lighttpd1.4/src/chunk.c

608 lines
12 KiB
C
Raw Normal View History

/**
* the network chunk-API
*
*
*/
#include "chunk.h"
#include "base.h"
#include "log.h"
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <stdlib.h>
#include <fcntl.h>
#include <unistd.h>
#include <stdio.h>
#include <errno.h>
#include <string.h>
chunkqueue *chunkqueue_init(void) {
chunkqueue *cq;
cq = calloc(1, sizeof(*cq));
cq->first = NULL;
cq->last = NULL;
cq->unused = NULL;
return cq;
}
static chunk *chunk_init(void) {
chunk *c;
c = calloc(1, sizeof(*c));
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c->type = MEM_CHUNK;
c->mem = buffer_init();
c->file.name = buffer_init();
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c->file.start = c->file.length = c->file.mmap.offset = 0;
c->file.fd = -1;
c->file.mmap.start = MAP_FAILED;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c->file.mmap.length = 0;
c->file.is_temp = 0;
c->offset = 0;
c->next = NULL;
return c;
}
static void chunk_reset(chunk *c) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (NULL == c) return;
c->type = MEM_CHUNK;
buffer_reset(c->mem);
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (c->file.is_temp && !buffer_string_is_empty(c->file.name)) {
unlink(c->file.name->ptr);
}
buffer_reset(c->file.name);
if (c->file.fd != -1) {
close(c->file.fd);
c->file.fd = -1;
}
if (MAP_FAILED != c->file.mmap.start) {
munmap(c->file.mmap.start, c->file.mmap.length);
c->file.mmap.start = MAP_FAILED;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c->file.start = c->file.length = c->file.mmap.offset = 0;
c->file.mmap.length = 0;
c->file.is_temp = 0;
c->offset = 0;
c->next = NULL;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
static void chunk_free(chunk *c) {
if (NULL == c) return;
chunk_reset(c);
buffer_free(c->mem);
buffer_free(c->file.name);
free(c);
}
void chunkqueue_free(chunkqueue *cq) {
chunk *c, *pc;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (NULL == cq) return;
for (c = cq->first; c; ) {
pc = c;
c = c->next;
chunk_free(pc);
}
for (c = cq->unused; c; ) {
pc = c;
c = c->next;
chunk_free(pc);
}
free(cq);
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
static void chunkqueue_push_unused_chunk(chunkqueue *cq, chunk *c) {
force_assert(NULL != cq && NULL != c);
/* keep at max 4 chunks in the 'unused'-cache */
if (cq->unused_chunks > 4) {
chunk_free(c);
} else {
chunk_reset(c);
c->next = cq->unused;
cq->unused = c;
cq->unused_chunks++;
}
}
static chunk *chunkqueue_get_unused_chunk(chunkqueue *cq) {
chunk *c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
force_assert(NULL != cq);
/* check if we have a unused chunk */
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (0 == cq->unused) {
c = chunk_init();
} else {
/* take the first element from the list (a stack) */
c = cq->unused;
cq->unused = c->next;
c->next = NULL;
cq->unused_chunks--;
}
return c;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
static void chunkqueue_prepend_chunk(chunkqueue *cq, chunk *c) {
c->next = cq->first;
cq->first = c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (NULL == cq->last) {
cq->last = c;
}
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
static void chunkqueue_append_chunk(chunkqueue *cq, chunk *c) {
if (cq->last) {
cq->last->next = c;
}
cq->last = c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (NULL == cq->first) {
cq->first = c;
}
}
void chunkqueue_reset(chunkqueue *cq) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
chunk *cur = cq->first;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
cq->first = cq->last = NULL;
while (NULL != cur) {
chunk *next = cur->next;
chunkqueue_push_unused_chunk(cq, cur);
cur = next;
}
cq->bytes_in = 0;
cq->bytes_out = 0;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_append_file(chunkqueue *cq, buffer *fn, off_t offset, off_t len) {
chunk *c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (0 == len) return;
c = chunkqueue_get_unused_chunk(cq);
c->type = FILE_CHUNK;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
buffer_copy_buffer(c->file.name, fn);
c->file.start = offset;
c->file.length = len;
c->offset = 0;
chunkqueue_append_chunk(cq, c);
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_append_buffer(chunkqueue *cq, buffer *mem) {
chunk *c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (buffer_string_is_empty(mem)) return;
c = chunkqueue_get_unused_chunk(cq);
c->type = MEM_CHUNK;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
force_assert(NULL != c->mem);
buffer_move(c->mem, mem);
chunkqueue_append_chunk(cq, c);
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_prepend_buffer(chunkqueue *cq, buffer *mem) {
chunk *c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (buffer_string_is_empty(mem)) return;
c = chunkqueue_get_unused_chunk(cq);
c->type = MEM_CHUNK;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
force_assert(NULL != c->mem);
buffer_move(c->mem, mem);
chunkqueue_prepend_chunk(cq, c);
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_append_mem(chunkqueue *cq, const char * mem, size_t len) {
chunk *c;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (0 == len) return;
c = chunkqueue_get_unused_chunk(cq);
c->type = MEM_CHUNK;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
buffer_copy_string_len(c->mem, mem, len);
chunkqueue_append_chunk(cq, c);
}
void chunkqueue_get_memory(chunkqueue *cq, char **mem, size_t *len, size_t min_size, size_t alloc_size) {
static const size_t REALLOC_MAX_SIZE = 256;
chunk *c;
buffer *b;
char *dummy_mem;
size_t dummy_len;
force_assert(NULL != cq);
if (NULL == mem) mem = &dummy_mem;
if (NULL == len) len = &dummy_len;
/* default values: */
if (0 == min_size) min_size = 1024;
if (0 == alloc_size) alloc_size = 4096;
if (alloc_size < min_size) alloc_size = min_size;
if (NULL != cq->last && MEM_CHUNK == cq->last->type) {
size_t have;
b = cq->last->mem;
have = buffer_string_space(b);
/* unused buffer: allocate space */
if (buffer_string_is_empty(b)) {
buffer_string_prepare_copy(b, alloc_size);
have = buffer_string_space(b);
}
/* if buffer is really small just make it bigger */
else if (have < min_size && b->size <= REALLOC_MAX_SIZE) {
size_t new_size = b->used + min_size, append;
if (new_size < alloc_size) new_size = alloc_size;
append = new_size - b->used;
if (append >= min_size) {
buffer_string_prepare_append(b, append);
have = buffer_string_space(b);
}
}
/* return pointer into existing buffer if large enough */
if (have >= min_size) {
*mem = b->ptr + buffer_string_length(b);
*len = have;
return;
}
}
/* allocate new chunk */
c = chunkqueue_get_unused_chunk(cq);
c->type = MEM_CHUNK;
chunkqueue_append_chunk(cq, c);
b = c->mem;
buffer_string_prepare_append(b, alloc_size);
*mem = b->ptr + buffer_string_length(b);
*len = buffer_string_space(b);
}
void chunkqueue_use_memory(chunkqueue *cq, size_t len) {
buffer *b;
force_assert(NULL != cq);
force_assert(NULL != cq->last && MEM_CHUNK == cq->last->type);
b = cq->last->mem;
force_assert(b->used > 0);
force_assert(len <= buffer_string_space(b));
if (len > 0) {
b->used += len;
b->ptr[b->used - 1] = '\0';
} else if (buffer_string_is_empty(b)) {
/* unused buffer: can't remove chunk easily from
* end of list, so just reset the buffer
*/
buffer_reset(b);
}
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_set_tempdirs(chunkqueue *cq, array *tempdirs) {
force_assert(NULL != cq);
cq->tempdirs = tempdirs;
}
void chunkqueue_steal(chunkqueue *dest, chunkqueue *src, off_t len) {
while (len > 0) {
chunk *c = src->first;
off_t clen = 0, use;
if (NULL == c) break;
switch (c->type) {
case MEM_CHUNK:
clen = buffer_string_length(c->mem);
break;
case FILE_CHUNK:
clen = c->file.length;
break;
}
force_assert(clen >= c->offset);
clen -= c->offset;
use = len >= clen ? clen : len;
src->bytes_out += use;
dest->bytes_in += use;
len -= use;
if (0 == clen) {
/* drop empty chunk */
src->first = c->next;
if (c == src->last) src->last = NULL;
chunkqueue_push_unused_chunk(src, c);
continue;
}
if (use == clen) {
/* move complete chunk */
src->first = c->next;
if (c == src->last) src->last = NULL;
chunkqueue_append_chunk(dest, c);
continue;
}
/* partial chunk with length "use" */
switch (c->type) {
case MEM_CHUNK:
chunkqueue_append_mem(dest, c->mem->ptr + c->offset, use);
break;
case FILE_CHUNK:
/* tempfile flag is in "last" chunk after the split */
chunkqueue_append_file(dest, c->file.name, c->file.start + c->offset, use);
break;
}
c->offset += use;
force_assert(0 == len);
}
}
static chunk *chunkqueue_get_append_tempfile(chunkqueue *cq) {
chunk *c;
buffer *template = buffer_init_string("/var/tmp/lighttpd-upload-XXXXXX");
int fd;
if (cq->tempdirs && cq->tempdirs->used) {
size_t i;
/* we have several tempdirs, only if all of them fail we jump out */
for (i = 0; i < cq->tempdirs->used; i++) {
data_string *ds = (data_string *)cq->tempdirs->data[i];
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
buffer_copy_buffer(template, ds->value);
buffer_append_slash(template);
buffer_append_string_len(template, CONST_STR_LEN("lighttpd-upload-XXXXXX"));
if (-1 != (fd = mkstemp(template->ptr))) break;
}
} else {
fd = mkstemp(template->ptr);
}
if (-1 == fd) {
buffer_free(template);
return NULL;
}
c = chunkqueue_get_unused_chunk(cq);
c->type = FILE_CHUNK;
c->file.fd = fd;
c->file.is_temp = 1;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
buffer_copy_buffer(c->file.name, template);
c->file.length = 0;
chunkqueue_append_chunk(cq, c);
buffer_free(template);
return c;
}
static int chunkqueue_append_to_tempfile(server *srv, chunkqueue *dest, const char *mem, size_t len) {
chunk *dst_c = NULL;
ssize_t written;
/* copy everything to max 1Mb sized tempfiles */
/*
* if the last chunk is
* - smaller than 1Mb (size < 1Mb)
* - not read yet (offset == 0)
* -> append to it
* otherwise
* -> create a new chunk
*
* */
if (NULL != dest->last
&& FILE_CHUNK != dest->last->type
&& dest->last->file.is_temp
&& -1 != dest->last->file.fd
&& 0 == dest->last->offset) {
/* ok, take the last chunk for our job */
dst_c = dest->last;
if (dest->last->file.length >= 1 * 1024 * 1024) {
/* the chunk is too large now, close it */
if (-1 != dst_c->file.fd) {
close(dst_c->file.fd);
dst_c->file.fd = -1;
}
dst_c = chunkqueue_get_append_tempfile(dest);
}
} else {
dst_c = chunkqueue_get_append_tempfile(dest);
}
if (NULL == dst_c) {
/* we don't have file to write to,
* EACCES might be one reason.
*
* Instead of sending 500 we send 413 and say the request is too large
*/
log_error_write(srv, __FILE__, __LINE__, "sbs",
"denying upload as opening temp-file for upload failed:",
dst_c->file.name, strerror(errno));
return -1;
}
if (0 > (written = write(dst_c->file.fd, mem, len)) || (size_t) written != len) {
/* write failed for some reason ... disk full ? */
log_error_write(srv, __FILE__, __LINE__, "sbs",
"denying upload as writing to file failed:",
dst_c->file.name, strerror(errno));
close(dst_c->file.fd);
dst_c->file.fd = -1;
return -1;
}
dst_c->file.length += len;
return 0;
}
int chunkqueue_steal_with_tempfiles(server *srv, chunkqueue *dest, chunkqueue *src, off_t len) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
while (len > 0) {
chunk *c = src->first;
off_t clen = 0, use;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
if (NULL == c) break;
switch (c->type) {
case MEM_CHUNK:
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
clen = buffer_string_length(c->mem);
break;
case FILE_CHUNK:
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
clen = c->file.length;
break;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
}
force_assert(clen >= c->offset);
clen -= c->offset;
use = len >= clen ? clen : len;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
src->bytes_out += use;
dest->bytes_in += use;
len -= use;
if (0 == clen) {
/* drop empty chunk */
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
src->first = c->next;
if (c == src->last) src->last = NULL;
chunkqueue_push_unused_chunk(src, c);
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
continue;
}
if (FILE_CHUNK == c->type) {
if (use == clen) {
/* move complete chunk */
src->first = c->next;
if (c == src->last) src->last = NULL;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
chunkqueue_append_chunk(dest, c);
} else {
/* partial chunk with length "use" */
/* tempfile flag is in "last" chunk after the split */
chunkqueue_append_file(dest, c->file.name, c->file.start + c->offset, use);
c->offset += use;
force_assert(0 == len);
}
continue;
}
/* store "use" bytes from memory chunk in tempfile */
if (0 != chunkqueue_append_to_tempfile(srv, dest, c->mem->ptr + c->offset, use)) {
/* undo counters */
src->bytes_out -= use;
dest->bytes_in -= use;
return -1;
}
c->offset += use;
if (use == clen) {
/* finished chunk */
src->first = c->next;
if (c == src->last) src->last = NULL;
chunkqueue_push_unused_chunk(src, c);
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
}
return 0;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
off_t chunkqueue_length(chunkqueue *cq) {
off_t len = 0;
chunk *c;
for (c = cq->first; c; c = c->next) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
off_t c_len = 0;
switch (c->type) {
case MEM_CHUNK:
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c_len = buffer_string_length(c->mem);
break;
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
case FILE_CHUNK:
c_len = c->file.length;
break;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
force_assert(c_len >= c->offset);
len += c_len - c->offset;
}
return len;
}
int chunkqueue_is_empty(chunkqueue *cq) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
return NULL == cq->first;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
void chunkqueue_remove_finished_chunks(chunkqueue *cq) {
chunk *c;
for (c = cq->first; c; c = cq->first) {
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
off_t c_len = 0;
switch (c->type) {
case MEM_CHUNK:
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c_len = buffer_string_length(c->mem);
break;
case FILE_CHUNK:
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
c_len = c->file.length;
break;
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
2015-02-08 12:37:10 +00:00
force_assert(c_len >= c->offset);