hyung-hwan/qse
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

enabled qse_http_feed() to get trailing headers after 0

Browse Source
This commit is contained in:
hyung-hwan
2010-11-27 09:00:57 +00:00
parent 84376d6d92
commit 5b4845db55
11 changed files with 339 additions and 183 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
qse/doc/Doxyfile.in
View File

@ -186,7 +186,7 @@ ALIASES =
# For instance, some of the names that are used will be different. The list
# of all members will be omitted, etc.
OPTIMIZE_OUTPUT_FOR_C = NO
OPTIMIZE_OUTPUT_FOR_C = YES
# Set the OPTIMIZE_OUTPUT_JAVA tag to YES if your project consists of Java
# sources only. Doxygen will then generate output that is more tailored for
@ -1397,12 +1397,12 @@ CLASS_GRAPH = YES
# indirect implementation dependencies (inheritance, containment, and
# class references variables) of the class with other documented classes.
COLLABORATION_GRAPH = YES
COLLABORATION_GRAPH = NO
# If the GROUP_GRAPHS and HAVE_DOT tags are set to YES then doxygen
# will generate a graph for groups, showing the direct groups dependencies
GROUP_GRAPHS = YES
GROUP_GRAPHS = NO
# If the UML_LOOK tag is set to YES doxygen will generate inheritance and
# collaboration diagrams in a style similar to the OMG's Unified Modeling

2
qse/doc/page/awk.doc
View File

@ -1,4 +1,4 @@
/** @page awk AWK
/** @defgroup awk AWK
@section awk_intro INTRODUCTION

2
qse/doc/page/cut.doc
View File

@ -1,4 +1,4 @@
/** @page cut TEXT CUTTER
/** @defgroup cut TEXT CUTTER
Text Cutting Utility
*/

2
qse/doc/page/sed.doc
View File

@ -1,4 +1,4 @@
/** @page sed STREAM EDITOR
/** @defgroup sed Stream Editor
@section sed_contents CONTENTS
- \ref sed_intro

13
qse/include/qse/cut/cut.h
View File

@ -37,7 +37,8 @@
*/
/** @struct qse_cut_t
* The qes_cut_t type defines a text cutter.
* The qse_cut_t type defines a text cutter. The details are hidden as it is
* a large complex structure vulnerable to unintended changes.
*/
typedef struct qse_cut_t qse_cut_t;
@ -110,7 +111,7 @@ typedef enum qse_cut_io_cmd_t qse_cut_io_cmd_t;
*/
struct qse_cut_io_arg_t
{
void* handle; /**< I/O handle */
void* handle; /**< I/O handle */
};
typedef struct qse_cut_io_arg_t qse_cut_io_arg_t;
@ -133,12 +134,12 @@ extern "C" {
QSE_DEFINE_COMMON_FUNCTIONS (cut)
/**
* The qse_cut_open() function creates a text cutter object.
* @return A pointer to a text cutter on success, QSE_NULL on failure
* The qse_cut_open() function creates a text cutter.
* @return A pointer to a text cutter on success, #QSE_NULL on failure
*/
qse_cut_t* qse_cut_open (
qse_mmgr_t* mmgr, /**< a memory manager */
qse_size_t xtn /**< the size of extension in bytes */
qse_mmgr_t* mmgr, /**< memory manager */
qse_size_t xtnsize /**< extension size in bytes */
);
/**

17
qse/include/qse/sed/Sed.hpp
View File

@ -1,5 +1,5 @@
/*
* $Id: Sed.hpp 321 2009-12-21 12:44:33Z hyunghwan.chung $
* $Id: Sed.hpp 373 2010-11-26 15:00:57Z hyunghwan.chung $
*
Copyright 2006-2009 Chung, Hyung-Hwan.
This file is part of QSE.
@ -24,6 +24,18 @@
#include <qse/cmn/Mmged.hpp>
#include <qse/sed/sed.h>
/**
* @defgroup sed_cxx C++
* @ingroup sed
* @{
*/
/**
* @defgroup sed_cxx_core Core interface
* @ingroup sed_cxx
* @{
*/
/** @file
* Stream Editor
*/
@ -310,4 +322,7 @@ private:
QSE_END_NAMESPACE(QSE)
/////////////////////////////////
/* @} */
/* @} */
#endif

24
qse/include/qse/sed/sed.h
View File

@ -1,5 +1,5 @@
/*
* $Id: sed.h 344 2010-08-17 13:15:14Z hyunghwan.chung $
* $Id: sed.h 373 2010-11-26 15:00:57Z hyunghwan.chung $
*
Copyright 2006-2009 Chung, Hyung-Hwan.
This file is part of QSE.
@ -24,6 +24,25 @@
#include <qse/types.h>
#include <qse/macros.h>
/**
* @defgroup sed_c C
* @ingroup sed
* @{
* hello
*/
/**
* @defgroup sed_c_core Core interface
* @ingroup sed_c
* @{
* The sed library provides data types and functions for creating a custom
* stream editor commonly available on platforms. It is a non-interactive
* text editing tool that reads text from an input stream, stores it to
* pattern space, manipulates the pattern space by applying a set of editing
* commands, and writes the pattern space to an output stream. Typically,
* the input and output streams are a console or a file.
*/
/** @file
* A stream editor performs text transformation on a text stream.
*
@ -423,4 +442,7 @@ void qse_sed_setlinnum (
}
#endif
/* @} */
/* @} */
#endif

39
qse/include/qse/sed/std.h
View File

@ -21,10 +21,20 @@
#ifndef _QSE_SED_STD_H_
#define _QSE_SED_STD_H_
/**
* @defgroup sed_c_std Helper functions
* @ingroup sed_c
* @{
* If you don't care about the details of memory management and I/O,
* you can choose to use the helper functions provided here. It is
* a higher-level interface that is easier to use as it implements
* default handlers for I/O and memory management.
*/
#include <qse/sed/sed.h>
/** @file
* This file provides easier-to-use versions of selected API functions
* This file provides easier-to-use interface
* by implementing default handlers for I/O and memory management.
*
* @example sed01.c
@ -38,8 +48,9 @@ extern "C" {
/**
* The qse_sed_openstd() function creates a stream editor with the default
* memory manager and initialized it for other qse_sed_xxxxstd functions.
* @return pointer to a stream editor on success, QSE_NULL on failure.
* memory manager and initializes it. Use qse_getxtnstd(), not qse_getxtn()
* to get the pointer to the extension area.
* @return pointer to a stream editor on success, #QSE_NULL on failure.
*/
qse_sed_t* qse_sed_openstd (
qse_size_t xtnsize /**< extension size in bytes */
@ -48,8 +59,9 @@ qse_sed_t* qse_sed_openstd (
/**
* The qse_sed_openstdwithmmgr() function creates a stream editor with a
* user-defined memory manager. It is equivalent to qse_sed_openstd(),
* except that you can specify your own memory manager.
* @return pointer to a stream editor on success, QSE_NULL on failure.
* except that you can specify your own memory manager.Use qse_getxtnstd(),
* not qse_getxtn() to get the pointer to the extension area.
* @return pointer to a stream editor on success, #QSE_NULL on failure.
*/
qse_sed_t* qse_sed_openstdwithmmgr (
qse_mmgr_t* mmgr, /**< memory manager */
@ -62,7 +74,7 @@ qse_sed_t* qse_sed_openstdwithmmgr (
* created with qse_sed_openstd() and qse_sed_openstdwithmmgr().
*/
void* qse_sed_getxtnstd (
qse_sed_t* sed
qse_sed_t* sed /**< stream editor */
);
/**
@ -71,26 +83,27 @@ void* qse_sed_getxtnstd (
* @return 0 on success, -1 on failure
*/
int qse_sed_compstd (
qse_sed_t* sed,
const qse_char_t* str
qse_sed_t* sed, /**< stream editor */
const qse_char_t* str /**< null-terminated script */
);
/**
* The qse_sed_execstd() function executes the compiled script
* over an input file @a infile and an output file @a outfile.
* If @a infile is QSE_NULL, the standard console input is used.
* If @a outfile is QSE_NULL, the standard console output is used.
* If @a infile is #QSE_NULL, the standard console input is used.
* If @a outfile is #QSE_NULL, the standard console output is used.
* @return 0 on success, -1 on failure
*/
int qse_sed_execstd (
qse_sed_t* sed,
const qse_char_t* infile,
const qse_char_t* outfile
qse_sed_t* sed, /**< stream editor */
const qse_char_t* infile, /**< input file */
const qse_char_t* outfile /**< output file */
);
#ifdef __cplusplus
}
#endif
/* @} */
#endif

1
qse/include/qse/utl/http.h
View File

@ -57,6 +57,7 @@ struct qse_http_t
qse_http_octb_t raw;
qse_http_octb_t con;
qse_http_octb_t tra;
enum
{

27
qse/lib/cut/cut.h
View File

@ -76,32 +76,31 @@ struct qse_cut_t
qse_char_t buf[2048];
qse_size_t len;
int eof;
} out;
} out;
/** data needed for input streams */
struct
{
qse_cut_io_fun_t fun; /**< an input handler */
qse_cut_io_arg_t arg; /**< input handling data */
{
qse_cut_io_fun_t fun; /**< an input handler */
qse_cut_io_arg_t arg; /**< input handling data */
qse_char_t xbuf[1]; /**< a read-ahead buffer */
int xbuf_len; /**< data length in the buffer */
qse_char_t xbuf[1]; /**< a read-ahead buffer */
int xbuf_len; /**< data length in the buffer */
qse_char_t buf[2048]; /**< input buffer */
qse_size_t len; /**< data length in the buffer */
qse_size_t pos; /**< current position in the buffer */
int eof; /**< EOF indicator */
qse_char_t buf[2048]; /**< input buffer */
qse_size_t len; /**< data length in the buffer */
qse_size_t pos; /**< current position in the buffer */
int eof; /**< EOF indicator */
qse_str_t line; /**< pattern space */
qse_size_t num; /**< current line number */
qse_str_t line; /**< pattern space */
qse_size_t num; /**< current line number */
qse_size_t nflds; /**< the number of fields */
qse_size_t cflds; /**< capacity of flds field */
qse_cstr_t sflds[128]; /**< static field buffer */
qse_cstr_t* flds;
int delimited;
} in;
} in;
} e;
};

389
qse/lib/utl/http.c
View File

@ -24,6 +24,8 @@
QSE_IMPLEMENT_COMMON_FUNCTIONS (http)
static const qse_byte_t NUL = '\0';
static QSE_INLINE int is_http_space (qse_char_t c)
{
return QSE_ISSPACE(c) && c != QSE_T('\r') && c != QSE_T('\n');
@ -354,8 +356,12 @@ static QSE_INLINE void clear_request (qse_http_t* http)
* reading the next request */
QSE_MEMSET (&http->req.state, 0, QSE_SIZEOF(http->req.state));
QSE_MEMSET (&http->req.attr, 0, QSE_SIZEOF(http->req.attr));
qse_htb_clear (&http->req.hdr.tab);
clear_combined_headers (http);
clear_buffer (http, &http->req.tra);
clear_buffer (http, &http->req.con);
clear_buffer (http, &http->req.raw);
}
@ -407,9 +413,12 @@ qse_http_t* qse_http_init (qse_http_t* http, qse_mmgr_t* mmgr)
init_buffer (http, &http->req.raw);
init_buffer (http, &http->req.con);
init_buffer (http, &http->req.tra);
if (qse_htb_init (&http->req.hdr.tab, mmgr, 60, 70, 1, 1) == QSE_NULL)
{
fini_buffer (http, &http->req.tra);
fini_buffer (http, &http->req.con);
fini_buffer (http, &http->req.raw);
return QSE_NULL;
}
@ -421,6 +430,7 @@ void qse_http_fini (qse_http_t* http)
{
qse_htb_fini (&http->req.hdr.tab);
clear_combined_headers (http);
fini_buffer (http, &http->req.tra);
fini_buffer (http, &http->req.con);
fini_buffer (http, &http->req.raw);
}
@ -951,14 +961,13 @@ qse_printf (QSE_T("HEADER OK %d[%S] %d[%S]\n"), (int)pair->klen, pair->kptr, (i
static QSE_INLINE int parse_request (
qse_http_t* http, const qse_byte_t* req, qse_size_t rlen)
{
static const qse_byte_t nul = '\0';
qse_byte_t* p;
/* add the actual request */
if (push_to_buffer (http, &http->req.raw, req, rlen) <= -1) return -1;
/* add the terminating null for easier parsing */
if (push_to_buffer (http, &http->req.raw, &nul, 1) <= -1) return -1;
if (push_to_buffer (http, &http->req.raw, &NUL, 1) <= -1) return -1;
p = http->req.raw.data;
@ -1037,8 +1046,8 @@ static const qse_byte_t* getchunklen (qse_http_t* http, const qse_byte_t* ptr, q
{
/* length explicity specified to 0
get trailing headers .... */
/*TODO: => http->req.state.chunk.phase = GET_CHUNK_TRAILERS;*/
http->req.state.chunk.phase = GET_CHUNK_DATA;
http->req.state.chunk.phase = GET_CHUNK_TRAILERS;
//qse_printf (QSE_T("SWITCH TO GET_CHUNK_TRAILERS....\n"));
}
else
{
@ -1061,11 +1070,86 @@ static const qse_byte_t* getchunklen (qse_http_t* http, const qse_byte_t* ptr, q
return ptr;
}
/* feed the percent encoded string */
int qse_http_feed (qse_http_t* http, const qse_byte_t* ptr, qse_size_t len)
static const qse_byte_t* get_trailing_headers (
qse_http_t* http, const qse_byte_t* req, const qse_byte_t* end)
{
const qse_byte_t* end = ptr + len;
const qse_byte_t* req = ptr;
const qse_byte_t* ptr = req;
while (ptr < end)
{
register qse_byte_t b = *ptr++;
switch (b)
{
case '\0':
/* guarantee that the request does not contain a null
* character */
http->errnum = QSE_HTTP_EBADREQ;
return -1;
case '\n':
if (http->req.state.crlf <= 1)
{
http->req.state.crlf = 2;
break;
}
else
{
qse_byte_t* p;
QSE_ASSERT (http->req.state.crlf <= 3);
http->req.state.crlf = 0;
if (push_to_buffer (
http, &http->req.tra, req, ptr - req) <= -1)
return QSE_NULL;
if (push_to_buffer (
http, &http->req.tra, &NUL, 1) <= -1)
return QSE_NULL;
p = http->req.tra.data;
do
{
while (is_whspace_octet(*p)) p++;
if (*p == '\0') break;
/* TODO: return error if protocol is 0.9.
* HTTP/0.9 must not get headers... */
p = parse_header_fields (http, p);
if (p == QSE_NULL) return QSE_NULL;
}
while (1);
http->req.state.chunk.phase = GET_CHUNK_DONE;
goto done;
}
case '\r':
if (http->req.state.crlf == 0 || http->req.state.crlf == 2)
http->req.state.crlf++;
else http->req.state.crlf = 1;
break;
default:
/* mark that neither CR nor LF was seen */
http->req.state.crlf = 0;
}
}
if (push_to_buffer (http, &http->req.tra, req, ptr - req) <= -1) return QSE_NULL;
done:
return ptr;
}
/* feed the percent encoded string */
int qse_http_feed (qse_http_t* http, const qse_byte_t* req, qse_size_t len)
{
const qse_byte_t* end = req + len;
const qse_byte_t* ptr = req;
/* does this goto drop code maintainability? */
if (http->req.state.need > 0) goto content_resume;
@ -1082,10 +1166,8 @@ int qse_http_feed (qse_http_t* http, const qse_byte_t* ptr, qse_size_t len)
case GET_CHUNK_CRLF:
goto dechunk_crlf;
/*
case GET_CHUNK_TRAILERS:
goto ....
*/
goto dechunk_get_trailers;
}
while (ptr < end)
@ -1100,128 +1182,156 @@ int qse_http_feed (qse_http_t* http, const qse_byte_t* ptr, qse_size_t len)
continue;
}
if (b == '\n')
switch (b)
{
if (http->req.state.crlf <= 1)
{
/* http->req.state.crlf == 0, CR was not seen
* http->req.state.crlf == 1, CR was seen
* whatever the current case is, mark the
* first LF is seen here.
*/
http->req.state.crlf = 2;
}
else
{
/* http->req.state.crlf == 2, no 2nd CR before LF
* http->req.state.crlf == 3, 2nd CR before LF
*/
case '\0':
/* guarantee that the request does not contain a null
* character */
http->errnum = QSE_HTTP_EBADREQ;
return -1;
/* we got a complete request. */
QSE_ASSERT (http->req.state.crlf <= 3);
/* reset the crlf state */
http->req.state.crlf = 0;
/* reset the raw request length */
http->req.state.plen = 0;
if (parse_request (http, req, ptr - req) <= -1)
return -1;
if (http->req.attr.chunked)
case '\n':
if (http->req.state.crlf <= 1)
{
/* transfer-encoding: chunked */
QSE_ASSERT (http->req.attr.content_length <= 0);
dechunk_start:
http->req.state.chunk.phase = GET_CHUNK_LEN;
http->req.state.chunk.len = 0;
http->req.state.chunk.count = 0;
dechunk_resume:
ptr = getchunklen (http, ptr, end - ptr);
if (ptr == QSE_NULL) return -1;
if (http->req.state.chunk.phase == GET_CHUNK_LEN)
{
/* still in the GET_CHUNK_LEN state.
* the length has been partially read. */
goto feedme_more;
}
/* http->req.state.crlf == 0
* => CR was not seen
* http->req.state.crlf == 1
* => CR was seen
* whatever the current case is,
* mark the first LF is seen here.
*/
http->req.state.crlf = 2;
}
else
{
/* we need to read as many octets as Content-Length */
http->req.state.need = http->req.attr.content_length;
}
if (http->req.state.need > 0)
{
/* content-length or chunked data length specified */
qse_size_t avail;
content_resume:
avail = end - ptr;
if (avail < http->req.state.need)
/* http->req.state.crlf == 2
* => no 2nd CR before LF
* http->req.state.crlf == 3
* => 2nd CR before LF
*/
/* we got a complete request. */
QSE_ASSERT (http->req.state.crlf <= 3);
/* reset the crlf state */
http->req.state.crlf = 0;
/* reset the raw request length */
http->req.state.plen = 0;
if (parse_request (http, req, ptr - req) <= -1)
return -1;
if (http->req.attr.chunked)
{
/* the data is not as large as needed */
if (push_to_buffer (http, &http->req.con, ptr, avail) <= -1) return -1;
http->req.state.need -= avail;
/* we didn't get a complete content yet */
goto feedme_more;
}
else
{
/* we are given all needed or more than needed */
if (push_to_buffer (http, &http->req.con, ptr, http->req.state.need) <= -1) return -1;
ptr += http->req.state.need;
http->req.state.need = 0;
}
}
if (http->req.state.chunk.phase == GET_CHUNK_DATA)
{
QSE_ASSERT (http->req.state.need == 0);
http->req.state.chunk.phase = GET_CHUNK_CRLF;
dechunk_crlf:
while (ptr < end && is_space_octet(*ptr)) ptr++;
if (ptr < end)
{
if (*ptr == '\n')
/* transfer-encoding: chunked */
QSE_ASSERT (http->req.attr.content_length <= 0);
dechunk_start:
http->req.state.chunk.phase = GET_CHUNK_LEN;
http->req.state.chunk.len = 0;
http->req.state.chunk.count = 0;
dechunk_resume:
ptr = getchunklen (http, ptr, end - ptr);
if (ptr == QSE_NULL) return -1;
if (http->req.state.chunk.phase == GET_CHUNK_LEN)
{
/* end of chunk data. */
ptr++;
/* more octets still available.
* let it decode the next chunk */
if (ptr < end) goto dechunk_start;
/* no more octets available after chunk data.
* the chunk state variables need to be
* reset when a jump is made to dechunk_resume
* upon the next call */
http->req.state.chunk.phase = GET_CHUNK_LEN;
http->req.state.chunk.len = 0;
http->req.state.chunk.count = 0;
/* still in the GET_CHUNK_LEN state.
* the length has been partially read. */
goto feedme_more;
}
else
else if (http->req.state.chunk.phase == GET_CHUNK_TRAILERS)
{
/* redundant character ... */
http->errnum = QSE_HTTP_EBADREQ;
return -1;
dechunk_get_trailers:
ptr = get_trailing_headers (http, ptr, end);
if (ptr == QSE_NULL) return -1;
if (http->req.state.chunk.phase == GET_CHUNK_TRAILERS)
{
/* still in the same state.
* the trailers have not been processed fully */
goto feedme_more;
}
}
}
else
{
/* data not enough */
goto feedme_more;
/* we need to read as many octets as Content-Length */
http->req.state.need = http->req.attr.content_length;
}
if (http->req.state.need > 0)
{
/* content-length or chunked data length specified */
qse_size_t avail;
content_resume:
avail = end - ptr;
if (avail < http->req.state.need)
{
/* the data is not as large as needed */
if (push_to_buffer (http, &http->req.con, ptr, avail) <= -1) return -1;
http->req.state.need -= avail;
/* we didn't get a complete content yet */
goto feedme_more;
}
else
{
/* we got all or more than needed */
if (push_to_buffer (
http, &http->req.con, ptr,
http->req.state.need) <= -1) return -1;
ptr += http->req.state.need;
http->req.state.need = 0;
}
}
if (http->req.state.chunk.phase == GET_CHUNK_DATA)
{
QSE_ASSERT (http->req.state.need == 0);
http->req.state.chunk.phase = GET_CHUNK_CRLF;
dechunk_crlf:
while (ptr < end && is_space_octet(*ptr)) ptr++;
if (ptr < end)
{
if (*ptr == '\n')
{
/* end of chunk data. */
ptr++;
/* more octets still available.
* let it decode the next chunk
*/
if (ptr < end) goto dechunk_start;
/* no more octets available after
* chunk data. the chunk state variables
* need to be reset when a jump is made
* to dechunk_resume upon the next call
*/
http->req.state.chunk.phase = GET_CHUNK_LEN;
http->req.state.chunk.len = 0;
http->req.state.chunk.count = 0;
goto feedme_more;
}
else
{
/* redundant character ... */
http->errnum = QSE_HTTP_EBADREQ;
return -1;
}
}
else
{
/* data not enough */
goto feedme_more;
}
}
}
qse_htb_walk (&http->req.hdr.tab, walk, QSE_NULL);
if (http->req.con.size > 0)
@ -1230,32 +1340,26 @@ if (http->req.con.size > 0)
}
/* TODO: do the main job here... before the raw buffer is cleared out... */
clear_request (http);
clear_request (http);
/* let ptr point to the next character to LF or the optional contents */
req = ptr;
}
}
else if (b == '\r')
{
if (http->req.state.crlf == 0 || http->req.state.crlf == 2)
http->req.state.crlf++;
else http->req.state.crlf = 1;
}
else if (b == '\0')
{
/* guarantee that the request does not contain a null
* character */
http->errnum = QSE_HTTP_EBADREQ;
return -1;
}
else
{
/* increment length of a request in raw
* excluding crlf */
http->req.state.plen++;
/* mark that neither CR nor LF was seen */
http->req.state.crlf = 0;
/* let ptr point to the next character to LF or
* the optional contents */
req = ptr;
}
break;
case '\r':
if (http->req.state.crlf == 0 || http->req.state.crlf == 2)
http->req.state.crlf++;
else http->req.state.crlf = 1;
break;
default:
/* increment length of a request in raw
* excluding crlf */
http->req.state.plen++;
/* mark that neither CR nor LF was seen */
http->req.state.crlf = 0;
}
}
@ -1267,5 +1371,6 @@ if (http->req.con.size > 0)
feedme_more:
return 0;
}