIGABRT, SIG_DFL) is called before SIGABRT is raised again.

abs

Returns absolute value of 𝑥.

This function is a footgun since your argument may be narrrowed. Consider using labs(), llabs(), or better yet a macro like this:

#define ABS(X) ((X) >= 0 ? (X) : -(X))

Note that passing x as INT_MIN is undefined behavior, which depends on whether or not your c library as well as the objects that call it were built using the -fwrapv or -ftrapv flags.

accept

Creates client socket file descriptor for incoming connection.

On Windows, when this function blocks, there may be a 10 millisecond delay on the handling of signals or thread cancelation.

accept4

Creates client socket file descriptor for incoming connection.

When fd is in O_NONBLOCK mode, this function will raise EAGAIN when no client is available to accept. To wait until a client exists the poll() function may be called using POLLIN.

On Linux, your SO_RCVTIMEO will timeout accept4(). Other OSes (i.e. Windows, MacOS, and BSDs) do not support this and will block forever.

On Windows, when this function blocks, there may be a 10 millisecond delay on the handling of signals or thread cancelation.

access

Checks if effective user can access path in particular ways.

This is equivalent to saying:

faccessat(AT_FDCWD, path, mode, 0);

acos

Returns arc cosine of 𝑥.

acosh

Returns inverse hyperbolic cosine of 𝑥.

acoshf

Returns inverse hyperbolic cosine of 𝑥.

acosl

Returns arc cosine of 𝑥.

AcquireToken

Atomically decrements signed byte index if it's positive.

Multiple threads are able to call this method, to determine if enough tokens exist to perform an operation. Return values greater than zero mean a token was atomically acquired. Values less than, or equal zero means the bucket is empty. There must exist 1 << c signed bytes (or buckets) in the b array.

Since this design uses signed bytes, your returned number may be used to control how much burstiness is allowed. For example:

int t = AcquireToken(tok.b, ip, 22);
if (t < 64) {
  if (t > 8) write(client, "HTTP/1.1 429 \r\n\r\n", 17);
  close(client);
  return;
}

Could be used to send rejections to clients that exceed their tokens, whereas clients who've grossly exceeded their tokens, could simply be dropped.

addmntent

aligned_alloc

Same as memalign(a, n) but requires IS2POW(a).

alphasort

AppendResourceReport

Generates process resource usage report.

appends

Appends string to buffer, e.g.

char *b = 0;
appends(&b, "hello");
free(b);

The resulting buffer is guaranteed to be NUL-terminated, i.e. !b[appendz(b).i] will be the case.

AppendStrList

arm_optimized_routines_notice

asctime

asctime_r

asin

Returns arc sine of 𝑥.

asinhf

Returns inverse hyperbolic sine of 𝑥.

asinl

Returns arc sine of 𝑥.

asprintf

Formats string, allocating needed memory.

__assert_fail

Handles assert() failure.

at_quick_exit

atanh

Returns inverse hyperbolic tangent of 𝑥.

atanhf

Returns inverse hyperbolic tangent of 𝑥.

atanhl

Returns inverse hyperbolic tangent of 𝑥.

atexit

Adds global destructor.

Destructors are called in reverse order. They won't be called if the program aborts or _exit() is called. Invocations of this function are usually generated by the C++ compiler.

atof

Converts string to double.

atoi

Turns string into int.

Decimal is the only radix supported. Leading whitespace (as specified by the isspace() function) is skipped over. Unlike strtol(), the atoi function has undefined behavior on error and it never changes errno

atol

Turns string into long.

Decimal is the only radix supported. Leading whitespace (as specified by the isspace() function) is skipped over. Unlike strtol(), the atoi function has undefined behavior on error and it never changes errno

basename

Returns pointer to last filename component in path, e.g.

path     │ dirname() │ basename()
─────────────────────────────────
0        │ .         │ .
.        │ .         │ .
..       │ .         │ ..
/        │ /         │ /
usr      │ .         │ usr
/usr/    │ /         │ usr
/usr/lib │ /usr      │ lib

Both / and \ are are considered valid component separators on all platforms. Trailing slashes are ignored. We don't grant special consideration to things like foo/., c:/, \\?\Volume, etc.

bcmp

Tests inequality of first 𝑛 bytes of 𝑝 and 𝑞.

bcopy

Moves memory the BSD way.

Please use memmove() instead. Note the order of arguments.

bind

Assigns local address and port number to socket, e.g.

struct sockaddr_in in = {AF_INET, htons(12345), {htonl(0x7f000001)}};
int fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
bind(fd, (struct sockaddr *)&in, sizeof(in));

On Windows, Cosmopolitan's implementation of bind() takes care of always setting the WIN32-specific SO_EXCLUSIVEADDRUSE option on inet stream sockets in order to safeguard your servers from tests

bingblit

bsf

Returns position of first bit set.

                      ctz(𝑥)         31^clz(𝑥)   clz(𝑥)
  uint32 𝑥   bsf(𝑥) tzcnt(𝑥)   ffs(𝑥)   bsr(𝑥) lzcnt(𝑥)
0x00000000      wut       32        0      wut       32
0x00000001        0        0        1        0       31
0x80000001        0        0        1       31        0
0x80000000       31       31       32       31        0
0x00000010        4        4        5        4       27
0x08000010        4        4        5       27        4
0x08000000       27       27       28       27        4
0xffffffff        0        0        1       31        0

bsfl

Returns position of first bit set.

                      ctz(𝑥)         31^clz(𝑥)   clz(𝑥)
  uint32 𝑥   bsf(𝑥) tzcnt(𝑥)   ffs(𝑥)   bsr(𝑥) lzcnt(𝑥)
0x00000000      wut       32        0      wut       32
0x00000001        0        0        1        0       31
0x80000001        0        0        1       31        0
0x80000000       31       31       32       31        0
0x00000010        4        4        5        4       27
0x08000010        4        4        5       27        4
0x08000000       27       27       28       27        4
0xffffffff        0        0        1       31        0

bsr

Returns binary logarithm of 𝑥.

                      ctz(𝑥)         31^clz(𝑥)   clz(𝑥)
  uint32 𝑥   bsf(𝑥) tzcnt(𝑥)   ffs(𝑥)   bsr(𝑥) lzcnt(𝑥)
0x00000000      wut       32        0      wut       32
0x00000001        0        0        1        0       31
0x80000001        0        0        1       31        0
0x80000000       31       31       32       31        0
0x00000010        4        4        5        4       27
0x08000010        4        4        5       27        4
0x08000000       27       27       28       27        4
0xffffffff        0        0        1       31        0

bsrl

Returns binary logarithm of 𝑥.

                      ctz(𝑥)         31^clz(𝑥)   clz(𝑥)
  uint32 𝑥   bsf(𝑥) tzcnt(𝑥)   ffs(𝑥)   bsr(𝑥) lzcnt(𝑥)
0x00000000      wut       32        0      wut       32
0x00000001        0        0        1        0       31
0x80000001        0        0        1       31        0
0x80000000       31       31       32       31        0
0x00000010        4        4        5        4       27
0x08000010        4        4        5       27        4
0x08000000       27       27       28       27        4
0xffffffff        0        0        1       31        0

bswap_16

bswap_32

bswap_64

bulk_free

Frees and clears (sets to NULL) each non-null pointer in given array.

This is twice as fast as freeing them one-by-one. If footers are used, pointers that have been allocated in different mspaces are not freed or cleared, and the count of all such pointers is returned. For large arrays of pointers with poor locality, it may be worthwhile to sort this array before calling bulk_free.

bzero

Sets memory to zero.

bzero n=0                          661 picoseconds
bzero n=1                          661 ps/byte          1,476 mb/s
bzero n=2                          330 ps/byte          2,952 mb/s
bzero n=3                          220 ps/byte          4,428 mb/s
bzero n=4                          165 ps/byte          5,904 mb/s
bzero n=7                           94 ps/byte         10,333 mb/s
bzero n=8                           41 ps/byte         23,618 mb/s
bzero n=15                          44 ps/byte         22,142 mb/s
bzero n=16                          20 ps/byte         47,236 mb/s
bzero n=31                          21 ps/byte         45,760 mb/s
bzero n=32                          20 ps/byte         47,236 mb/s
bzero n=63                          10 ps/byte         92,997 mb/s
bzero n=64                          15 ps/byte         62,982 mb/s
bzero n=127                         15 ps/byte         62,490 mb/s
bzero n=128                         10 ps/byte         94,473 mb/s
bzero n=255                         14 ps/byte         68,439 mb/s
bzero n=256                          9 ps/byte            105 gb/s
bzero n=511                         15 ps/byte         62,859 mb/s
bzero n=512                         11 ps/byte         83,976 mb/s
bzero n=1023                        15 ps/byte         61,636 mb/s
bzero n=1024                        10 ps/byte         88,916 mb/s
bzero n=2047                         9 ps/byte            105 gb/s
bzero n=2048                         8 ps/byte            109 gb/s
bzero n=4095                         8 ps/byte            115 gb/s
bzero n=4096                         8 ps/byte            118 gb/s
bzero n=8191                         7 ps/byte            129 gb/s
bzero n=8192                         7 ps/byte            130 gb/s
bzero n=16383                        6 ps/byte            136 gb/s
bzero n=16384                        6 ps/byte            137 gb/s
bzero n=32767                        6 ps/byte            140 gb/s
bzero n=32768                        6 ps/byte            141 gb/s
bzero n=65535                       15 ps/byte         64,257 mb/s
bzero n=65536                       15 ps/byte         64,279 mb/s
bzero n=131071                      15 ps/byte         63,166 mb/s
bzero n=131072                      15 ps/byte         63,115 mb/s
bzero n=262143                      15 ps/byte         62,052 mb/s
bzero n=262144                      15 ps/byte         62,097 mb/s
bzero n=524287                      15 ps/byte         61,699 mb/s
bzero n=524288                      15 ps/byte         61,674 mb/s
bzero n=1048575                     16 ps/byte         60,179 mb/s
bzero n=1048576                     15 ps/byte         61,330 mb/s
bzero n=2097151                     15 ps/byte         61,071 mb/s
bzero n=2097152                     15 ps/byte         61,065 mb/s
bzero n=4194303                     16 ps/byte         60,942 mb/s
bzero n=4194304                     16 ps/byte         60,947 mb/s
bzero n=8388607                     16 ps/byte         60,872 mb/s
bzero n=8388608                     16 ps/byte         60,879 mb/s

cachestat

Query the page cache statistics of a file.

calloc

Allocates n * itemsize bytes, initialized to zero.

CategorizeIp

Classifies IP address.

cbrt

Returns cube root of 𝑥.

cbrtf

Returns cube root of 𝑥.

cbrtl

Returns cube root of 𝑥.

cfgetispeed

Returns input baud rate.

cfgetospeed

Returns output baud rate.

cfmakeraw

cfsetispeed

Sets input baud rate.

cfsetospeed

Sets output baud rate.

cfsetspeed

Sets input and output baud rate.

chmod

Changes permissions on file, e.g.:

CHECK_NE(-1, chmod("foo/bar.txt", 0644));
CHECK_NE(-1, chmod("o/default/program", 0755));
CHECK_NE(-1, chmod("privatefolder/", 0700));

The esoteric bits generally available on System Five are:

CHECK_NE(-1, chmod("/opt/", 01000));          // sticky bit
CHECK_NE(-1, chmod("/usr/bin/sudo", 04755));  // setuid bit
CHECK_NE(-1, chmod("/usr/bin/wall", 02755));  // setgid bit

chomp

Mutates line to remove line-ending characters.

chomp16

Mutates line to remove line-ending characters.

chown

Changes owner and/or group of pathname.

chromium_notice

chroot

Changes root directory.

Please consider using unveil() instead of chroot(). If you use this system call then consider using both chdir() and closefrom() before calling this function. Otherwise there's a small risk that fchdir() could be used to escape the chroot() environment. Cosmopolitan Libc focuses on static binaries which make chroot() infinitely easier to use since you don't need to construct an entire userspace each time however unveil() is still better to use on modern Linux and OpenBSD because it doesn't require root privileges.

clearerr

Clears eof and error state indicators on stream.

clearerr_unlocked

Clears eof and error state indicators on stream.

__clk_tck

Returns system clock ticks per second.

The returned value is memoized. This function is intended to be used via the CLK_TCK macro wrapper.

The returned value is always greater than zero. It's usually 100 hertz which means each clock tick is 10 milliseconds long.

clock

Returns sum of CPU time consumed by current process since birth.

This function provides a basic idea of how computationally expensive your program is, in terms of both the userspace and kernel processor resources it's hitherto consumed. Here's an example of how you might display this information:

printf("consumed %g seconds of cpu time\n",
       (double)clock() / CLOCKS_PER_SEC);

This function offers at best microsecond accuracy on all supported platforms. Please note the reported values might be a bit chunkier depending on the kernel scheduler sampling interval see CLK_TCK.

clock_getres

Returns granularity of clock.

clock_gettime

Returns nanosecond time.

The clock parameter may bo set to:

• CLOCK_REALTIME returns a wall clock timestamp represented in nanoseconds since the UNIX epoch (~1970). It'll count time in the suspend state. This clock is subject to being smeared by various adjustments made by NTP. These timestamps can have unpredictable discontinuous jumps when clock_settime() is used. Therefore this clock is the default clock for everything, even pthread condition variables. Cosmopoiltan guarantees this clock will never raise EINVAL and also guarantees CLOCK_REALTIME == 0 will always be the case. On Windows this maps to GetSystemTimePreciseAsFileTime(). On platforms with vDSOs like Linux, Windows, and MacOS ARM64 this should take about 20 nanoseconds.
• CLOCK_MONOTONIC returns a timestamp with an unspecified epoch, that should be when the system was powered on. These timestamps shouldn't go backwards. Timestamps shouldn't count time spent in the sleep, suspend, and hibernation states. These timestamps won't be impacted by clock_settime(). These timestamps may be impacted by frequency adjustments made by NTP. Cosmopoiltan guarantees this clock will never raise EINVAL. MacOS and BSDs use the word "uptime" to describe this clock. On Windows this maps to QueryUnbiasedInterruptTimePrecise().
• CLOCK_BOOTTIME is a monotonic clock returning a timestamp with an unspecified epoch, that should be relative to when the host system was powered on. These timestamps shouldn't go backwards. Timestamps should also include time spent in a sleep, suspend, or hibernation state. These timestamps aren't impacted by clock_settime(), but they may be impacted by frequency adjustments made by NTP. This clock will raise an EINVAL error on extremely old Linux distros like RHEL5. MacOS and BSDs use the word "monotonic" to describe this clock. On Windows this maps to QueryInterruptTimePrecise().
• CLOCK_MONOTONIC_RAW returns a timestamp from an unspecified epoch. These timestamps don't count time spent in the sleep, suspend, and hibernation states. This clock is not impacted by clock_settime(). Unlike CLOCK_MONOTONIC this clock is guaranteed to not be impacted by frequency adjustments. Providing this level of assurances may make this clock 10x slower than the monotonic clock. Furthermore this clock may cause EINVAL to be raised if running on a host system that doesn't provide those guarantees, e.g. OpenBSD and MacOS on AMD64.
• CLOCK_REALTIME_COARSE is the same as CLOCK_REALTIME except it'll go faster if the host OS provides a cheaper way to read the wall time. Please be warned that coarse can be really coarse. Rather than nano precision, you're looking at CLK_TCK precision, which can lag as far as 30 milliseconds behind or possibly more. Cosmopolitan may fallback to CLOCK_REALTIME if a faster less accurate clock isn't provided by the system. This clock will raise an EINVAL error on extremely old Linux distros like RHEL5. On platforms with vDSOs like Linux, Windows, and MacOS ARM64 this should take about 5 nanoseconds.
• CLOCK_MONOTONIC_COARSE is the same as CLOCK_MONOTONIC except it'll go faster if the host OS provides a cheaper way to read the unbiased time. Please be warned that coarse can be really coarse. Rather than nano precision, you're looking at CLK_TCK precision, which can lag as far as 30 milliseconds behind or possibly more. Cosmopolitan may fallback to CLOCK_REALTIME if a faster less accurate clock isn't provided by the system. This clock will raise an EINVAL error on extremely old Linux distros like RHEL5. On platforms with vDSOs like Linux, Windows, and MacOS ARM64 this should take about 5 nanoseconds.
• CLOCK_PROCESS_CPUTIME_ID returns the amount of time this process was actively scheduled. This is similar to getrusage() and clock().
• CLOCK_THREAD_CPUTIME_ID returns the amount of time this thread was actively scheduled. This is similar to getrusage() and clock().

clock_nanosleep

Sleeps for particular amount of time.

Here's how you could sleep for one second:

clock_nanosleep(0, 0, &(struct timespec){1}, 0);

Your sleep will be interrupted automatically if you do something like press ctrl-c during the wait. That's an EINTR error and it lets you immediately react to status changes. This is always the case, even if you're using SA_RESTART since this is a @norestart system call.

void OnCtrlC(int sig) {} // EINTR only happens after delivery
signal(SIGINT, OnCtrlC); // do delivery rather than kill proc
printf("save me from sleeping forever by pressing ctrl-c\n");
clock_nanosleep(0, 0, &(struct timespec){INT_MAX}, 0);
printf("you're my hero\n");

If you want to perform an uninterruptible sleep without having to use sigprocmask() to block all signals then this function provides a good solution to that problem. For example:

struct timespec rel, now, abs;
clock_gettime(CLOCK_REALTIME, &now);
rel = timespec_frommillis(100);
abs = timespec_add(now, rel);
while (clock_nanosleep(CLOCK_REALTIME, TIMER_ABSTIME, &abs, 0));

will accurately spin on EINTR errors. That way you're not impeding signal delivery and you're not loosing precision on the wait timeout. This function has first-class support on Linux, FreeBSD, and NetBSD; on OpenBSD it's good; on XNU it's bad; and on Windows it's ugly.

clock_settime

Changes time.

close_range

Closes inclusive range of file descriptors, e.g.

// close all non-stdio file descriptors
if (close_range(3, -1, 0) == -1) {
  for (int i = 3; i < 256; ++i) {
    close(i);
  }
}

The following flags are available:

• CLOSE_RANGE_UNSHARE (Linux-only)
• CLOSE_RANGE_CLOEXEC (Linux-only)

This is only supported on Linux 5.9+ and FreeBSD 13+. Consider using closefrom() which will work on OpenBSD too.

closefrom

Closes extra file descriptors, e.g.

if (closefrom(3))
  for (int i = 3; i < 256; ++i)
    close(i);

CloseSymbolTable

Frees symbol table.

commandv

Resolves full pathname of executable.

commandvenv

Finds full executable path in overridable way.

This is a higher level version of the commandv() function. Programs that spawn subprocesses can use this function to determine the path at startup. Here's an example how you could use it:

if ((strace = commandvenv("STRACE", "strace"))) {
  strace = strdup(strace);
} else {
  fprintf(stderr, "error: please install strace\n");
  exit(1);
}

CompareSlices

CompareSlicesCase

confstr

connect

Connects socket to remote end.

When fd is in O_NONBLOCK mode, this raises EINPROGRESS. To wait for establishment poll() function may be called using POLLOUT. Then SO_ERROR may be used to check for errors.

Connectionless sockets, e.g. UDP, can be connected too. The benefit is not needing to specify the remote address on each send. It also means getsockname() can be called to retrieve routing details.

On Linux, your SO_SNDTIMEO will timeout connect(). Other OSes (i.e. Windows, MacOS, and BSDs) do not support this and will block forever.

On Windows, when this function blocks, there may be a 10 millisecond delay on the handling of signals or thread cancelation.

copy_file_range

Transfers data between files.

If this system call is available (Linux c. 2018 or FreeBSD c. 2021) and the file system supports it (e.g. ext4) and the source and dest files are on the same file system, then this system call shall make copies go about 2x faster.

This implementation requires Linux 5.9+ even though the system call was introduced in Linux 4.5. That's to ensure ENOSYS works reliably due to a faulty backport, that happened in RHEL7. FreeBSD detection on the other hand will work fine.

copysign

Returns 𝑥 with same sign as 𝑦.

copysignf

Returns 𝑥 with same sign as 𝑦.

copysignl

Returns 𝑥 with same sign as 𝑦.

cos

Returns cosine of 𝑥.

cosf

Returns cosine of y.

This is a fast cosf implementation. The worst-case ULP is 0.5607, and the maximum relative error is 0.5303 * 2^-23. A single-step range reduction is used for small values. Large inputs have their range reduced using fast integer arithmetic.

cosh

Returns hyperbolic cosine of 𝑥.

cosh(x) = (exp(x) + 1/exp(x))/2
        = 1 + 0.5*(exp(x)-1)*(exp(x)-1)/exp(x)
        = 1 + x*x/2 + o(x^4)

coshf

Returns hyperbolic cosine of 𝑥.

cosh(x) = (exp(x) + 1/exp(x))/2
        = 1 + 0.5*(exp(x)-1)*(exp(x)-1)/exp(x)
        = 1 + x*x/2 + o(x^4)

cosl

Returns cosine of 𝑥.

cosmo

Cosmopolitan runtime.

cosmo2flock

cosmo_once

Ensures initialization function is called exactly once.

This is the same as pthread_once except that it always uses a tiny spinlock implementation and won't make any system calls. It's needed since this function is an upstream dependency of both pthread_once() and nsync_once(). Most code should favor calling those functions.

cosmoaudio_describe_open_options

cosmoaudio_describe_poll_frames

cosmoaudio_describe_status

_countbits

Returns population count of array.

CountTokens

Returns current number of tokens in bucket.

CPU_AND

CPU_COUNT

CPU_COUNT_S

CPU_EQUAL

__cpu_march

Returns microarchitecture name, e.g.

puts(__cpu_march(__cpu_model.__cpu_subtype));

CPU_OR

CPU_XOR

CPU_ZERO

crc32c

Computes 32-bit Castagnoli Cyclic Redundancy Check.

x^32+x^26+x^23+x^22+x^16+x^12+x^11+x^10+x^8+x^7+x^5+x^4+x^2+x+1
0b00011110110111000110111101000001

crc32init

creat

Creates file.

This is equivalent to saying:

int fd = openat(AT_FDCWD, file, O_CREAT | O_WRONLY | O_TRUNC, mode);

critbit0_allprefixed

Invokes callback for all items with prefix.

critbit0_clear

Removes all items from 𝑡.

critbit0_contains

Returns non-zero iff 𝑢 ∈ 𝑡.

critbit0_delete

Removes 𝑢 from 𝑡.

critbit0_emplace

Inserts 𝑢 into 𝑡 without copying.

critbit0_get

Returns first item in 𝑡 with prefix 𝑢.

critbit0_insert

Inserts 𝑢 into 𝑡.

crypt

Encrypts password the old fashioned way.

The method of encryption depends on the first three chars of salt:

• $1$ is MD5
• $2$ is Blowfish
• $5$ is SHA-256
• $6$ is SHA-512
• Otherwise DES

crypt_r

Encrypts password the old fashioned way.

The method of encryption depends on the first three chars of salt:

• $1$ is MD5
• $2$ is Blowfish
• $5$ is SHA-256
• $6$ is SHA-512
• Otherwise DES

ctermid

Generates path of controlling terminal.

This function always returns /dev/tty since that's supported by all supported platforms, and polyfilled on Windows.

ctime

Represents time as string.

ctime_r

__cxa_atexit

Adds global destructor.

Destructors are called in reverse order. They won't be called if the program aborts or _exit() is called. Invocations of this function are usually generated by the C++ compiler. Behavior is limitless if some other module has linked calloc().

__cxa_finalize

Triggers global destructors.

They're called in LIFO order. If a destructor adds more destructors, then those destructors will be called immediately following, before iteration continues.

__cxa_printexits

Prints global destructors.

_Cz_adler32

Updates running Adler-32 checksum with the bytes buf[0..len-1] and return the updated checksum. If buf is Z_NULL, this function returns the required initial value for the checksum.

_Cz_adler32_combine

Combine two Adler-32 checksums into one. For two sequences of bytes, seq1 and seq2 with lengths len1 and len2, Adler-32 checksums were calculated for each, adler1 and adler2. adler32_combine() returns the Adler-32 checksum of seq1 and seq2 concatenated, requiring only adler1, adler2, and len2. Note that the off_t type (like off_t) is a signed integer. If len2 is negative, the result has no meaning or utility.

_Cz_adler32_z

Same as adler32(), but with a size_t length.

_Cz_compress

Compresses source buffer into the destination buffer. sourceLen is the byte length of the source buffer. Upon entry, destLen is the total size of the destination buffer, which must be at least the value returned by compressBound(sourceLen). Upon exit, destLen is the actual size of the compressed data. compress() is equivalent to compress2() with a level parameter of Z_DEFAULT_COMPRESSION.

_Cz_compress2

Compresses source buffer into the destination buffer. The level parameter has the same meaning as in deflateInit. sourceLen is the byte length of the source buffer. Upon entry, destLen is the total size of the destination buffer, which must be at least the value returned by compressBound(sourceLen). Upon exit, destLen is the actual size of the compressed data.

_Cz_compressBound

Returns an upper bound on the compressed size after compress() or compress2() on sourceLen bytes. It would be used before a compress() or compress2() call to allocate the destination buffer.

_Cz_crc32

Update a running CRC-32 with the bytes buf[0..len-1] and return the updated CRC-32. If buf is Z_NULL, this function returns the required initial value for the crc. Pre- and post-conditioning (one's complement) is performed within this function so it shouldn't be done by the application.

Usage example:

uLong crc = crc32(0L, Z_NULL, 0);
while (read_buffer(buffer, length) != EOF) {
  crc = crc32(crc, buffer, length);
}
if (crc != original_crc) error();

_Cz_crc32_combine

Combine two CRC-32 check values into one. For two sequences of bytes, seq1 and seq2 with lengths len1 and len2, CRC-32 check values were calculated for each, crc1 and crc2. crc32_combine() returns the CRC-32 check value of seq1 and seq2 concatenated, requiring only crc1, crc2, and len2.

_Cz_crc32_z

Same as crc32(), but with a size_t length.

_Cz_deflate

deflate compresses as much data as possible, and stops when the input buffer becomes empty or the output buffer becomes full. It may introduce some output latency (reading input without producing any output) except when forced to flush.

The detailed semantics are as follows. deflate performs one or both of the following actions:

• Compress more input starting at next_in and update next_in and

avail_in accordingly. If not all input can be processed (because there is not enough room in the output buffer), next_in and avail_in are updated and processing will resume at this point for the next call of deflate().

• Generate more output starting at next_out and update next_out and

avail_out accordingly. This action is forced if the parameter flush is non zero. Forcing flush frequently degrades the compression ratio, so this parameter should be set only when necessary. Some output may be provided even if flush is zero.

Before the call of deflate(), the application should ensure that at least one of the actions is possible, by providing more input and/or consuming more output, and updating avail_in or avail_out accordingly; avail_out should never be zero before the call. The application can consume the compressed output when it wants, for example when the output buffer is full (avail_out == 0), or after each call of deflate(). If deflate returns Z_OK and with zero avail_out, it must be called again after making room in the output buffer because there might be more output pending. See deflatePending(), which can be used if desired to determine whether or not there is more ouput in that case.

Normally the parameter flush is set to Z_NO_FLUSH, which allows deflate to decide how much data to accumulate before producing output, in order to maximize compression.

If the parameter flush is set to Z_SYNC_FLUSH, all pending output is flushed to the output buffer and the output is aligned on a byte boundary, so that the decompressor can get all input data available so far. (In particular avail_in is zero after the call if enough output space has been provided before the call.) Flushing may degrade compression for some compression algorithms and so it should be used only when necessary. This completes the current deflate block and follows it with an empty stored block that is three bits plus filler bits to the next byte, followed by four bytes (00 00 ff ff).

If flush is set to Z_PARTIAL_FLUSH, all pending output is flushed to the output buffer, but the output is not aligned to a byte boundary. All of the input data so far will be available to the decompressor, as for Z_SYNC_FLUSH. This completes the current deflate block and follows it with an empty fixed codes block that is 10 bits long. This assures that enough bytes are output in order for the decompressor to finish the block before the empty fixed codes block.

If flush is set to Z_BLOCK, a deflate block is completed and emitted, as for Z_SYNC_FLUSH, but the output is not aligned on a byte boundary, and up to seven bits of the current block are held to be written as the next byte after the next deflate block is completed. In this case, the decompressor may not be provided enough bits at this point in order to complete decompression of the data provided so far to the compressor. It may need to wait for the next block to be emitted. This is for advanced applications that need to control the emission of deflate blocks.

If flush is set to Z_FULL_FLUSH, all output is flushed as with Z_SYNC_FLUSH, and the compression state is reset so that decompression can restart from this point if previous compressed data has been damaged or if random access is desired. Using Z_FULL_FLUSH too often can seriously degrade compression.

If deflate returns with avail_out == 0, this function must be called again with the same value of the flush parameter and more output space (updated avail_out), until the flush is complete (deflate returns with non-zero avail_out). In the case of a Z_FULL_FLUSH or Z_SYNC_FLUSH, make sure that avail_out is greater than six to avoid repeated flush markers due to avail_out == 0 on return.

If the parameter flush is set to Z_FINISH, pending input is processed, pending output is flushed and deflate returns with Z_STREAM_END if there was enough output space. If deflate returns with Z_OK or Z_BUF_ERROR, this function must be called again with Z_FINISH and more output space (updated avail_out) but no more input data, until it returns with Z_STREAM_END or an error. After deflate has returned Z_STREAM_END, the only possible operations on the stream are deflateReset or deflateEnd.

Z_FINISH can be used in the first deflate call after deflateInit if all the compression is to be done in a single step. In order to complete in one call, avail_out must be at least the value returned by deflateBound (see below). Then deflate is guaranteed to return Z_STREAM_END. If not enough output space is provided, deflate will not return Z_STREAM_END, and it must be called again as described above.

deflate() sets strm->adler to the Adler-32 checksum of all input read so far (that is, total_in bytes). If a gzip stream is being generated, then strm->adler will be the CRC-32 checksum of the input read so far. (See deflateInit2 below.)

deflate() may update strm->data_type if it can make a good guess about the input data type (Z_BINARY or Z_TEXT). If in doubt, the data is considered binary. This field is only for information purposes and does not affect the compression algorithm in any manner.

_Cz_deflateBound

deflateBound() returns an upper bound on the compressed size after deflation of sourceLen bytes. It must be called after deflateInit() or deflateInit2(), and after deflateSetHeader(), if used. This would be used to allocate an output buffer for deflation in a single pass, and so would be called before deflate(). If that first deflate() call is provided the sourceLen input bytes, an output buffer allocated to the size returned by deflateBound(), and the flush value Z_FINISH, then deflate() is guaranteed to return Z_STREAM_END. Note that it is possible for the compressed size to be larger than the value returned by deflateBound() if flush options other than Z_FINISH or Z_NO_FLUSH are used.

_Cz_deflateCopy

Sets destination stream as a complete copy of the source stream.

This function can be useful when several compression strategies will be tried, for example when there are several ways of pre-processing the input data with a filter. The streams that will be discarded should then be freed by calling deflateEnd. Note that deflateCopy duplicates the internal compression state which can be quite large, so this strategy is slow and can consume lots of memory.

_Cz_deflateEnd

All dynamically allocated data structures for this stream are freed. This function discards any unprocessed input and does not flush any pending output.

_Cz_deflateGetDictionary

Returns the sliding dictionary being maintained by deflate. dictLength is set to the number of bytes in the dictionary, and that many bytes are copied to dictionary. dictionary must have enough space, where 32768 bytes is always enough. If deflateGetDictionary() is called with dictionary equal to Z_NULL, then only the dictionary length is returned, and nothing is copied. Similary, if dictLength is Z_NULL, then it is not set.

deflateGetDictionary() may return a length less than the window size, even when more than the window size in input has been provided. It may return up to 258 bytes less in that case, due to how zlib's implementation of deflate manages the sliding window and lookahead for matches, where matches can be up to 258 bytes long. If the application needs the last window-size bytes of input, then that would need to be saved by the application outside of zlib.

_Cz_deflateInit

Initializes the internal stream state for compression. The fields zalloc, zfree and opaque must be initialized before by the caller. If zalloc and zfree are set to Z_NULL, deflateInit updates them to use default allocation functions.

The compression level must be Z_DEFAULT_COMPRESSION, or between 0 and 9: 1 gives best speed, 9 gives best compression, 0 gives no compression at all (the input data is simply copied a block at a time). Z_DEFAULT_COMPRESSION requests a default compromise between speed and compression (currently equivalent to level 6).

_Cz_deflateInit2

This is another version of deflateInit with more compression options. The fields next_in, zalloc, zfree and opaque must be initialized before by the caller.

The method parameter is the compression method. It must be Z_DEFLATED in this version of the library.

The windowBits parameter is the base two logarithm of the window size (the size of the history buffer). It should be in the range 8..15 for this version of the library. Larger values of this parameter result in better compression at the expense of memory usage. The default value is 15 if deflateInit is used instead.

For the current implementation of deflate(), a windowBits value of 8 (a window size of 256 bytes) is not supported. As a result, a request for 8 will result in 9 (a 512-byte window). In that case, providing 8 to inflateInit2() will result in an error when the zlib header with 9 is checked against the initialization of inflate(). The remedy is to not use 8 with deflateInit2() with this initialization, or at least in that case use 9 with inflateInit2().

windowBits can also be -8..-15 for raw deflate. In this case, -windowBits determines the window size. deflate() will then generate raw deflate data with no zlib header or trailer, and will not compute a check value.

windowBits can also be greater than 15 for optional gzip encoding. Add 16 to windowBits to write a simple gzip header and trailer around the compressed data instead of a zlib wrapper. The gzip header will have no file name, no extra data, no comment, no modification time (set to zero), no header crc, and the operating system will be set to the appropriate value, if the operating system was determined at compile time. If a gzip stream is being written, strm->adler is a CRC-32 instead of an Adler-32.

For raw deflate or gzip encoding, a request for a 256-byte window is rejected as invalid, since only the zlib header provides a means of transmitting the window size to the decompressor.

The memLevel parameter specifies how much memory should be allocated for the internal compression state. memLevel=1 uses minimum memory but is slow and reduces compression ratio; memLevel=9 uses maximum memory for optimal speed. The default value is 8. See zconf.h for total memory usage as a function of windowBits and memLevel.

The strategy parameter is used to tune the compression algorithm. Use the value Z_DEFAULT_STRATEGY for normal data, Z_FILTERED for data produced by a filter (or predictor), Z_HUFFMAN_ONLY to force Huffman encoding only (no string match), or Z_RLE to limit match distances to one (run-length encoding). Filtered data consists mostly of small values with a somewhat random distribution. In this case, the compression algorithm is tuned to compress them better. The effect of Z_FILTERED is to force more Huffman coding and less string matching; it is somewhat intermediate between Z_DEFAULT_STRATEGY and Z_HUFFMAN_ONLY. Z_RLE is designed to be almost as fast as Z_HUFFMAN_ONLY, but give better compression for PNG image data. The strategy parameter only affects the compression ratio but not the correctness of the compressed output even if it is not set appropriately. Z_FIXED prevents the use of dynamic Huffman codes, allowing for a simpler decoder for special applications.

_Cz_deflateParams

Dynamically update the compression level and compression strategy. The interpretation of level and strategy is as in deflateInit2(). This can be used to switch between compression and straight copy of the input data, or to switch to a different kind of input data requiring a different strategy. If the compression approach (which is a function of the level) or the strategy is changed, and if any input has been consumed in a previous deflate() call, then the input available so far is compressed with the old level and strategy using deflate(strm, Z_BLOCK). There are three approaches for the compression levels 0, 1..3, and 4..9 respectively. The new level and strategy will take effect at the next call of deflate().

If a deflate(strm, Z_BLOCK) is performed by deflateParams(), and it does not have enough output space to complete, then the parameter change will not take effect. In this case, deflateParams() can be called again with the same parameters and more output space to try again.

In order to assure a change in the parameters on the first try, the deflate stream should be flushed using deflate() with Z_BLOCK or other flush request until strm.avail_out is not zero, before calling deflateParams(). Then no more input data should be provided before the deflateParams() call. If this is done, the old level and strategy will be applied to the data compressed before deflateParams(), and the new level and strategy will be applied to the the data compressed after deflateParams().

deflateParams returns Z_OK on success, Z_STREAM_ERROR if the source stream state was inconsistent or if a parameter was invalid, or Z_BUF_ERROR if there was not enough output space to complete the compression of the available input data before a change in the strategy or approach. Note that in the case of a Z_BUF_ERROR, the parameters are not changed. A return value of Z_BUF_ERROR is not fatal, in which case deflateParams() can be retried with more output space.

_Cz_deflatePending

deflatePending() returns the number of bytes and bits of output that have been generated, but not yet provided in the available output. The bytes not provided would be due to the available output space having being consumed. The number of bits of output not provided are between 0 and 7, where they await more bits to join them in order to fill out a full byte. If pending or bits are Z_NULL, then those values are not set.

_Cz_deflatePrime

deflatePrime() inserts bits in the deflate output stream. The intent is that this function is used to start off the deflate output with the bits leftover from a previous deflate stream when appending to it. As such, this function can only be used for raw deflate, and must be used before the first deflate() call after a deflateInit2() or deflateReset(). bits must be less than or equal to 16, and that many of the least significant bits of value will be inserted in the output.

_Cz_deflateReset

This function is equivalent to deflateEnd followed by deflateInit, but does not free and reallocate the internal compression state. The stream will leave the compression level and any other attributes that may have been set unchanged.

deflateReset returns Z_OK if success, or Z_STREAM_ERROR if the source stream state was inconsistent (such as zalloc or state being Z_NULL).

_Cz_deflateSetDictionary

Initializes the compression dictionary from the given byte sequence without producing any compressed output. When using the zlib format, this function must be called immediately after deflateInit, deflateInit2 or deflateReset, and before any call of deflate. When doing raw deflate, this function must be called either before any call of deflate, or immediately after the completion of a deflate block, i.e. after all input has been consumed and all output has been delivered when using any of the flush options Z_BLOCK, Z_PARTIAL_FLUSH, Z_SYNC_FLUSH, or Z_FULL_FLUSH. The compressor and decompressor must use exactly the same dictionary (see inflateSetDictionary).

The dictionary should consist of strings (byte sequences) that are likely to be encountered later in the data to be compressed, with the most commonly used strings preferably put towards the end of the dictionary. Using a dictionary is most useful when the data to be compressed is short and can be predicted with good accuracy; the data can then be compressed better than with the default empty dictionary.

Depending on the size of the compression data structures selected by deflateInit or deflateInit2, a part of the dictionary may in effect be discarded, for example if the dictionary is larger than the window size provided in deflateInit or deflateInit2. Thus the strings most likely to be useful should be put at the end of the dictionary, not at the front. In addition, the current implementation of deflate will use at most the window size minus 262 bytes of the provided dictionary.

Upon return of this function, strm->adler is set to the Adler-32 value of the dictionary; the decompressor may later use this value to determine which dictionary has been used by the compressor. (The Adler-32 value applies to the whole dictionary even if only a subset of the dictionary is actually used by the compressor.) If a raw deflate was requested, then the Adler-32 value is not computed and strm->adler is not set.

_Cz_deflateSetHeader

Provides gzip header information for when a gzip stream is requested by deflateInit2(). deflateSetHeader() may be called after deflateInit2() or deflateReset() and before the first call of deflate(). The text, time, os, extra field, name, and comment information in the provided gz_header structure are written to the gzip header (xflag is ignored -- the extra flags are set according to the compression level). The caller must assure that, if not Z_NULL, name and comment are terminated with a zero byte, and that if extra is not Z_NULL, that extra_len bytes are available there. If hcrc is true, a gzip header crc is included. Note that the current versions of the command-line version of gzip (up through version 1.3.x) do not support header crc's, and will report that it is a "multi-part gzip file" and give up.

If deflateSetHeader is not used, the default gzip header has text false, the time set to zero, and os set to 255, with no extra, name, or comment fields. The gzip header is returned to the default state by deflateReset().

deflateSetHeader returns Z_OK if success, or Z_STREAM_ERROR if the source stream state was inconsistent.

_Cz_deflateTune

Fine tune deflate's internal compression parameters. This should only be used by someone who understands the algorithm used by zlib's deflate for searching for the best matching string, and even then only by the most fanatic optimizer trying to squeeze out the last compressed bit for their specific input data. Read the deflate.c source code for the meaning of the max_lazy, good_length, nice_length, and max_chain parameters.

deflateTune() can be called after deflateInit() or deflateInit2(), and returns Z_OK on success, or Z_STREAM_ERROR for an invalid deflate stream.

_Cz_gzbuffer

Sets internal buffer size used by this library's functions. The default buffer size is 8192 bytes. This function must be called after gzopen() or gzdopen(), and before any other calls that read or write the file. The buffer memory allocation is always deferred to the first read or write. Three times that size in buffer space is allocated. A larger buffer size of, for example, 64K or 128K bytes will noticeably increase the speed of decompression (reading).

The new buffer size also affects the maximum length for gzprintf().

_Cz_gzclearerr

Clears the error and end-of-file flags for file. This is analogous to the clearerr() function in stdio. This is useful for continuing to read a gzip file that is being written concurrently.

_Cz_gzclose

Flushes all pending output if necessary, closes the compressed file and deallocates the (de)compression state. Note that once file is closed, you cannot call gzerror with file, since its structures have been deallocated. gzclose must not be called more than once on the same file, just as free must not be called more than once on the same allocation.

_Cz_gzclose_r

Same as gzclose(), but gzclose_r() is only for use when reading, and gzclose_w() is only for use when writing or appending. The advantage to using these instead of gzclose() is that they avoid linking in zlib compression or decompression code that is not used when only reading or only writing respectively. If gzclose() is used, then both compression and decompression code will be included the application when linking to a static zlib library.

_Cz_gzdirect

Returns true (1) if file is being copied directly while reading, or false (0) if file is a gzip stream being decompressed.

If the input file is empty, gzdirect() will return true, since the input does not contain a gzip stream.

If gzdirect() is used immediately after gzopen() or gzdopen() it will cause buffers to be allocated to allow reading the file to determine if it is a gzip file. Therefore if gzbuffer() is used, it should be called before gzdirect().

When writing, gzdirect() returns true (1) if transparent writing was requested ("wT" for the gzopen() mode), or false (0) otherwise. (Note: gzdirect() is not needed when writing. Transparent writing must be explicitly requested, so the application already knows the answer. When linking statically, using gzdirect() will include all of the zlib code for gzip file reading and decompression, which may not be desired.)

_Cz_gzdopen

Associates gzFile with the file descriptor.

File descriptors are obtained from calls like open, dup, creat, pipe or fileno (if the file has been previously opened with fopen). The mode parameter is as in gzopen.

The next call of gzclose on the returned gzFile will also close the file descriptor fd, just like fclose(fdopen(fd, mode)) closes the file descriptor fd. If you want to keep fd open, use fd = dup(fd_keep); gz = gzdopen(fd, mode);. The duplicated descriptor should be saved to avoid a leak, since gzdopen does not close fd if it fails. If you are using fileno() to get the file descriptor from a FILE *, then you will have to use dup() to avoid double-close()ing the file descriptor. Both gzclose() and fclose() will close the associated file descriptor, so they need to have different file descriptors.

_Cz_gzeof

Returns true (1) if the end-of-file indicator has been set while reading, false (0) otherwise. Note that the end-of-file indicator is set only if the read tried to go past the end of the input, but came up short. Therefore, just like feof(), gzeof() may return false even if there is no more data to read, in the event that the last read request was for the exact number of bytes remaining in the input file. This will happen if the input file size is an exact multiple of the buffer size.

If gzeof() returns true, then the read functions will return no more data, unless the end-of-file indicator is reset by gzclearerr() and the input file has grown since the previous end of file was detected.

_Cz_gzerror

Returns the error message for the last error which occurred on the given compressed file. errnum is set to zlib error number. If an error occurred in the file system and not in the compression library, errnum is set to Z_ERRNO and the application may consult errno to get the exact error code.

The application must not modify the returned string. Future calls to this function may invalidate the previously returned string. If file is closed, then the string previously returned by gzerror will no longer be available.

gzerror() should be used to distinguish errors from end-of-file for those functions above that do not distinguish those cases in their return values.

_Cz_gzflush

Flushes all pending output into the compressed file. The parameter flush is as in the deflate() function. The return value is the zlib error number (see function gzerror below). gzflush is only permitted when writing.

If the flush parameter is Z_FINISH, the remaining data is written and the gzip stream is completed in the output. If gzwrite() is called again, a new gzip stream will be started in the output. gzread() is able to read such concatenated gzip streams.

gzflush should be called only when strictly necessary because it will degrade compression if called too often.

_Cz_gzfread

Read up to nitems items of size size from file to buf, otherwise operating as gzread() does. This duplicates the interface of stdio's fread(), with size_t request and return types. If the library defines size_t, then size_t is identical to size_t. If not, then size_t is an unsigned integer type that can contain a pointer.

gzfread() returns the number of full items read of size size, or zero if the end of the file was reached and a full item could not be read, or if there was an error. gzerror() must be consulted if zero is returned in order to determine if there was an error. If the multiplication of size and nitems overflows, i.e. the product does not fit in a size_t, then nothing is read, zero is returned, and the error state is set to Z_STREAM_ERROR.

In the event that the end of file is reached and only a partial item is available at the end, i.e. the remaining uncompressed data length is not a multiple of size, then the final partial item is nevetheless read into buf and the end-of-file flag is set. The length of the partial item read is not provided, but could be inferred from the result of gztell(). This behavior is the same as the behavior of fread() implementations in common libraries, but it prevents the direct use of gzfread() to read a concurrently written file, reseting and retrying on end-of-file, when size is not 1.

_Cz_gzfwrite

Writes nitems items of size size from buf to file, duplicating the interface of stdio's fwrite(), with size_t request and return types. If the library defines size_t, then size_t is identical to size_t. If not, then size_t is an unsigned integer type that can contain a pointer.

gzfwrite() returns the number of full items written of size size, or zero if there was an error. If the multiplication of size and nitems overflows, i.e. the product does not fit in a size_t, then nothing is written, zero is returned, and the error state is set to Z_STREAM_ERROR.

_Cz_gzgetc

Reads one byte from the compressed file. gzgetc returns this byte or -1 in case of end of file or error. This is implemented as a macro for speed. As such, it does not do all of the checking the other functions do. I.e. it does not check to see if file is NULL, nor whether the structure file points to has been clobbered or not.

_Cz_gzgets

Reads bytes from the compressed file until len-1 characters are read, or a newline character is read and transferred to buf, or an end-of-file condition is encountered. If any characters are read or if len == 1, the string is terminated with a null character. If no characters are read due to an end-of-file or len < 1, then the buffer is left untouched.

_Cz_gzoffset

Returns current offset in the file being read or written. This offset includes the count of bytes that precede the gzip stream, for example when appending or when using gzdopen() for reading. When reading, the offset does not include as yet unused buffered input. This information can be used for a progress indicator. On error, gzoffset() returns -1.

_Cz_gzopen

Opens a gzip (.gz) file for reading or writing.

The mode parameter is as in fopen ("rb" or "wb") but can also include a compression level ("wb9") or a strategy: 'f' for filtered data as in "wb6f", 'h' for Huffman-only compression as in "wb1h", 'R' for run-length encoding as in "wb1R", or 'F' for fixed code compression as in "wb9F". (See the description of deflateInit2 for more information about the strategy parameter.) 'T' will request transparent writing or appending with no compression and not using the gzip format.

"a" can be used instead of "w" to request that the gzip stream that will be written be appended to the file. "+" will result in an error, since reading and writing to the same gzip file is not supported. The addition of "x" when writing will create the file exclusively, which fails if the file already exists. On systems that support it, the addition of "e" when reading or writing will set the flag to close the file on an execve() call.

These functions, as well as gzip, will read and decode a sequence of gzip streams in a file. The append function of gzopen() can be used to create such a file. (Also see gzflush() for another way to do this.) When appending, gzopen does not test whether the file begins with a gzip stream, nor does it look for the end of the gzip streams to begin appending. gzopen will simply append a gzip stream to the existing file.

gzopen can be used to read a file which is not in gzip format; in this case gzread will directly read from the file without decompression. When reading, this will be detected automatically by looking for the magic two- byte gzip header.

_Cz_gzprintf

Converts, formats, and writes the arguments to the compressed file under control of the format string, as in fprintf. gzprintf returns the number of uncompressed bytes actually written, or a negative zlib error code in case of error. The number of uncompressed bytes written is limited to 8191, or one less than the buffer size given to gzbuffer(). The caller should assure that this limit is not exceeded. If it is exceeded, then gzprintf() will return an error (0) with nothing written. In this case, there may also be a buffer overflow with unpredictable consequences, which is possible only if zlib was compiled with the insecure functions sprintf() or vsprintf() because the secure snprintf() or vsnprintf() functions were not available. This can be determined using zlibCompileFlags().

_Cz_gzputc

Writes character converted to an unsigned char into compressed file.

_Cz_gzputs

Writes the given null-terminated string to the compressed file, excluding the terminating null character.

_Cz_gzread

Reads given number of uncompressed bytes from the compressed file. If the input file is not in gzip format, gzread copies the given number of bytes into the buffer directly from the file.

After reaching the end of a gzip stream in the input, gzread will continue to read, looking for another gzip stream. Any number of gzip streams may be concatenated in the input file, and will all be decompressed by gzread(). If something other than a gzip stream is encountered after a gzip stream, that remaining trailing garbage is ignored (and no error is returned).

gzread can be used to read a gzip file that is being concurrently written. Upon reaching the end of the input, gzread will return with the available data. If the error code returned by gzerror is Z_OK or Z_BUF_ERROR, then gzclearerr can be used to clear the end of file indicator in order to permit gzread to be tried again. Z_OK indicates that a gzip stream was completed on the last gzread. Z_BUF_ERROR indicates that the input file ended in the middle of a gzip stream. Note that gzread does not return -1 in the event of an incomplete gzip stream. This error is deferred until gzclose(), which will return Z_BUF_ERROR if the last gzread ended in the middle of a gzip stream. Alternatively, gzerror can be used before gzclose to detect this case.

_Cz_gzrewind

Rewinds file.

This function is supported only for reading.

_Cz_gzseek

Sets starting position for the next gzread or gzwrite on the given compressed file. The offset represents a number of bytes in the uncompressed data stream. The whence parameter is defined as in lseek(2); the value SEEK_END is not supported.

If the file is opened for reading, this function is emulated but can be extremely slow. If the file is opened for writing, only forward seeks are supported; gzseek then compresses a sequence of zeroes up to the new starting position.

_Cz_gzsetparams

Dynamically update the compression level or strategy. See the description of deflateInit2 for the meaning of these parameters. Previously provided data is flushed before the parameter change.

_Cz_gztell

Returns starting position for the next gzread or gzwrite on the given compressed file. This position represents a number of bytes in the uncompressed data stream, and is zero when starting, even if appending or reading a gzip stream from the middle of a file using gzdopen().

_Cz_gzungetc

Pushes one character back onto the stream to be read as the first character on the next read. At least one character of push-back is allowed. gzungetc() returns the character pushed, or -1 on failure. gzungetc() will fail if c is -1, and may fail if a character has been pushed but not read yet. If gzungetc is used immediately after gzopen or gzdopen, at least the output buffer size of pushed characters is allowed. (See gzbuffer above.) The pushed character will be discarded if the stream is repositioned with gzseek() or gzrewind().

_Cz_gzwrite

Writes given number of uncompressed bytes into the compressed file. gzwrite returns the number of uncompressed bytes written or 0 in case of error.

_Cz_inflate

inflate decompresses as much data as possible, and stops when the input buffer becomes empty or the output buffer becomes full. It may introduce some output latency (reading input without producing any output) except when forced to flush.

The detailed semantics are as follows. inflate performs one or both of the following actions:

• Decompress more input starting at next_in and update next_in and avail_in accordingly. If not all input can be processed (because there is not enough room in the output buffer), then next_in and avail_in are updated accordingly, and processing will resume at this point for the next call of inflate().
• Generate more output starting at next_out and update next_out and avail_out accordingly. inflate() provides as much output as possible, until there is no more input data or no more space in the output buffer (see below about the flush parameter).

Before the call of inflate(), the application should ensure that at least one of the actions is possible, by providing more input and/or consuming more output, and updating the next_* and avail_* values accordingly. If the caller of inflate() does not provide both available input and available output space, it is possible that there will be no progress made. The application can consume the uncompressed output when it wants, for example when the output buffer is full (avail_out == 0), or after each call of inflate(). If inflate returns Z_OK and with zero avail_out, it must be called again after making room in the output buffer because there might be more output pending.

The flush parameter of inflate() can be Z_NO_FLUSH, Z_SYNC_FLUSH, Z_FINISH, Z_BLOCK, or Z_TREES. Z_SYNC_FLUSH requests that inflate() flush as much output as possible to the output buffer. Z_BLOCK requests that inflate() stop if and when it gets to the next deflate block boundary. When decoding the zlib or gzip format, this will cause inflate() to return immediately after the header and before the first block. When doing a raw inflate, inflate() will go ahead and process the first block, and will return when it gets to the end of that block, or when it runs out of data.

The Z_BLOCK option assists in appending to or combining deflate streams. To assist in this, on return inflate() always sets strm->data_type to the number of unused bits in the last byte taken from strm->next_in, plus 64 if inflate() is currently decoding the last block in the deflate stream, plus 128 if inflate() returned immediately after decoding an end-of-block code or decoding the complete header up to just before the first byte of the deflate stream. The end-of-block will not be indicated until all of the uncompressed data from that block has been written to strm->next_out. The number of unused bits may in general be greater than seven, except when bit 7 of data_type is set, in which case the number of unused bits will be less than eight. data_type is set as noted here every time inflate() returns for all flush options, and so can be used to determine the amount of currently consumed input in bits.

The Z_TREES option behaves as Z_BLOCK does, but it also returns when the end of each deflate block header is reached, before any actual data in that block is decoded. This allows the caller to determine the length of the deflate block header for later use in random access within a deflate block. 256 is added to the value of strm->data_type when inflate() returns immediately after reaching the end of the deflate block header.

inflate() should normally be called until it returns Z_STREAM_END or an error. However if all decompression is to be performed in a single step (a single call of inflate), the parameter flush should be set to Z_FINISH. In this case all pending input is processed and all pending output is flushed; avail_out must be large enough to hold all of the uncompressed data for the operation to complete. (The size of the uncompressed data may have been saved by the compressor for this purpose.) The use of Z_FINISH is not required to perform an inflation in one step. However it may be used to inform inflate that a faster approach can be used for the single inflate() call. Z_FINISH also informs inflate to not maintain a sliding window if the stream completes, which reduces inflate's memory footprint. If the stream does not complete, either because not all of the stream is provided or not enough output space is provided, then a sliding window will be allocated and inflate() can be called again to continue the operation as if Z_NO_FLUSH had been used.

In this implementation, inflate() always flushes as much output as possible to the output buffer, and always uses the faster approach on the first call. So the effects of the flush parameter in this implementation are on the return value of inflate() as noted below, when inflate() returns early when Z_BLOCK or Z_TREES is used, and when inflate() avoids the allocation of memory for a sliding window when Z_FINISH is used.

If a preset dictionary is needed after this call (see inflateSetDictionary below), inflate sets strm->adler to the Adler-32 checksum of the dictionary chosen by the compressor and returns Z_NEED_DICT; otherwise it sets strm->adler to the Adler-32 checksum of all output produced so far (that is, total_out bytes) and returns Z_OK, Z_STREAM_END or an error code as described below. At the end of the stream, inflate() checks that its computed Adler-32 checksum is equal to that saved by the compressor and returns Z_STREAM_END only if the checksum is correct.

inflate() can decompress and check either zlib-wrapped or gzip-wrapped deflate data. The header type is detected automatically, if requested when initializing with inflateInit2(). Any information contained in the gzip header is not retained unless inflateGetHeader() is used. When processing gzip-wrapped deflate data, strm->adler32 is set to the CRC-32 of the output produced so far. The CRC-32 is checked against the gzip trailer, as is the uncompressed length, modulo 2^32.

_Cz_inflateBack

inflateBack() does a raw inflate with a single call using a call-back interface for input and output. This is potentially more efficient than inflate() for file i/o applications, in that it avoids copying between the output and the sliding window by simply making the window itself the output buffer. inflate() can be faster on modern CPUs when used with large buffers. inflateBack() trusts the application to not change the output buffer passed by the output function, at least until inflateBack() returns.

inflateBackInit() must be called first to allocate the internal state and to initialize the state with the user-provided window buffer. inflateBack() may then be used multiple times to inflate a complete, raw deflate stream with each call. inflateBackEnd() is then called to free the allocated state.

A raw deflate stream is one with no zlib or gzip header or trailer. This routine would normally be used in a utility that reads zip or gzip files and writes out uncompressed files. The utility would decode the header and process the trailer on its own, hence this routine expects only the raw deflate stream to decompress. This is different from the default behavior of inflate(), which expects a zlib header and trailer around the deflate stream.

inflateBack() uses two subroutines supplied by the caller that are then called by inflateBack() for input and output. inflateBack() calls those routines until it reads a complete deflate stream and writes out all of the uncompressed data, or until it encounters an error. The function's parameters and return types are defined above in the in_func and out_func typedefs. inflateBack() will call in(in_desc, &buf) which should return the number of bytes of provided input, and a pointer to that input in buf. If there is no input available, in() must return zero -- buf is ignored in that case -- and inflateBack() will return a buffer error. inflateBack() will call out(out_desc, buf, len) to write the uncompressed data buf[0..len-1]. out() should return zero on success, or non-zero on failure. If out() returns non-zero, inflateBack() will return with an error. Neither in() nor out() are permitted to change the contents of the window provided to inflateBackInit(), which is also the buffer that out() uses to write from. The length written by out() will be at most the window size. Any non-zero amount of input may be provided by in().

For convenience, inflateBack() can be provided input on the first call by setting strm->next_in and strm->avail_in. If that input is exhausted, then in() will be called. Therefore strm->next_in must be initialized before calling inflateBack(). If strm->next_in is Z_NULL, then in() will be called immediately for input. If strm->next_in is not Z_NULL, then strm->avail_in must also be initialized, and then if strm->avail_in is not zero, input will initially be taken from strm->next_in[0 .. strm->avail_in - 1].

The in_desc and out_desc parameters of inflateBack() is passed as the first parameter of in() and out() respectively when they are called. These descriptors can be optionally used to pass any information that the caller- supplied in() and out() functions need to do their job.

On return, inflateBack() will set strm->next_in and strm->avail_in to pass back any unused input that was provided by the last in() call. The return values of inflateBack() can be Z_STREAM_END on success, Z_BUF_ERROR if in() or out() returned an error, Z_DATA_ERROR if there was a format error in the deflate stream (in which case strm->msg is set to indicate the nature of the error), or Z_STREAM_ERROR if the stream was not properly initialized. In the case of Z_BUF_ERROR, an input or output error can be distinguished using strm->next_in which will be Z_NULL only if in() returned an error. If strm->next_in is not Z_NULL, then the Z_BUF_ERROR was due to out() returning non-zero. (in() will always be called before out(), so strm->next_in is assured to be defined if out() returns non-zero.) Note that inflateBack() cannot return Z_OK.

_Cz_inflateBackEnd

All memory allocated by inflateBackInit() is freed.

_Cz_inflateCopy

Sets the destination stream as a complete copy of the source stream.

This function can be useful when randomly accessing a large stream. The first pass through the stream can periodically record the inflate state, allowing restarting inflate at those points when randomly accessing the stream.

_Cz_inflateEnd

All dynamically allocated data structures for this stream are freed. This function discards any unprocessed input and does not flush any pending output.

_Cz_inflateGetDictionary

Returns the sliding dictionary being maintained by inflate. dictLength is set to the number of bytes in the dictionary, and that many bytes are copied to dictionary. dictionary must have enough space, where 32768 bytes is always enough. If inflateGetDictionary() is called with dictionary equal to Z_NULL, then only the dictionary length is returned, and nothing is copied. Similary, if dictLength is Z_NULL, then it is not set.

_Cz_inflateGetHeader

inflateGetHeader() requests that gzip header information be stored in the provided gz_header structure. inflateGetHeader() may be called after inflateInit2() or inflateReset(), and before the first call of inflate(). As inflate() processes the gzip stream, head->done is zero until the header is completed, at which time head->done is set to one. If a zlib stream is being decoded, then head->done is set to -1 to indicate that there will be no gzip header information forthcoming. Note that Z_BLOCK or Z_TREES can be used to force inflate() to return immediately after header processing is complete and before any actual data is decompressed.

The text, time, xflags, and os fields are filled in with the gzip header contents. hcrc is set to true if there is a header CRC. (The header CRC was valid if done is set to one.) If extra is not Z_NULL, then extra_max contains the maximum number of bytes to write to extra. Once done is true, extra_len contains the actual extra field length, and extra contains the extra field, or that field truncated if extra_max is less than extra_len. If name is not Z_NULL, then up to name_max characters are written there, terminated with a zero unless the length is greater than name_max. If comment is not Z_NULL, then up to comm_max characters are written there, terminated with a zero unless the length is greater than comm_max. When any of extra, name, or comment are not Z_NULL and the respective field is not present in the header, then that field is set to Z_NULL to signal its absence. This allows the use of deflateSetHeader() with the returned structure to duplicate the header. However if those fields are set to allocated memory, then the application will need to save those pointers elsewhere so that they can be eventually freed.

If inflateGetHeader is not used, then the header information is simply discarded. The header is always checked for validity, including the header CRC if present. inflateReset() will reset the process to discard the header information. The application would need to call inflateGetHeader() again to retrieve the header from the next gzip stream.

_Cz_inflateInit

Initializes the internal stream state for decompression. The fields next_in, avail_in, zalloc, zfree and opaque must be initialized before by the caller. In the current version of inflate, the provided input is not read or consumed. The allocation of a sliding window will be deferred to the first call of inflate (if the decompression does not complete on the first call). If zalloc and zfree are set to Z_NULL, inflateInit updates them to use default allocation functions.

_Cz_inflateInit2

This is another version of inflateInit with an extra parameter. The fields next_in, avail_in, zalloc, zfree and opaque must be initialized before by the caller.

The windowBits parameter is the base two logarithm of the maximum window size (the size of the history buffer). It should be in the range 8..15 for this version of the library. The default value is 15 if inflateInit is used instead. windowBits must be greater than or equal to the windowBits value provided to deflateInit2() while compressing, or it must be equal to 15 if deflateInit2() was not used. If a compressed stream with a larger window size is given as input, inflate() will return with the error code Z_DATA_ERROR instead of trying to allocate a larger window.

windowBits can also be zero to request that inflate use the window size in the zlib header of the compressed stream.

windowBits can also be -8..-15 for raw inflate. In this case, -windowBits determines the window size. inflate() will then process raw deflate data, not looking for a zlib or gzip header, not generating a check value, and not looking for any check values for comparison at the end of the stream. This is for use with other formats that use the deflate compressed data format such as zip. Those formats provide their own check values. If a custom format is developed using the raw deflate format for compressed data, it is recommended that a check value such as an Adler-32 or a CRC-32 be applied to the uncompressed data as is done in the zlib, gzip, and zip formats. For most applications, the zlib format should be used as is. Note that comments above on the use in deflateInit2() applies to the magnitude of windowBits.

windowBits can also be greater than 15 for optional gzip decoding. Add 32 to windowBits to enable zlib and gzip decoding with automatic header detection, or add 16 to decode only the gzip format (the zlib format will return a Z_DATA_ERROR). If a gzip stream is being decoded, strm->adler is a CRC-32 instead of an Adler-32. Unlike the gunzip utility and gzread() (see below), inflate() will not automatically decode concatenated gzip streams. inflate() will return Z_STREAM_END at the end of the gzip stream. The state would need to be reset to continue decoding a subsequent gzip stream.

_Cz_inflateMark

Returns two values, one in the lower 16 bits of the return value, and the other in the remaining upper bits, obtained by shifting the return value down 16 bits. If the upper value is -1 and the lower value is zero, then inflate() is currently decoding information outside of a block. If the upper value is -1 and the lower value is non-zero, then inflate is in the middle of a stored block, with the lower value equaling the number of bytes from the input remaining to copy. If the upper value is not -1, then it is the number of bits back from the current bit position in the input of the code (literal or length/distance pair) currently being processed. In that case the lower value is the number of bytes already emitted for that code.

A code is being processed if inflate is waiting for more input to complete decoding of the code, or if it has completed decoding but is waiting for more output space to write the literal or match data.

inflateMark() is used to mark locations in the input data for random access, which may be at bit positions, and to note those cases where the output of a code may span boundaries of random access blocks. The current location in the input stream can be determined from avail_in and data_type as noted in the description for the Z_BLOCK flush parameter for inflate.

_Cz_inflatePrime

This function inserts bits in the inflate input stream. The intent is that this function is used to start inflating at a bit position in the middle of a byte. The provided bits will be used before any bytes are used from next_in. This function should only be used with raw inflate, and should be used before the first inflate() call after inflateInit2() or inflateReset(). bits must be less than or equal to 16, and that many of the least significant bits of value will be inserted in the input.

If bits is negative, then the input stream bit buffer is emptied. Then inflatePrime() can be called again to put bits in the buffer. This is used to clear out bits leftover after feeding inflate a block description prior to feeding inflate codes.

_Cz_inflateReset

This function is equivalent to inflateEnd followed by inflateInit, but does not free and reallocate the internal decompression state. The stream will keep attributes that may have been set by inflateInit2.

_Cz_inflateReset2

This function is the same as inflateReset, but it also permits changing the wrap and window size requests. The windowBits parameter is interpreted the same as it is for inflateInit2. If the window size is changed, then the memory allocated for the window is freed, and the window will be reallocated by inflate() if needed.

_Cz_inflateSetDictionary

Initializes the decompression dictionary from the given uncompressed byte sequence. This function must be called immediately after a call of inflate, if that call returned Z_NEED_DICT. The dictionary chosen by the compressor can be determined from the Adler-32 value returned by that call of inflate. The compressor and decompressor must use exactly the same dictionary (see deflateSetDictionary). For raw inflate, this function can be called at any time to set the dictionary. If the provided dictionary is smaller than the window and there is already data in the window, then the provided dictionary will amend what's there. The application must insure that the dictionary that was used for compression is provided.

_Cz_inflateSync

Skips invalid compressed data until a possible full flush point (see above for the description of deflate with Z_FULL_FLUSH) can be found, or until all available input is skipped. No output is provided.

inflateSync searches for a 00 00 FF FF pattern in the compressed data. All full flush points have this pattern, but not all occurrences of this pattern are full flush points.

_Cz_uncompress

Decompresses the source buffer into the destination buffer. sourceLen is the byte length of the source buffer. Upon entry, destLen is the total size of the destination buffer, which must be large enough to hold the entire uncompressed data. (The size of the uncompressed data must have been saved previously by the compressor and transmitted to the decompressor by some mechanism outside the scope of this compression library.) Upon exit, destLen is the actual size of the uncompressed data.

_Cz_uncompress2

Same as uncompress, except that sourceLen is a pointer, where the length of the source is *sourceLen. On return, *sourceLen is the number of source bytes consumed.

_Cz_zlibVersion

The application can compare zlibVersion and ZLIB_VERSION for consistency. If the first character differs, the library code actually used is not compatible with the zlib.h header file used by the application. This check is automatically made by deflateInit and inflateInit.

daylight

DecodeBase64

Decodes base64 ascii representation to binary.

This supports the following alphabets:

• ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/
• ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_

DecodeLatin1

Decodes ISO-8859-1 to UTF-8.

_DescribeClockName

Describes clock_gettime() clock argument.

_DescribeRlimitName

Describes setrlimit() / getrlimit() argument.

_DescribeSchedParam

Describes clock_gettime() clock argument.

_DescribeSchedPolicy

Describes clock_gettime() clock argument.

_DescribeSiCode

Returns symbolic name for siginfo::si_code value.

_DescribeSleepFlags

Describes clock_nanosleep() flags argument.

_DescribeSockLevel

Describes setsockopt() level arguments.

_DescribeSockOptname

Describes setsockopt() optname arguments.

difftime

dirname

Returns directory portion of path, e.g.

path     │ dirname() │ basename()
─────────────────────────────────
.        │ .         │ .
..       │ .         │ ..
/        │ /         │ /
usr      │ .         │ usr
/usr/    │ /         │ usr
/usr/lib │ /usr      │ lib

div

Divides integers yielding numerator and denominator.

__divmodti4

Divides 128-bit signed integers w/ remainder.

__divti3

Divides 128-bit signed integers.

djbsort

D.J. Bernstein's outrageously fast integer sorting algorithm.

djbsort_avx2

D.J. Bernstein's outrageously fast integer sorting algorithm.

dl_iterate_phdr

dlmalloc_requires_more_vespene_gas

Acquires more system memory for dlmalloc.

dn_comp

dn_skipname

double_conversion_notice

dprintf

Formats string directly to file descriptor.

drand48

dup

Duplicates file descriptor.

The O_CLOEXEC flag shall be cleared from the resulting file descriptor; see dup3() to preserve it.

One use case for duplicating file descriptors is to be able to reassign an open()'d file or pipe() to the stdio of an executed subprocess. On Windows, in order for this to work, the subprocess needs to be a Cosmopolitan program that has socket() linked.

Only small programs should duplicate sockets. That's because this implementation uses DuplicateHandle() on Windows, which Microsoft says might cause its resources to leak internally. Thus it likely isn't a good idea to design a server that does it a lot and lives a long time, without contributing a patch to this implementation.

duplocale

E2BIG

Argument list too long.

EACCES

Permission denied.

eaccess

Performs access() check using effective user/group id.

EADDRINUSE

Address already in use.

EADDRNOTAVAIL

Address not available.

EAFNOSUPPORT

Address family not supported.

EAGAIN

Resource temporarily unavailable (e.g. SO_RCVTIMEO expired, too many processes, too much memory locked, read or write with O_NONBLOCK needs polling, etc.).

EALREADY

Connection already in progress.

EBADF

Bad file descriptor.

EBADMSG

Bad message.

EBUSY

Device or resource busy.

ECANCELED

Operation canceled.

ECHILD

No child process.

ECONNABORTED

Connection reset before accept.

ECONNREFUSED

Connection refused error.

ECONNRESET

Connection reset by client.

EDEADLK

Resource deadlock avoided.

EDESTADDRREQ

Destination address required.

EDOM

Mathematics argument out of domain of function.

EDQUOT

Disk quota exceeded.

EEXIST

File exists.

EFAULT

Pointer passed to system call that would otherwise segfault.

EFBIG

File too large.

_EfiPostboot

Start the Cosmopolitan runtime after exiting UEFI Boot Services.

EFTYPE

Inappropriate file type or format.

EHOSTDOWN

Host down error.

EHOSTUNREACH

Host unreachable error.

EIDRM

Identifier removed.

EILSEQ

Unicode decoding error.

EINPROGRESS

Operation already in progress.

EINTR

The greatest of all errnos.

EINVAL

Invalid argument.

EIO

Unix consensus.

EISCONN

Socket is connected.

EISDIR

Is a a directory.

ElevateSeDebugPrivilege

ELOOP

Too many levels of symbolic links.

EMEDIUMTYPE

Wrong medium type.

EMFILE

Too many open files.

EMLINK

Too many links.

EMSGSIZE

Message too long.

EMULTIHOP

Multihop attempted.

ENAMETOOLONG

Filename too long.

EncodeBase64

Encodes binary to base64 ascii representation.

EncodeHttpHeaderValue

Encodes HTTP header value.

This operation involves the following:

1. Trim whitespace.
2. Turn UTF-8 into ISO-8859-1.
3. Make sure no C0 or C1 control codes are present (except tab).

If the input value isn't thompson-pike encoded then this implementation will fall back to latin1 in most cases.

EncodeLatin1

Encodes UTF-8 to ISO-8859-1.

EncodeUrl

Encodes URL.

encrypt

endgrent

Closes group database.

endhostent

endmntent

endprotoent

endpwent

Closes global handle to password database.

endservent

endswith

Returns true if s has suffix.

endswith16

Returns true if s has suffix.

endutxent

ENETDOWN

Network is down.

ENETRESET

Connection reset by network.

ENETUNREACH

Host is unreachable.

ENFILE

Too many open files in system.

ENOBUFS

No buffer space available.

ENODATA

No data.

ENODEV

No such device.

ENOENT

No such file or directory.

ENOEXEC

Exec format error.

ENOLCK

No locks available.

ENOLINK

Link severed.

ENOMEDIUM

No medium found.

ENOMEM

We require more vespene gas.

ENOMSG

No message error.

ENONET

No network.

ENOPROTOOPT

Protocol not available.

ENOSPC

No space left on device.

ENOSR

Out of streams resources.

ENOSTR

No string.

ENOSYS

System call unavailable.

ENOTBLK

Block device required.

ENOTCONN

Socket is not connected.

ENOTDIR

Not a directory.

ENOTEMPTY

Directory not empty.

ENOTRECOVERABLE

State not recoverable.

ENOTSOCK

Not a socket.

ENOTSUP

Operation not supported.

ENOTTY

Inappropriate i/o control operation.

ENXIO

No such device or address.

EOPNOTSUPP

Socket operation not supported.

EOVERFLOW

Overflow error.

EOWNERDEAD

Owner died.

EPERM

Operation not permitted.

EPFNOSUPPORT

Protocol family not supported.

EPIPE

Broken pipe.

EPROTO

Protocol error.

EPROTONOSUPPORT

Protocol not supported.

EPROTOTYPE

Protocol wrong type for socket.

erand48

ERANGE

Result too large.

EREMOTE

Remote error.

ERESTART

Please restart syscall.

erf

Returns error function of x.

Highest measured error is 1.01 ULPs at 0x1.39956ac43382fp+0.

erff

EROFS

Read-only filesystem.

__errno

Global variable for last error.

The system call wrappers update this with WIN32 error codes. Unlike traditional libraries, Cosmopolitan error codes are defined as variables. By convention, system calls and other functions do not update this variable when nothing's broken.

__errno_location

Returns address of errno variable.

EscapeFragment

Escapes URL fragment.

EscapeHost

Escapes URL host or registry name.

EscapeHtml

Escapes HTML entities.

EscapeIp

Escapes URL IP-literal.

This is the same as EscapeHost except colon is permitted.

EscapeJsStringLiteral

Escapes UTF-8 data for JavaScript or JSON string literal.

HTML entities and forward slash are escaped too for added safety. Single quote (') is \uxxxx-encoded for consistency, as it's allowed in JavaScript, but not in JSON strings.

We assume the UTF-8 is well-formed and can be represented as UTF-16. Things that can't be decoded will fall back to binary. Things that can't be encoded will use invalid codepoint markers. This function is agnostic to numbers that have been used with malicious intent in the past under buggy software. Noncanonical encodings such as overlong NUL are canonicalized as NUL. Therefore it isn't necessary to say EscapeJsStringLiteral(Underlong(𝑥)) since EscapeJsStringLiteral(𝑥) will do the same thing.

EscapeParam

Escapes query/form name/parameter.

EscapePass

Escapes URL password.

EscapePath

Escapes URL path.

This is the same as EscapePathSegment() except slash is allowed.

EscapeSegment

Escapes URL path segment.

Please note this will URI encode the slash character. That's because segments are the labels between the slashes in a path.

EscapeUrl

Escapes URL component using generic table.

This function is agnostic to the underlying charset. Always using UTF-8 is a good idea.

EscapeUrlView

Escapes URL component using generic table w/ stpcpy() api.

EscapeUser

Escapes URL user name.

ESHUTDOWN

Cannot send after transport endpoint shutdown.

ESOCKTNOSUPPORT

Socket type not supported.

ESPIPE

Invalid seek.

ESRCH

No such process.

ESTALE

Stale error.

ETIME

Timer expired.

ETIMEDOUT

Connection timed out.

ETOOMANYREFS

Too many references: cannot splice.

ETXTBSY

Won't open executable that's executing in write mode.

euidaccess

Performs access() check using effective user/group id.

EUSERS

Too many users.

EXDEV

Improper link.

execlp

Executes program, with PATH search and current environment.

The current process is replaced with the executed one.

execve

Replaces current process with program.

On Windows, argv and envp can't contain binary strings. They need to be valid UTF-8 in order to round-trip the WIN32 API, without being corrupted.

On Windows, only file descriptors 0, 1 and 2 can be passed to a child process in such a way that allows them to be automatically discovered when the child process initializes. Cosmpolitan currently treats your other file descriptors as implicitly O_CLOEXEC.

execvpe

Executes program, with path environment search.

This function is a wrapper of the execve() system call that does path resolution. The PATH environment variable is taken from your global environ rather than the envp argument.

exit

Exits process with grace.

This calls functions registered by atexit() before terminating the current process, and any associated threads. It also calls all the legacy linker registered destructors in reversed order

This implementation allows exit() to be called recursively via atexit() handlers.

exp10

Returns 10ˣ.

The largest observed error is ~0.513 ULP.

exp10f

Returns 10ˣ.

exp2l

Returns 2^𝑥.

expm1

Returns 𝑒^𝑥-𝟷.

fabs

Returns absolute value of floating point number.

fabsf

Returns absolute value of floating point number.

fabsl

Returns absolute value of floating point number.

faccessat

Checks if effective user can access path in particular ways.

fadvise

Drops hints to O/S about intended I/O behavior.

It makes a huge difference. For example, when copying a large file, it can stop the system from persisting GBs of useless memory content.

__fbufsize

Returns capacity of stdio stream buffer.

fchdir

Sets current directory based on file descriptor.

This does *not* update the PWD environment variable.

fchmod

Changes file permissions via open()'d file descriptor.

fchmodat

Changes permissions on file, e.g.:

CHECK_NE(-1, fchmodat(AT_FDCWD, "foo/bar.txt", 0644));
CHECK_NE(-1, fchmodat(AT_FDCWD, "o/default/program", 0755));
CHECK_NE(-1, fchmodat(AT_FDCWD, "privatefolder/", 0700));

fchown

Changes owner and/or group of file, via open()'d descriptor.

fchownat

Changes owner and/or group of path.

fclose

Closes standard i/o stream and its underlying thing.

fcntl

Does things with file descriptor, e.g.

CHECK_NE(-1, fcntl(fd, F_SETFD, FD_CLOEXEC));

This function lets you duplicate file descriptors without running into an edge case where they take over stdio handles:

CHECK_GE((newfd = fcntl(oldfd, F_DUPFD,         3)), 3);
CHECK_GE((newfd = fcntl(oldfd, F_DUPFD_CLOEXEC, 3)), 3);

This function implements file record locking, which lets independent processes (and on Linux 3.15+, threads too!) lock arbitrary ranges associated with a file. See test/libc/calls/lock_test.c and other locking related tests in that folder.

On Windows, the Cosmopolitan Libc polyfill for POSIX advisory locks only implements enough of its nuances to support SQLite's needs. Some possibilities, e.g. punching holes in lock, will raise ENOTSUP.

fdexists

Returns true if file descriptor exists.

This function is equivalent to:

fcntl(fd, F_GETFL) != -1

Except it won't clobber errno and has minimal linkage.

fdim

Returns positive difference.

fdimf

Returns positive difference.

fdiml

Returns positive difference.

fdlibm_notice

fdopen

Allocates stream object for already-opened file descriptor.

feclearexcept

Clears floating point exception status, e.g.

feclearexcept(FE_ALL_EXCEPT);

fedisableexcept

Disables floating point exception trapping, e.g.

feenableexcept(FE_INVALID | FE_DIVBYZERO |
               FE_OVERFLOW | FE_UNDERFLOW);

When trapping is enabled, something should handle SIGFPE. Calling ShowCrashReports() at startup will install a generic handler with backtraces and the symbol of the si->si_code which UNIX defines

• FPE_INTOVF: integer overflow
• FPE_INTDIV: integer divide by zero
• FPE_FLTDIV: floating point divide by zero
• FPE_FLTOVF: floating point overflow
• FPE_FLTUND: floating point underflow
• FPE_FLTRES: floating point inexact
• FPE_FLTINV: invalid floating point operation
• FPE_FLTSUB: subscript out of range

It's important to not use the -ffast-math or -Ofast flags when compiling code that needs to be debugged. Using -fsignaling-nans will also help, since GCC doesn't enable that by default.

feenableexcept

Enables floating point exception trapping, e.g.

feenableexcept(FE_INVALID | FE_DIVBYZERO |
               FE_OVERFLOW | FE_UNDERFLOW);

When trapping is enabled, something should handle SIGFPE. Calling ShowCrashReports() at startup will install a generic handler with backtraces and the symbol of the si->si_code which UNIX defines

• FPE_INTOVF: integer overflow
• FPE_INTDIV: integer divide by zero
• FPE_FLTDIV: floating point divide by zero
• FPE_FLTOVF: floating point overflow
• FPE_FLTUND: floating point underflow
• FPE_FLTRES: floating point inexact
• FPE_FLTINV: invalid floating point operation
• FPE_FLTSUB: subscript out of range

It's important to not use the -ffast-math or -Ofast flags when compiling code that needs to be debugged. Using -fsignaling-nans will also help, since GCC doesn't enable that by default.

feholdexcept

Saves floating-point environment and clears current exceptions.

feof

Returns true if stream is in end-of-file state.

feof_unlocked

Returns true if stream is in end-of-file state.

ferror

Returns nonzero if stream is in error state.

ferror_unlocked

Returns nonzero if stream is in error state.

fesetround

Sets rounding mode.

This configures the x87 FPU as well as SSE.

fetestexcept

feupdateenv

Restores floating point environment and raises exceptions.

fflush

Blocks until data from stream buffer is written out.

ffs

Finds lowest set bit in word.

ffsl

Finds lowest set bit in word.

fgetc

Reads byte from stream.

fgetc_unlocked

Reads byte from stream.

fgetln

Retrieves line from stream, e.g.

char *line;
while ((line = chomp(fgetln(stdin, 0)))) {
  printf("%s\n", line);
}

The returned memory is owned by the stream. It'll be reused when fgetln() is called again. It's free()'d upon fclose() / fflush()

When reading from the console on Windows in ICANON mode, the returned line will end with \r\n rather than \n.

fgetpos

fgets

Reads line from stream.

This function is similar to getline() except it'll truncate lines exceeding size. The line ending marker is included and may be removed using chomp().

When reading from the console on Windows in ICANON mode, the returned line will end with \r\n rather than \n.

fgets_unlocked

Reads line from stream.

This function is similar to getline() except it'll truncate lines exceeding size. The line ending marker is included and may be removed using chomp().

fgetwc

Reads UTF-8 character from stream.

fgetwc_unlocked

Reads UTF-8 character from stream.

fgetws

Reads UTF-8 content from stream into UTF-32 buffer.

This function is similar to getline() except it'll truncate lines exceeding size. The line ending marker is included and may be removed using chomp().

fgetws_unlocked

Reads UTF-8 content from stream into UTF-32 buffer.

This function is similar to getline() except it'll truncate lines exceeding size. The line ending marker is included and may be removed using chomp().

fileno

Returns file descriptor associated with stream.

fileno_unlocked

Returns file descriptor associated with stream.

FindComBinary

Returns path of binary without debug information, or null.

FindContentType

Returns Content-Type for file extension.

FindDebugBinary

Returns path of binary with the debug information, or null.

You can specify the COMDBG environment variable, with the path of the debug binary, in case the automatic heuristics fail. What we look for is GetProgramExecutableName() with ".dbg", ".com.dbg", etc. appended.

FindElfSectionByName

Returns ELF section header for section with name.

finite

Returns nonzero if 𝑥 isn't INFINITY or NAN.

finitef

Returns nonzero if 𝑥 isn't INFINITY or NAN.

finitel

Returns nonzero if 𝑥 isn't INFINITY or NAN.

__flbf

Returns nonzero if stream is line buffered.

flock

Acquires lock on file.

Please note multiple file descriptors means multiple locks.