sourcemod-central: core/TextParsers.cpp@2a83a521287c (annotated)

ds@1952	1	/**
ds@1952	2	* vim: set ts=4 :
ds@1952	3	* =============================================================================
ds@1952	4	* SourceMod
ds@1952	5	* Copyright (C) 2004-2008 AlliedModders LLC. All rights reserved.
ds@1952	6	* =============================================================================
ds@1952	7	*
ds@1952	8	* This program is free software; you can redistribute it and/or modify it under
ds@1952	9	* the terms of the GNU General Public License, version 3.0, as published by the
ds@1952	10	* Free Software Foundation.
ds@1952	11	*
ds@1952	12	* This program is distributed in the hope that it will be useful, but WITHOUT
ds@1952	13	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
ds@1952	14	* FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
ds@1952	15	* details.
ds@1952	16	*
ds@1952	17	* You should have received a copy of the GNU General Public License along with
ds@1952	18	* this program. If not, see <http://www.gnu.org/licenses/>.
ds@1952	19	*
ds@1952	20	* As a special exception, AlliedModders LLC gives you permission to link the
ds@1952	21	* code of this program (as well as its derivative works) to "Half-Life 2," the
ds@1952	22	* "Source Engine," the "SourcePawn JIT," and any Game MODs that run on software
ds@1952	23	* by the Valve Corporation. You must obey the GNU General Public License in
ds@1952	24	* all respects for all other code used. Additionally, AlliedModders LLC grants
ds@1952	25	* this exception to all derivative works. AlliedModders LLC defines further
ds@1952	26	* exceptions, found in LICENSE.txt (as of this writing, version JULY-31-2007),
ds@1952	27	* or <http://www.sourcemod.net/license.php>.
ds@1952	28	*
ds@1952	29	* Version: $Id$
ds@1952	30	*/
ds@1952	31
ds@1952	32	#include <stdio.h>
ds@1952	33	#include <ctype.h>
ds@1952	34	#include <wctype.h>
ds@1952	35	#include <string.h>
ds@1952	36	#include <stdlib.h>
ds@1952	37	#include <assert.h>
ds@1952	38	#include "TextParsers.h"
ds@1952	39	#include "ShareSys.h"
ds@1952	40	#include "sm_stringutil.h"
ds@1952	41	#include "LibrarySys.h"
ds@1952	42
ds@1952	43	TextParsers g_TextParser;
ds@1952	44	ITextParsers *textparsers = &g_TextParser;
ds@1952	45
ds@1952	46	static int g_ini_chartable1[255] = {0};
ds@1952	47	static int g_ws_chartable[255] = {0};
ds@1952	48
ds@1952	49	bool TextParsers::IsWhitespace(const char *stream)
ds@1952	50	{
ds@1952	51	return g_ws_chartable[(unsigned)*stream] == 1;
ds@1952	52	}
ds@1952	53
ds@1952	54	TextParsers::TextParsers()
ds@1952	55	{
ds@1952	56	g_ini_chartable1[(unsigned)'_'] = 1;
ds@1952	57	g_ini_chartable1[(unsigned)'-'] = 1;
ds@1952	58	g_ini_chartable1[(unsigned)','] = 1;
ds@1952	59	g_ini_chartable1[(unsigned)'+'] = 1;
ds@1952	60	g_ini_chartable1[(unsigned)'.'] = 1;
ds@1952	61	g_ini_chartable1[(unsigned)'$'] = 1;
ds@1952	62	g_ini_chartable1[(unsigned)'?'] = 1;
ds@1952	63	g_ini_chartable1[(unsigned)'/'] = 1;
ds@1952	64	g_ws_chartable[(unsigned)'\n'] = 1;
ds@1952	65	g_ws_chartable[(unsigned)'\v'] = 1;
ds@1952	66	g_ws_chartable[(unsigned)'\r'] = 1;
ds@1952	67	g_ws_chartable[(unsigned)'\t'] = 1;
ds@1952	68	g_ws_chartable[(unsigned)'\f'] = 1;
ds@1952	69	g_ws_chartable[(unsigned)' '] = 1;
ds@1952	70	}
ds@1952	71
ds@1952	72	void TextParsers::OnSourceModAllInitialized()
ds@1952	73	{
ds@1952	74	g_ShareSys.AddInterface(NULL, this);
ds@1952	75	}
ds@1952	76
ds@1952	77	unsigned int TextParsers::GetUTF8CharBytes(const char *stream)
ds@1952	78	{
ds@1952	79	return _GetUTF8CharBytes(stream);
ds@1952	80	}
ds@1952	81
ds@1952	82	/**
ds@1952	83	* Character streams
ds@1952	84	*/
ds@1952	85
ds@1952	86	struct CharStream
ds@1952	87	{
ds@1952	88	const char *curpos;
ds@1952	89	};
ds@1952	90
ds@1952	91	bool CharStreamReader(void stream, char buffer, size_t maxlength, unsigned int *read)
ds@1952	92	{
ds@1952	93	CharStream srdr = (CharStream )stream;
ds@1952	94
ds@1952	95	const char *ptr = srdr->curpos;
ds@1952	96	for (size_t i=0; i<maxlength; i++)
ds@1952	97	{
ds@1952	98	if (*ptr == '\0')
ds@1952	99	{
ds@1952	100	break;
ds@1952	101	}
ds@1952	102	buffer++ = ptr++;
ds@1952	103	}
ds@1952	104
ds@1952	105	*read = ptr - srdr->curpos;
ds@1952	106
ds@1952	107	srdr->curpos = ptr;
ds@1952	108
ds@1952	109	return true;
ds@1952	110	}
ds@1952	111
ds@1952	112	SMCError TextParsers::ParseString_SMC(const char *stream,
ds@1952	113	ITextListener_SMC *smc,
ds@1952	114	SMCStates *states)
ds@1952	115	{
ds@1952	116	CharStream srdr = { stream };
ds@1952	117
ds@1952	118	return ParseStream_SMC(&srdr, CharStreamReader, smc, states);
ds@1952	119	}
ds@1952	120
ds@1952	121	/**
ds@1952	122	* File streams
ds@1952	123	*/
ds@1952	124
ds@1952	125	bool FileStreamReader(void stream, char buffer, size_t maxlength, unsigned int *read)
ds@1952	126	{
ds@1952	127	size_t num = fread(buffer, 1, maxlength, (FILE *)stream);
ds@1952	128
ds@1952	129	*read = static_cast<unsigned int>(num);
ds@1952	130
ds@1952	131	if (num == 0 && feof((FILE *)stream))
ds@1952	132	{
ds@1952	133	return true;
ds@1952	134	}
ds@1952	135
ds@1952	136	return (ferror((FILE *)stream) == 0);
ds@1952	137	}
ds@1952	138
ds@1952	139	SMCError TextParsers::ParseFile_SMC(const char file, ITextListener_SMC smc, SMCStates *states)
ds@1952	140	{
ds@1952	141	FILE *fp = fopen(file, "rt");
ds@1952	142
ds@1952	143	if (!fp)
ds@1952	144	{
ds@1952	145	if (states != NULL)
ds@1952	146	{
ds@1952	147	states->line = 0;
ds@1952	148	states->col = 0;
ds@1952	149	}
ds@1952	150	return SMCError_StreamOpen;
ds@1952	151	}
ds@1952	152
ds@1952	153	SMCError result = ParseStream_SMC(fp, FileStreamReader, smc, states);
ds@1952	154
ds@1952	155	fclose(fp);
ds@1952	156
ds@1952	157	return result;
ds@1952	158	}
ds@1952	159
ds@1952	160	SMCError TextParsers::ParseSMCFile(const char *file,
ds@1952	161	ITextListener_SMC *smc_listener,
ds@1952	162	SMCStates *states,
ds@1952	163	char *buffer,
ds@1952	164	size_t maxsize)
ds@1952	165	{
ds@1952	166	const char *errstr;
ds@1952	167	FILE *fp = fopen(file, "rt");
ds@1952	168
ds@1952	169	if (fp == NULL)
ds@1952	170	{
ds@1952	171	char error[256] = "unknown";
ds@1952	172	if (states != NULL)
ds@1952	173	{
ds@1952	174	states->line = 0;
ds@1952	175	states->col = 0;
ds@1952	176	}
ds@1952	177	g_LibSys.GetPlatformError(error, sizeof(error));
ds@1952	178	UTIL_Format(buffer, maxsize, "File could not be opened: %s", error);
ds@1952	179	return SMCError_StreamOpen;
ds@1952	180	}
ds@1952	181
ds@1952	182	SMCError result = ParseStream_SMC(fp, FileStreamReader, smc_listener, states);
ds@1952	183
ds@1952	184	fclose(fp);
ds@1952	185
ds@1952	186	errstr = GetSMCErrorString(result);
ds@1952	187	UTIL_Format(buffer, maxsize, "%s", errstr != NULL ? errstr : "Unknown error");
ds@1952	188
ds@1952	189	return result;
ds@1952	190	}
ds@1952	191
ds@1952	192	/**
ds@1952	193	* Raw parsing of streams with helper functions
ds@1952	194	*/
ds@1952	195
ds@1952	196	struct StringInfo
ds@1952	197	{
ds@1952	198	StringInfo() : quoted(false), ptr(NULL), end(NULL), special(false) { }
ds@1952	199	bool quoted;
ds@1952	200	char *ptr;
ds@1952	201	char *end;
ds@1952	202	bool special;
ds@1952	203	};
ds@1952	204
ds@1952	205	const char *FixupString(StringInfo &data)
ds@1952	206	{
ds@1952	207	if (!data.ptr)
ds@1952	208	{
ds@1952	209	return NULL;
ds@1952	210	}
ds@1952	211
ds@1952	212	if (data.quoted)
ds@1952	213	{
ds@1952	214	data.ptr++;
ds@1952	215	}
ds@1952	216	#if defined _DEBUG
ds@1952	217	else {
ds@1952	218	/* A string will never have beginning whitespace because we ignore it in the stream.
ds@1952	219	* Furthermore, if there is trailing whitespace, the end ptr will point to it, so it is valid
ds@1952	220	* to overwrite! Lastly, the last character must be whitespace or a comment/invalid character.
ds@1952	221	*/
ds@1952	222	}
ds@1952	223	#endif
ds@1952	224
ds@1952	225	/* Do some extra work on strings that have special quoted characters. */
ds@1952	226	if (data.special)
ds@1952	227	{
ds@1952	228	char *outptr = data.ptr;
ds@1952	229	size_t len = data.end - data.ptr;
ds@1952	230	if (len >= 2)
ds@1952	231	{
ds@1952	232	for (size_t i=0; i<len; i++)
ds@1952	233	{
ds@1952	234	if (data.ptr[i] == '\\' && i < len - 1)
ds@1952	235	{
ds@1952	236	/* Resolve the next character. */
ds@1952	237	i++;
ds@1952	238	if (data.ptr[i] == 'n')
ds@1952	239	{
ds@1952	240	data.ptr[i] = '\n';
ds@1952	241	} else if (data.ptr[i] == 't') {
ds@1952	242	data.ptr[i] = '\t';
ds@1952	243	} else if (data.ptr[i] == 'r') {
ds@1952	244	data.ptr[i] = '\r';
ds@1952	245	} else if (data.ptr[i] != '\\'
ds@1952	246	&& data.ptr[i] != '"') {
ds@1952	247	/* This character is invalid, so go back one */
ds@1952	248	i--;
ds@1952	249	}
ds@1952	250	}
ds@1952	251	*outptr++ = data.ptr[i];
ds@1952	252	}
ds@1952	253	*outptr = '\0';
ds@1952	254	}
ds@1952	255	}
ds@1952	256
ds@1952	257	if (data.end)
ds@1952	258	{
ds@1952	259	*(data.end) = '\0';
ds@1952	260	}
ds@1952	261
ds@1952	262	return data.ptr;
ds@1952	263	}
ds@1952	264
ds@1952	265	const char *rotate(StringInfo info[3])
ds@1952	266	{
ds@1952	267	if (info[2].ptr != NULL)
ds@1952	268	{
ds@1952	269	return info[2].ptr;
ds@1952	270	}
ds@1952	271
ds@1952	272	if (info[0].ptr != NULL)
ds@1952	273	{
ds@1952	274	info[2] = info[1];
ds@1952	275	info[1] = info[0];
ds@1952	276	info[0] = StringInfo();
ds@1952	277	}
ds@1952	278
ds@1952	279	return NULL;
ds@1952	280	}
ds@1952	281
ds@1952	282	void scrap(StringInfo info[3])
ds@1952	283	{
ds@1952	284	info[2] = StringInfo();
ds@1952	285	info[1] = StringInfo();
ds@1952	286	info[0] = StringInfo();
ds@1952	287	}
ds@1952	288
ds@1952	289	void reloc(StringInfo &data, unsigned int bytes)
ds@1952	290	{
ds@1952	291	if (data.ptr)
ds@1952	292	{
ds@1952	293	data.ptr -= bytes;
ds@1952	294	}
ds@1952	295	if (data.end)
ds@1952	296	{
ds@1952	297	data.end -= bytes;
ds@1952	298	}
ds@1952	299	}
ds@1952	300
ds@1952	301	char *lowstring(StringInfo info[3])
ds@1952	302	{
ds@1952	303	for (int i=2; i>=0; i--)
ds@1952	304	{
ds@1952	305	if (info[i].ptr)
ds@1952	306	{
ds@1952	307	return info[i].ptr;
ds@1952	308	}
ds@1952	309	}
ds@1952	310
ds@1952	311	return NULL;
ds@1952	312	}
ds@1952	313
ds@1952	314	SMCError TextParsers::ParseStream_SMC(void *stream,
ds@1952	315	STREAMREADER srdr,
ds@1952	316	ITextListener_SMC *smc,
ds@1952	317	SMCStates *pStates)
ds@1952	318	{
ds@1952	319	char *reparse_point = NULL;
ds@1952	320	char in_buf[4096];
ds@1952	321	char *parse_point = in_buf;
ds@1952	322	char *line_begin = in_buf;
ds@1952	323	unsigned int read;
ds@1952	324	unsigned int curlevel = 0;
ds@1952	325	bool in_quote = false;
ds@1952	326	bool ignoring = false;
ds@1952	327	bool eol_comment = false;
ds@1952	328	bool ml_comment = false;
ds@1952	329	unsigned int i;
ds@1952	330	SMCError err = SMCError_Okay;
ds@1952	331	SMCResult res;
ds@1952	332	SMCStates states;
ds@1952	333	char c;
ds@1952	334
ds@1952	335	StringInfo strings[3];
ds@1952	336	StringInfo emptystring;
ds@1952	337
ds@1952	338	states.line = 1;
ds@1952	339	states.col = 0;
ds@1952	340
ds@1952	341	smc->ReadSMC_ParseStart();
ds@1952	342
ds@1952	343	/**
ds@1952	344	* The stream reader reads in as much as it can fill the buffer with.
ds@1952	345	* It then processes the buffer. If the buffer cannot be fully processed, for example,
ds@1952	346	* a line is left hanging with no newline, then the contents of the buffer is shifted
ds@1952	347	* down, and the buffer is filled from the stream reader again.
ds@1952	348	*
ds@1952	349	* What makes this particularly annoying is that we cache pointers everywhere, so when
ds@1952	350	* the shifting process takes place, all those pointers must be shifted as well.
ds@1952	351	*/
ds@1952	352	while (srdr(stream, parse_point, sizeof(in_buf) - (parse_point - in_buf) - 1, &read))
ds@1952	353	{
ds@1952	354	if (!read)
ds@1952	355	{
ds@1952	356	break;
ds@1952	357	}
ds@1952	358
ds@1952	359	/* Check for BOM markings, which is only relevant on the first line.
ds@1952	360	* Not worth it, but it could be moved out of the loop.
ds@1952	361	*/
ds@1952	362	if (states.line == 1 &&
ds@1952	363	in_buf[0] == (char)0xEF &&
ds@1952	364	in_buf[1] == (char)0xBB &&
ds@1952	365	in_buf[2] == (char)0xBF)
ds@1952	366	{
ds@1952	367	/* Move EVERYTHING down :\ */
ds@1952	368	memmove(in_buf, &in_buf[3], read - 3);
ds@1952	369	read -= 3;
ds@1952	370	}
ds@1952	371
ds@1952	372	if (reparse_point)
ds@1952	373	{
ds@1952	374	read += (parse_point - reparse_point);
ds@1952	375	parse_point = reparse_point;
ds@1952	376	reparse_point = NULL;
ds@1952	377	}
ds@1952	378
ds@1952	379	for (i=0; i<read; i++)
ds@1952	380	{
ds@1952	381	c = parse_point[i];
ds@1952	382	if (c == '\n')
ds@1952	383	{
ds@1952	384	/* If we got a newline, there's a lot of things that could have happened in the interim.
ds@1952	385	* First, let's make sure the staged strings are rotated.
ds@1952	386	*/
ds@1952	387	if (strings[0].ptr)
ds@1952	388	{
ds@1952	389	strings[0].end = &parse_point[i];
ds@1952	390	if (rotate(strings) != NULL)
ds@1952	391	{
ds@1952	392	err = SMCError_InvalidTokens;
ds@1952	393	goto failed;
ds@1952	394	}
ds@1952	395	}
ds@1952	396
ds@1952	397	/* Next, let's clear some line-based values that may no longer have meaning */
ds@1952	398	eol_comment = false;
ds@1952	399	in_quote = false;
ds@1952	400	if (ignoring && !ml_comment)
ds@1952	401	{
ds@1952	402	ignoring = false;
ds@1952	403	}
ds@1952	404
ds@1952	405	/* Pass the raw line onto the listener. We terminate the line so the receiver
ds@1952	406	* doesn't get tons of useless info. We restore the newline after.
ds@1952	407	*/
ds@1952	408	parse_point[i] = '\0';
ds@1952	409	if ((res=smc->ReadSMC_RawLine(&states, line_begin)) != SMCResult_Continue)
ds@1952	410	{
ds@1952	411	err = (res == SMCResult_HaltFail) ? SMCError_Custom : SMCError_Okay;
ds@1952	412	goto failed;
ds@1952	413	}
ds@1952	414	parse_point[i] = '\n';
ds@1952	415
ds@1952	416	/* Now we check the sanity of our staged strings! */
ds@1952	417	if (strings[2].ptr)
ds@1952	418	{
ds@1952	419	if (!curlevel)
ds@1952	420	{
ds@1952	421	err = SMCError_InvalidProperty1;
ds@1952	422	goto failed;
ds@1952	423	}
ds@1952	424	/* Assume the next string is a property and pass the info on. */
ds@1952	425	if ((res=smc->ReadSMC_KeyValue(
ds@1952	426	&states,
ds@1952	427	FixupString(strings[2]),
ds@1952	428	FixupString(strings[1]))) != SMCResult_Continue)
ds@1952	429	{
ds@1952	430	err = (res == SMCResult_HaltFail) ? SMCError_Custom : SMCError_Okay;
ds@1952	431	goto failed;
ds@1952	432	}
ds@1952	433	scrap(strings);
ds@1952	434	}
ds@1952	435
ds@1952	436	/* Change the states for the next line */
ds@1952	437	states.col = 0;
ds@1952	438	states.line++;
ds@1952	439	line_begin = &parse_point[i+1]; //Note: safe because this gets relocated later
ds@1952	440	}
ds@1952	441	else if (ignoring)
ds@1952	442	{
ds@1952	443	if (in_quote)
ds@1952	444	{
ds@1952	445	/* If i was 0, we could have reparsed, so make sure there's no buffer underrun */
ds@1952	446	if ((&parse_point[i] != in_buf) && c == '"' && parse_point[i-1] != '\\')
ds@1952	447	{
ds@1952	448	/* If we reached a quote in an ignore phase,
ds@1952	449	* we're staging a string and we must rotate it out.
ds@1952	450	*/
ds@1952	451	in_quote = false;
ds@1952	452	ignoring = false;
ds@1952	453	/* Set our info */
ds@1952	454	strings[0].end = &parse_point[i];
ds@1952	455	strings[0].quoted = true;
ds@1952	456	if (rotate(strings) != NULL)
ds@1952	457	{
ds@1952	458	/* If we rotated too many strings, there was too much crap on one line */
ds@1952	459	err = SMCError_InvalidTokens;
ds@1952	460	goto failed;
ds@1952	461	}
ds@1952	462	}
ds@1952	463	else if (c == '\\')
ds@1952	464	{
ds@1952	465	strings[0].special = true;
ds@1952	466	if (i == (read - 1))
ds@1952	467	{
ds@1952	468	reparse_point = &parse_point[i];
ds@1952	469	break;
ds@1952	470	}
ds@1952	471	}
ds@1952	472	}
ds@1952	473	else if (ml_comment)
ds@1952	474	{
ds@1952	475	if (c == '*')
ds@1952	476	{
ds@1952	477	/* Check if we need to get more input first */
ds@1952	478	if (i == read - 1)
ds@1952	479	{
ds@1952	480	reparse_point = &parse_point[i];
ds@1952	481	break;
ds@1952	482	}
ds@1952	483	if (parse_point[i+1] == '/')
ds@1952	484	{
ds@1952	485	ml_comment = false;
ds@1952	486	ignoring = false;
ds@1952	487	/* We should not be staging anything right now. */
ds@1952	488	assert(strings[0].ptr == NULL);
ds@1952	489	/* Advance the input stream so we don't choke on this token */
ds@1952	490	i++;
ds@1952	491	states.col++;
ds@1952	492	}
ds@1952	493	}
ds@1952	494	}
ds@1952	495	}
ds@1952	496	else
ds@1952	497	{
ds@1952	498	/* Check if we're whitespace or not */
ds@1952	499	if (!g_ws_chartable[(unsigned)c])
ds@1952	500	{
ds@1952	501	bool restage = false;
ds@1952	502	/* Check various special tokens:
ds@1952	503	* ;
ds@1952	504	* //
ds@1952	505	* / *
ds@1952	506	* {
ds@1952	507	* }
ds@1952	508	*/
ds@1952	509	if (c == ';' \|\| c == '/')
ds@1952	510	{
ds@1952	511	/* If it's a line-based comment (that is, ; or //)
ds@1952	512	* we will need to scrap everything until the end of the line.
ds@1952	513	*/
ds@1952	514	if (c == '/')
ds@1952	515	{
ds@1952	516	if (i == read - 1)
ds@1952	517	{
ds@1952	518	/* If we reached the end of the look-ahead, we need to re-check our input.
ds@1952	519	* Breaking out will force this to be the new reparse point!
ds@1952	520	*/
ds@1952	521	reparse_point = &parse_point[i];
ds@1952	522	break;
ds@1952	523	}
ds@1952	524	if (parse_point[i + 1] == '/')
ds@1952	525	{
ds@1952	526	/* standard comment */
ds@1952	527	ignoring = true;
ds@1952	528	eol_comment = true;
ds@1952	529	restage = true;
ds@1952	530	}
ds@1952	531	else if (parse_point[i+1] == '*')
ds@1952	532	{
ds@1952	533	/* inline comment - start ignoring */
ds@1952	534	ignoring = true;
ds@1952	535	ml_comment = true;
ds@1952	536	/* yes, we restage, meaning that:
ds@1952	537	* STR/ stuff /ING (space because ml comments don't nest in C++)
ds@1952	538	* will not generate 'STRING', but rather 'STR' and 'ING'.
ds@1952	539	* This should be a rare occurrence and is done here for convenience.
ds@1952	540	*/
ds@1952	541	restage = true;
ds@1952	542	}
ds@1952	543	}
ds@1952	544	else
ds@1952	545	{
ds@1952	546	ignoring = true;
ds@1952	547	eol_comment = true;
ds@1952	548	restage = true;
ds@1952	549	}
ds@1952	550	}
ds@1952	551	else if (c == '{')
ds@1952	552	{
ds@1952	553	/* If we are staging a string, we must rotate here */
ds@1952	554	if (strings[0].ptr)
ds@1952	555	{
ds@1952	556	/* We have unacceptable tokens on this line */
ds@1952	557	if (rotate(strings) != NULL)
ds@1952	558	{
ds@1952	559	err = SMCError_InvalidSection1;
ds@1952	560	goto failed;
ds@1952	561	}
ds@1952	562	}
ds@1952	563	/* Sections must always be alone */
ds@1952	564	if (strings[2].ptr != NULL)
ds@1952	565	{
ds@1952	566	err = SMCError_InvalidSection1;
ds@1952	567	goto failed;
ds@1952	568	}
ds@1952	569	else if (strings[1].ptr == NULL)
ds@1952	570	{
ds@1952	571	err = SMCError_InvalidSection2;
ds@1952	572	goto failed;
ds@1952	573	}
ds@1952	574	if ((res=smc->ReadSMC_NewSection(&states, FixupString(strings[1])))
ds@1952	575	!= SMCResult_Continue)
ds@1952	576	{
ds@1952	577	err = (res == SMCResult_HaltFail) ? SMCError_Custom : SMCError_Okay;
ds@1952	578	goto failed;
ds@1952	579	}
ds@1952	580	strings[1] = emptystring;
ds@1952	581	curlevel++;
ds@1952	582	}
ds@1952	583	else if (c == '}')
ds@1952	584	{
ds@1952	585	/* Unlike our matching friend, this can be on the same line as something prior */
ds@1952	586	if (rotate(strings) != NULL)
ds@1952	587	{
ds@1952	588	err = SMCError_InvalidSection3;
ds@1952	589	goto failed;
ds@1952	590	}
ds@1952	591	if (strings[2].ptr)
ds@1952	592	{
ds@1952	593	if (!curlevel)
ds@1952	594	{
ds@1952	595	err = SMCError_InvalidProperty1;
ds@1952	596	goto failed;
ds@1952	597	}
ds@1952	598	if ((res=smc->ReadSMC_KeyValue(
ds@1952	599	&states,
ds@1952	600	FixupString(strings[2]),
ds@1952	601	FixupString(strings[1])))
ds@1952	602	!= SMCResult_Continue)
ds@1952	603	{
ds@1952	604	err = (res == SMCResult_HaltFail) ? SMCError_Custom : SMCError_Okay;
ds@1952	605	goto failed;
ds@1952	606	}
ds@1952	607	}
ds@1952	608	else if (strings[1].ptr)
ds@1952	609	{
ds@1952	610	err = SMCError_InvalidSection3;
ds@1952	611	goto failed;
ds@1952	612	}
ds@1952	613	else if (!curlevel)
ds@1952	614	{
ds@1952	615	err = SMCError_InvalidSection4;
ds@1952	616	goto failed;
ds@1952	617	}
ds@1952	618	/* Now it's safe to leave the section */
ds@1952	619	scrap(strings);
ds@1952	620	if ((res=smc->ReadSMC_LeavingSection(&states)) != SMCResult_Continue)
ds@1952	621	{
ds@1952	622	err = (res == SMCResult_HaltFail) ? SMCError_Custom : SMCError_Okay;
ds@1952	623	goto failed;
ds@1952	624	}
ds@1952	625	curlevel--;
ds@1952	626	}
ds@1952	627	else if (c == '"')
ds@1952	628	{
ds@1952	629	/* If we get a quote mark, we always restage, but we need to do it beforehand */
ds@1952	630	if (strings[0].ptr)
ds@1952	631	{
ds@1952	632	strings[0].end = &parse_point[i];
ds@1952	633	if (rotate(strings) != NULL)
ds@1952	634	{
ds@1952	635	err = SMCError_InvalidTokens;
ds@1952	636	goto failed;
ds@1952	637	}
ds@1952	638	}
ds@1952	639	strings[0].ptr = &parse_point[i];
ds@1952	640	in_quote = true;
ds@1952	641	ignoring = true;
ds@1952	642	}
ds@1952	643	else if (!strings[0].ptr)
ds@1952	644	{
ds@1952	645	/* If we have no string, we must start one */
ds@1952	646	strings[0].ptr = &parse_point[i];
ds@1952	647	}
ds@1952	648	if (restage && strings[0].ptr)
ds@1952	649	{
ds@1952	650	strings[0].end = &parse_point[i];
ds@1952	651	if (rotate(strings) != NULL)
ds@1952	652	{
ds@1952	653	err = SMCError_InvalidTokens;
ds@1952	654	goto failed;
ds@1952	655	}
ds@1952	656	}
ds@1952	657	}
ds@1952	658	else
ds@1952	659	{
ds@1952	660	/* If we're eating a string and get whitespace, we need to restage.
ds@1952	661	* (Note that if we are quoted, this is being ignored)
ds@1952	662	*/
ds@1952	663	if (strings[0].ptr)
ds@1952	664	{
ds@1952	665	/*
ds@1952	666	* The specification says the second string in a pair does not need to be quoted.
ds@1952	667	* Thus, we check if there's already a string on the stack.
ds@1952	668	* If there's a newline, we always rotate so the newline has an empty starter.
ds@1952	669	*/
ds@1952	670	if (!strings[1].ptr)
ds@1952	671	{
ds@1952	672	/* There's no string, so we must move this one down and eat up another */
ds@1952	673	strings[0].end = &parse_point[i];
ds@1952	674	rotate(strings);
ds@1952	675	}
ds@1952	676	else if (!strings[1].quoted)
ds@1952	677	{
ds@1952	678	err = SMCError_InvalidTokens;
ds@1952	679	goto failed;
ds@1952	680	}
ds@1952	681	}
ds@1952	682	}
ds@1952	683	}
ds@1952	684
ds@1952	685	/* Advance which token we're on */
ds@1952	686	states.col++;
ds@1952	687	}
ds@1952	688
ds@1952	689	if (line_begin != in_buf)
ds@1952	690	{
ds@1952	691	/* The line buffer has advanced, so it's safe to copy N bytes back to the beginning.
ds@1952	692	* What's N? N is the lowest point we're currently relying on.
ds@1952	693	*/
ds@1952	694	char *stage = lowstring(strings);
ds@1952	695	if (!stage \|\| stage > line_begin)
ds@1952	696	{
ds@1952	697	stage = line_begin;
ds@1952	698	}
ds@1952	699	unsigned int bytes = read - (stage - parse_point);
ds@1952	700
ds@1952	701	/* It is now safe to delete everything before the staged point */
ds@1952	702	memmove(in_buf, stage, bytes);
ds@1952	703
ds@1952	704	/* Calculate the number of bytes in the new buffer */
ds@1952	705	bytes = stage - in_buf;
ds@1952	706	/* Relocate all the cached pointers to our new base */
ds@1952	707	line_begin -= bytes;
ds@1952	708	reloc(strings[0], bytes);
ds@1952	709	reloc(strings[1], bytes);
ds@1952	710	reloc(strings[2], bytes);
ds@1952	711	if (reparse_point)
ds@1952	712	{
ds@1952	713	reparse_point -= bytes;
ds@1952	714	}
ds@1952	715	if (parse_point)
ds@1952	716	{
ds@1952	717	parse_point = &parse_point[read];
ds@1952	718	parse_point -= bytes;
ds@1952	719	}
ds@1952	720	}
ds@1952	721	else if (read == sizeof(in_buf) - 1)
ds@1952	722	{
ds@1952	723	err = SMCError_TokenOverflow;
ds@1952	724	goto failed;
ds@1952	725	}
ds@1952	726	}
ds@1952	727
ds@1952	728	/* If we're done parsing and there are tokens left over... */
ds@1952	729	if (curlevel)
ds@1952	730	{
ds@1952	731	err = SMCError_InvalidSection5;
ds@1952	732	goto failed;
ds@1952	733	}
ds@1952	734	else if (strings[0].ptr \|\| strings[1].ptr)
ds@1952	735	{
ds@1952	736	err = SMCError_InvalidTokens;
ds@1952	737	goto failed;
ds@1952	738	}
ds@1952	739
ds@1952	740	smc->ReadSMC_ParseEnd(false, false);
ds@1952	741
ds@1952	742	if (pStates != NULL)
ds@1952	743	{
ds@1952	744	*pStates = states;
ds@1952	745	}
ds@1952	746
ds@1952	747	return SMCError_Okay;
ds@1952	748
ds@1952	749	failed:
ds@1952	750	if (pStates != NULL)
ds@1952	751	{
ds@1952	752	*pStates = states;
ds@1952	753	}
ds@1952	754
ds@1952	755	smc->ReadSMC_ParseEnd(true, (err == SMCError_Custom));
ds@1952	756
ds@1952	757	return err;
ds@1952	758	}
ds@1952	759
ds@1952	760
ds@1952	761	/**
ds@1952	762	* INI parser
ds@1952	763	*/
ds@1952	764
ds@1952	765	bool TextParsers::ParseFile_INI(const char file, ITextListener_INI ini_listener, unsigned int line, unsigned int col)
ds@1952	766	{
ds@1952	767	FILE *fp = fopen(file, "rt");
ds@1952	768	unsigned int curline = 0;
ds@1952	769	unsigned int curtok;
ds@1952	770	size_t len;
ds@1952	771
ds@1952	772	if (!fp)
ds@1952	773	{
ds@1952	774	if (line)
ds@1952	775	{
ds@1952	776	*line = 0;
ds@1952	777	}
ds@1952	778
ds@1952	779	return false;
ds@1952	780	}
ds@1952	781
ds@1952	782	char buffer[2048];
ds@1952	783	char ptr, save_ptr;
ds@1952	784	bool in_quote;
ds@1952	785
ds@1952	786	while (!feof(fp))
ds@1952	787	{
ds@1952	788	curline++;
ds@1952	789	curtok = 0;
ds@1952	790	buffer[0] = '\0';
ds@1952	791	if (fgets(buffer, sizeof(buffer), fp) == NULL)
ds@1952	792	{
ds@1952	793	break;
ds@1952	794	}
ds@1952	795
ds@1952	796	//:TODO: this will only run once, so find a nice way to move it out of the while loop
ds@1952	797	/* If this is the first line, check the first three bytes for BOM */
ds@1952	798	if (curline == 1 &&
ds@1952	799	buffer[0] == (char)0xEF &&
ds@1952	800	buffer[1] == (char)0xBB &&
ds@1952	801	buffer[2] == (char)0xBF)
ds@1952	802	{
ds@1952	803	/* We have a UTF-8 marked file... skip these bytes */
ds@1952	804	ptr = &buffer[3];
ds@1952	805	} else {
ds@1952	806	ptr = buffer;
ds@1952	807	}
ds@1952	808
ds@1952	809	/***************************************************
ds@1952	810	* We preprocess the string before parsing tokens! *
ds@1952	811	***************************************************/
ds@1952	812
ds@1952	813	/* First strip beginning whitespace */
ds@1952	814	while (ptr != '\0' && g_ws_chartable[(unsigned)ptr] != 0)
ds@1952	815	{
ds@1952	816	ptr++;
ds@1952	817	}
ds@1952	818
ds@1952	819	len = strlen(ptr);
ds@1952	820
ds@1952	821	if (!len)
ds@1952	822	{
ds@1952	823	continue;
ds@1952	824	}
ds@1952	825
ds@1952	826	/* Now search for comment characters */
ds@1952	827	in_quote = false;
ds@1952	828	save_ptr = ptr;
ds@1952	829	for (size_t i=0; i<len; i++,ptr++)
ds@1952	830	{
ds@1952	831	if (!in_quote)
ds@1952	832	{
ds@1952	833	switch (*ptr)
ds@1952	834	{
ds@1952	835	case '"':
ds@1952	836	{
ds@1952	837	in_quote = true;
ds@1952	838	break;
ds@1952	839	}
ds@1952	840	case ';':
ds@1952	841	{
ds@1952	842	/* Stop the loop */
ds@1952	843	len = i;
ds@1952	844	/* Terminate the string here */
ds@1952	845	*ptr = '\0';
ds@1952	846	break;
ds@1952	847	}
ds@1952	848	}
ds@1952	849	} else {
ds@1952	850	if (*ptr == '"')
ds@1952	851	{
ds@1952	852	in_quote = false;
ds@1952	853	}
ds@1952	854	}
ds@1952	855	}
ds@1952	856
ds@1952	857	if (!len)
ds@1952	858	{
ds@1952	859	continue;
ds@1952	860	}
ds@1952	861
ds@1952	862	ptr = save_ptr;
ds@1952	863
ds@1952	864	/* Lastly, strip ending whitespace off */
ds@1952	865	for (size_t i=len-1; i>=0 && i<len; i--)
ds@1952	866	{
ds@1952	867	if (g_ws_chartable[(unsigned)ptr[i]])
ds@1952	868	{
ds@1952	869	ptr[i] = '\0';
ds@1952	870	len--;
ds@1952	871	} else {
ds@1952	872	break;
ds@1952	873	}
ds@1952	874	}
ds@1952	875
ds@1952	876	if (!len)
ds@1952	877	{
ds@1952	878	continue;
ds@1952	879	}
ds@1952	880
ds@1952	881	if (!ini_listener->ReadINI_RawLine(ptr, &curtok))
ds@1952	882	{
ds@1952	883	goto event_failed;
ds@1952	884	}
ds@1952	885
ds@1952	886	if (*ptr == '[')
ds@1952	887	{
ds@1952	888	bool invalid_tokens = false;
ds@1952	889	bool got_bracket = false;
ds@1952	890	bool extra_tokens = false;
ds@1952	891	char c;
ds@1952	892	bool alnum;
ds@1952	893	wchar_t wc;
ds@1952	894
ds@1952	895	for (size_t i=1; i<len; i++)
ds@1952	896	{
ds@1952	897	c = ptr[i];
ds@1952	898	alnum = false;
ds@1952	899
ds@1952	900	if (c & (1<<7))
ds@1952	901	{
ds@1952	902	if (mbtowc(&wc, &ptr[i], len-i) != -1)
ds@1952	903	{
ds@1952	904	alnum = (iswalnum(wc) != 0);
ds@1952	905	i += _GetUTF8CharBytes(&ptr[i]) - 1;
ds@1952	906	}
ds@1952	907	} else {
ds@1952	908	alnum = (isalnum(c) != 0) \|\| (g_ini_chartable1[(unsigned)c] != 0);
ds@1952	909	}
ds@1952	910	if (!alnum)
ds@1952	911	{
ds@1952	912	/* First check - is this a bracket? */
ds@1952	913	if (c == ']')
ds@1952	914	{
ds@1952	915	/* Yes! */
ds@1952	916	got_bracket = true;
ds@1952	917	/* If this isn't the last character... */
ds@1952	918	if (i != len - 1)
ds@1952	919	{
ds@1952	920	extra_tokens = true;
ds@1952	921	}
ds@1952	922	/* terminate */
ds@1952	923	ptr[i] = '\0';
ds@1952	924	break;
ds@1952	925	} else {
ds@1952	926	/* n...No! Continue copying. */
ds@1952	927	invalid_tokens = true;
ds@1952	928	}
ds@1952	929	}
ds@1952	930	}
ds@1952	931
ds@1952	932	/* Tell the handler */
ds@1952	933	if (!ini_listener->ReadINI_NewSection(&ptr[1], invalid_tokens, got_bracket, extra_tokens, &curtok))
ds@1952	934	{
ds@1952	935	goto event_failed;
ds@1952	936	}
ds@1952	937	} else {
ds@1952	938	char *key_ptr = ptr;
ds@1952	939	char *val_ptr = NULL;
ds@1952	940	char c;
ds@1952	941	size_t first_space = 0;
ds@1952	942	bool invalid_tokens = false;
ds@1952	943	bool equal_token = false;
ds@1952	944	bool quotes = false;
ds@1952	945	bool alnum;
ds@1952	946	wchar_t wc;
ds@1952	947
ds@1952	948	for (size_t i=0; i<len; i++)
ds@1952	949	{
ds@1952	950	c = ptr[i];
ds@1952	951	alnum = false;
ds@1952	952	/* is this an invalid char? */
ds@1952	953	if (c & (1<<7))
ds@1952	954	{
ds@1952	955	if (mbtowc(&wc, &ptr[i], len-i) != -1)
ds@1952	956	{
ds@1952	957	alnum = (iswalnum(wc) != 0);
ds@1952	958	i += _GetUTF8CharBytes(&ptr[i]) - 1;
ds@1952	959	}
ds@1952	960	} else {
ds@1952	961	alnum = (isalnum(c) != 0) \|\| (g_ini_chartable1[(unsigned)c] != 0);
ds@1952	962	}
ds@1952	963
ds@1952	964	if (!alnum)
ds@1952	965	{
ds@1952	966	if (g_ws_chartable[(unsigned)c])
ds@1952	967	{
ds@1952	968	/* if it's a space, keep track of the first occurring space */
ds@1952	969	if (!first_space)
ds@1952	970	{
ds@1952	971	first_space = i;
ds@1952	972	}
ds@1952	973	} else {
ds@1952	974	if (c == '=')
ds@1952	975	{
ds@1952	976	/* if it's an equal sign, we're done with the key */
ds@1952	977	if (first_space)
ds@1952	978	{
ds@1952	979	/* remove excess whitespace */
ds@1952	980	key_ptr[first_space] = '\0';
ds@1952	981	} else {
ds@1952	982	/* remove the equal sign */
ds@1952	983	key_ptr[i] = '\0';
ds@1952	984	}
ds@1952	985	if (ptr[++i] != '\0')
ds@1952	986	{
ds@1952	987	/* If this isn't the end, set next pointer */
ds@1952	988	val_ptr = &ptr[i];
ds@1952	989	}
ds@1952	990	equal_token = true;
ds@1952	991	break;
ds@1952	992	} else {
ds@1952	993	/* Mark that we got something invalid! */
ds@1952	994	invalid_tokens = true;
ds@1952	995	first_space = 0;
ds@1952	996	}
ds@1952	997	}
ds@1952	998	}
ds@1952	999	}
ds@1952	1000
ds@1952	1001	/* Now we need to parse the value, if any */
ds@1952	1002	if (val_ptr)
ds@1952	1003	{
ds@1952	1004	/* eat up spaces! there shouldn't be any h*/
ds@1952	1005	while ((val_ptr != '\0') && g_ws_chartable[(unsigned)val_ptr] != 0)
ds@1952	1006	{
ds@1952	1007	val_ptr++;
ds@1952	1008	}
ds@1952	1009	if (*val_ptr == '\0')
ds@1952	1010	{
ds@1952	1011	val_ptr = NULL;
ds@1952	1012	goto skip_value;
ds@1952	1013	}
ds@1952	1014	/* Do we have an initial quote? If so, the parsing rules change! */
ds@1952	1015	if (val_ptr == '"' && val_ptr != '\0')
ds@1952	1016	{
ds@1952	1017	len = strlen(val_ptr);
ds@1952	1018	if (val_ptr[len-1] == '"')
ds@1952	1019	{
ds@1952	1020	/* Strip quotes! */
ds@1952	1021	val_ptr[--len] = '\0';
ds@1952	1022	val_ptr++;
ds@1952	1023	quotes = true;
ds@1952	1024	}
ds@1952	1025	}
ds@1952	1026	}
ds@1952	1027	skip_value:
ds@1952	1028	/* We're done! */
ds@1952	1029	curtok = val_ptr - buffer;
ds@1952	1030	if (!ini_listener->ReadINI_KeyValue(key_ptr, val_ptr, invalid_tokens, equal_token, quotes, &curtok))
ds@1952	1031	{
ds@1952	1032	curtok = 0;
ds@1952	1033	goto event_failed;
ds@1952	1034	}
ds@1952	1035	}
ds@1952	1036	}
ds@1952	1037
ds@1952	1038	if (line)
ds@1952	1039	{
ds@1952	1040	*line = curline;
ds@1952	1041	}
ds@1952	1042
ds@1952	1043	fclose(fp);
ds@1952	1044
ds@1952	1045	return true;
ds@1952	1046
ds@1952	1047	event_failed:
ds@1952	1048	if (line)
ds@1952	1049	{
ds@1952	1050	*line = curline;
ds@1952	1051	}
ds@1952	1052
ds@1952	1053	if (col)
ds@1952	1054	{
ds@1952	1055	*col = curtok;
ds@1952	1056	}
ds@1952	1057
ds@1952	1058	fclose(fp);
ds@1952	1059
ds@1952	1060	return false;
ds@1952	1061	}
ds@1952	1062
ds@1952	1063	const char *TextParsers::GetSMCErrorString(SMCError err)
ds@1952	1064	{
ds@1952	1065	static const char *s_errors[] =
ds@1952	1066	{
ds@1952	1067	NULL,
ds@1952	1068	"Stream failed to open",
ds@1952	1069	"Stream returned read error",
ds@1952	1070	NULL,
ds@1952	1071	"Un-quoted section has invalid tokens",
ds@1952	1072	"Section declared without header",
ds@1952	1073	"Section declared with unknown tokens",
ds@1952	1074	"Section ending without a matching section beginning",
ds@1952	1075	"Section beginning without a matching ending",
ds@1952	1076	"Line contained too many invalid tokens",
ds@1952	1077	"Token buffer overflowed",
ds@1952	1078	"A property was declared outside of a section",
ds@1952	1079	};
ds@1952	1080
ds@1952	1081	if (err < SMCError_Okay \|\| err > SMCError_InvalidProperty1)
ds@1952	1082	{
ds@1952	1083	return NULL;
ds@1952	1084	}
ds@1952	1085
ds@1952	1086	return s_errors[err];
ds@1952	1087	}

author	Scott Ehlert <ds@alliedmods.net>
	Wed Jun 11 08:07:17 2008 +0000 (2008-06-11)
changeset 2112	2a83a521287c
parent 1952	b7aa2b626807
child 2289	44045ead61cd
permissions	-rw-r--r--