/**

 * vim: set ts=4 :

 * =============================================================================

 * SourceMod

 * Copyright (C) 2004-2007 AlliedModders LLC.  All rights reserved.

 * =============================================================================

 *

 * This program is free software; you can redistribute it and/or modify it under

 * the terms of the GNU General Public License, version 3.0, as published by the

 * Free Software Foundation.

 * 

 * This program is distributed in the hope that it will be useful, but WITHOUT

 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS

 * FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more

 * details.

 *

 * You should have received a copy of the GNU General Public License along with

 * this program.  If not, see <http://www.gnu.org/licenses/>.

 *

 * As a special exception, AlliedModders LLC gives you permission to link the

 * code of this program (as well as its derivative works) to "Half-Life 2," the

 * "Source Engine," the "SourcePawn JIT," and any Game MODs that run on software

 * by the Valve Corporation.  You must obey the GNU General Public License in

 * all respects for all other code used.  Additionally, AlliedModders LLC grants

 * this exception to all derivative works.  AlliedModders LLC defines further

 * exceptions, found in LICENSE.txt (as of this writing, version JULY-31-2007),

 * or <http://www.sourcemod.net/license.php>.

 *

 * Version: $Id$

 */

#include <stdarg.h>

#include <stdlib.h>

#include <ctype.h>

#include "Translator.h"

#include "Logger.h"

#include "LibrarySys.h"

#include "sm_stringutil.h"

#include "sourcemod.h"

#include "PlayerManager.h"

Translator g_Translator;

CPhraseFile *g_pCorePhrases = NULL;

unsigned int g_pCorePhraseID = 0;

struct trans_t

{

	int stridx;

	int fmt_order;

};

struct phrase_t

{

	int fmt_list;

	unsigned int fmt_count;

	unsigned int fmt_bytes;

	int trans_tbl;

	unsigned int translations;

};

CPhraseFile::CPhraseFile(Translator *pTranslator, const char *file)

{

	m_pStringTab = pTranslator->GetStringTable();

	m_pMemory = m_pStringTab->GetMemTable();

	m_LangCount = pTranslator->GetLanguageCount();

	m_File.assign(file);

	m_pTranslator = pTranslator;

	m_pPhraseLookup = NULL;

}

CPhraseFile::~CPhraseFile()

{

	if (m_pPhraseLookup)

	{

		sm_trie_destroy(m_pPhraseLookup);

	}

}

void CPhraseFile::ParseError(const char *message, ...)

{

	char buffer[1024];

	va_list ap;

	va_start(ap, message);

	vsnprintf(buffer, sizeof(buffer), message, ap);

	va_end(ap);

	m_ParseError.assign(buffer);

}

void CPhraseFile::ParseWarning(const char *message, ...)

{

	char buffer[1024];

	va_list ap;

	va_start(ap, message);

	vsnprintf(buffer, sizeof(buffer), message, ap);

	va_end(ap);

	if (!m_FileLogged)

	{

		g_Logger.LogError("[SM] Warning(s) encountered in translation file \"%s\"", m_File.c_str());

		m_FileLogged = true;

	}

	g_Logger.LogError("[SM] %s", message);

}

void CPhraseFile::ReparseFile()

{

	if (m_pPhraseLookup)

	{

		sm_trie_destroy(m_pPhraseLookup);

	}

	m_pPhraseLookup = sm_trie_create();

	m_LangCount = m_pTranslator->GetLanguageCount();

	if (!m_LangCount)

	{

		return;

	}

	SMCError err;

	char path[PLATFORM_MAX_PATH];

	g_SourceMod.BuildPath(Path_SM, path, PLATFORM_MAX_PATH, "translations/%s", m_File.c_str());

	//backwards compatibility shim

	/* :HACKHACK: Change .cfg/.txt and vice versa for compatibility */

	if (!g_LibSys.PathExists(path))

	{

		if (m_File.compare("common.cfg") == 0)

		{

			UTIL_ReplaceAll(path, sizeof(path), "common.cfg", "common.phrases.txt");

		} else if (strstr(path, ".cfg")) {

			UTIL_ReplaceAll(path, sizeof(path), ".cfg", ".txt");

		} else if (strstr(path, ".txt")) {

			UTIL_ReplaceAll(path, sizeof(path), ".txt", ".cfg");

		}

	}

	SMCStates states;

	if ((err=textparsers->ParseFile_SMC(path, this, &states)) != SMCError_Okay)

	{

		const char *msg = textparsers->GetSMCErrorString(err);

		if (!msg)

		{

			msg = m_ParseError.c_str();

		}

		g_Logger.LogError("[SM] Fatal error encountered parsing translation file \"%s\"", m_File.c_str());

		g_Logger.LogError("[SM] Error (line %d, column %d): %s", states.line, states.col, msg);

	}

}

void CPhraseFile::ReadSMC_ParseStart()

{

	m_CurPhrase = -1;

	m_ParseState = PPS_None;

	m_FileLogged = false;

	m_LastPhraseString.clear();

}

SMCResult CPhraseFile::ReadSMC_NewSection(const SMCStates *states, const char *name)

{

	bool recognized = false;

	if (m_ParseState == PPS_None)

	{

		if (strcmp(name, "Phrases") == 0)

		{

			m_ParseState = PPS_Phrases;

			recognized = true;

		}

	} 

	else if (m_ParseState == PPS_Phrases) 

	{

		m_ParseState = PPS_InPhrase;

		recognized = true;

		phrase_t *pPhrase;

		m_CurPhrase = m_pMemory->CreateMem(sizeof(phrase_t), (void **)&pPhrase);

		/* Create the reverse lookup */

		if (!sm_trie_insert(m_pPhraseLookup, name, reinterpret_cast<void *>(m_CurPhrase)))

		{

			ParseWarning("Skipping duplicate phrase \"%s\" on line %d.", name, states->line);

			/* :IDEA: prevent memory waste by seeking backwards in memtable? */

			m_CurPhrase = -1;

		} 

		else 

		{

			/* Initialize new phrase */

			trans_t *pTrans;

			pPhrase->fmt_count = 0;

			pPhrase->fmt_list = -1;

			int trans_tbl = m_pMemory->CreateMem(sizeof(trans_t) * m_LangCount, (void **)&pTrans);

			pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

			pPhrase->trans_tbl = trans_tbl;

			pPhrase->translations = 0;

			pPhrase->fmt_bytes = 0;

			for (unsigned int i=0; i<m_LangCount; i++)

			{

				pTrans[i].stridx = -1;

			}

			m_LastPhraseString.assign(name);

		}

	} 

	else if (m_ParseState == PPS_InPhrase) 

	{

		ParseError("Phrase sections may not have sub-sections");

		return SMCResult_HaltFail;

	}

	if (!recognized)

	{

		ParseWarning("Ignoring invalid section \"%s\" on line %d.", name, states->line);

	}

	return SMCResult_Continue;

}

SMCResult CPhraseFile::ReadSMC_KeyValue(const SMCStates *states, const char *key, const char *value)

{

	/* See if we are ignoring a phrase */

	if (m_CurPhrase == -1)

	{

		return SMCResult_Continue;

	}

	phrase_t *pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

	if (key[0] == '#' && strcmp(key, "#format") == 0)

	{

		if (pPhrase->fmt_list != -1)

		{

			ParseWarning("Ignoring duplicated #format property on line %d", states->line);

			return SMCResult_Continue;

		}

		if (pPhrase->translations > 0)

		{

			ParseWarning("#format property should come before translations on line %d, ignoring", states->line);

			return SMCResult_Continue;

		}

		/* Do an initial browsing for error checking and what not */

		enum ParseStates

		{

			Parse_None,

			Parse_Index,

			Parse_Format,

		};

		ParseStates state = Parse_None;

		const char *old_value = value;

		const char *last_value = value;

		while (*value != '\0')

		{

			if (state == Parse_None)

			{

				if (*value == '{')

				{

					pPhrase->fmt_count++;

					state = Parse_Index;

				} 

				else if (*value == ',') 

				{

					/* Do nothing */

				} 

				else 

				{

					unsigned int bytes = textparsers->GetUTF8CharBytes(value);

					if (bytes != 1 || !isalpha(*value))

					{

						ParseWarning("Invalid token '%c' in #format property on line %d.", *value, states->line);

					}

				}

			} 

			else if (state == Parse_Index) 

			{

				if (*value == ':')

				{

					state = Parse_Format;

					if (value - last_value >= 15)

					{

						ParseWarning("Too many digits in format index on line %d, phrase will be ignored.", states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

				} 

				else 

				{

					unsigned int bytes = textparsers->GetUTF8CharBytes(value);

					if (bytes != 1 || !isdigit(*value))

					{

						ParseWarning("Token '%c' in #format property on line %d is not a digit, phrase will be ignored.",

									*value,

									states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

				}

			} 

			else if (state == Parse_Format) 

			{

				if (*value == '}')

				{

					state = Parse_None;

					last_value = value + 1;

				}

			}

			value++;

		}

		if (state != Parse_None)

		{

			/* Moose clam cow. */

			ParseWarning("Unterminated format string on line %d, phrase will be ignored.", states->line);

			m_CurPhrase = -1;

			return SMCResult_Continue;

		}

		value = old_value;

		/* Allocate the format table */

		char fmt_buf[16];

		int *fmt_list;

		int tmp = m_pMemory->CreateMem(sizeof(int) * pPhrase->fmt_count, (void **)&fmt_list);

		/* Update the phrase pointer in case it changed */

		pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

		pPhrase->fmt_list = tmp;

		/* Initialize */

		for (size_t i=0; i<pPhrase->fmt_count; i++)

		{

			fmt_list[i] = -1;

		}

		/* Now we need to read again... this time into the format buffer */

		const char *in_ptr = value;

		const char *idx_ptr = NULL;

		char *out_ptr = NULL;

		unsigned int cur_idx = 0;

		state = Parse_None;

		while (*in_ptr != '\0')

		{

			if (state == Parse_None)

			{

				if (*in_ptr == '{')

				{

					state = Parse_Index;

					idx_ptr = NULL;

				}

			}

			else if (state == Parse_Index) 

			{

				if (*in_ptr == ':')

				{

					/* Check the number! */

					if (!idx_ptr)

					{

						ParseWarning("Format property contains unindexed format string on line %d, phrase will be ignored.", states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

					long idx = strtol(idx_ptr, NULL, 10);

					if (idx < 1 || idx > (long)pPhrase->fmt_count)

					{

						ParseWarning("Format property contains invalid index '%d' on line %d, phrase will be ignored.", idx, states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

					else if (fmt_list[idx - 1] != -1) 

					{

						ParseWarning("Format property contains duplicated index '%d' on line %d, phrase will be ignored.", idx, states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

					cur_idx = (unsigned int)idx;

					state = Parse_Format;

					out_ptr = NULL;

				}

				else if (!idx_ptr) 

				{

					idx_ptr = in_ptr;

				}

			}

			else if (state == Parse_Format) 

			{

				if (*in_ptr == '}')

				{

					if (!out_ptr)

					{

						ParseWarning("Format property contains empty format string on line %d, phrase will be ignored.", states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

					*out_ptr = '\0';

					state = Parse_None;

					/* Now, add this to our table */

					int tmp_idx = m_pStringTab->AddString(fmt_buf);

					/* Update pointers and update necessary variables */

					pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

					pPhrase->fmt_bytes += strlen(fmt_buf);

					fmt_list = (int *)m_pMemory->GetAddress(pPhrase->fmt_list);

					fmt_list[cur_idx - 1] = tmp_idx;

				} 

				else 

				{

					if (!out_ptr)

					{

						out_ptr = fmt_buf;

						*out_ptr++ = '%';

					}

					/* Check length ... */

					if ((unsigned)(out_ptr - fmt_buf) >= sizeof(fmt_buf) - 1)

					{

						ParseWarning("Format property contains format string that exceeds maximum length on line %d, phrase will be ignored.",

									 states->line);

						m_CurPhrase = -1;

						return SMCResult_Continue;

					}

					*out_ptr++ = *in_ptr;

				}

			}

			in_ptr++;

		}

		/* If we've gotten here, we only accepted unique indexes in a range.

		 * Therefore, the list has to be completely filled.  Double check anyway.

		 */

		for (size_t i=0; i<pPhrase->fmt_count; i++)

		{

			if (fmt_list[i] == -1)

			{

				ParseWarning("Format property contains no string for index %d on line %d, phrase will be ignored.",

							 i + 1,

							 states->line);

				m_CurPhrase = -1;

				return SMCResult_Continue;

			}

		}

	} 

	else 

	{

		size_t len = strlen(key);

		if (len != 2)

		{

			ParseWarning("Ignoring translation to invalid language \"%s\" on line %d.", key, states->line);

			return SMCResult_Continue;

		}

		unsigned int lang;

		if (!m_pTranslator->GetLanguageByCode(key, &lang))

		{

			/* Ignore if we don't have a language.

			 * :IDEA: issue a one-time alert?

			 */

			return SMCResult_Continue;

		}

		/* See how many bytes we need for this string, then allocate.

		 * NOTE: THIS SHOULD GUARANTEE THAT WE DO NOT NEED TO NEED TO SIZE CHECK 

		 */

		len = strlen(value) + pPhrase->fmt_bytes + 1;

		char *out_buf;

		int out_idx;

		out_idx = m_pMemory->CreateMem(len, (void **)&out_buf);

		/* Update pointer */

		pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

		int *fmt_order;

		int *fmt_list = (int *)m_pMemory->GetAddress(pPhrase->fmt_list);

		trans_t *pTrans = (trans_t *)m_pMemory->GetAddress(pPhrase->trans_tbl);

		pTrans = &pTrans[lang];

		pTrans->stridx = out_idx;

		bool params[MAX_TRANSLATE_PARAMS];

		/* Build the format order list, if necessary */

		if (fmt_list)

		{

			int tmp = m_pMemory->CreateMem(pPhrase->fmt_count * sizeof(int), (void **)&fmt_order);

			/* Update pointers */

			pPhrase = (phrase_t *)m_pMemory->GetAddress(m_CurPhrase);

			pTrans = (trans_t *)m_pMemory->GetAddress(pPhrase->trans_tbl);

			fmt_list = (int *)m_pMemory->GetAddress(pPhrase->fmt_list);

			out_buf = (char *)m_pMemory->GetAddress(out_idx);

			pTrans = &pTrans[lang];

			/* Now it's safe to save the index */

			pTrans->fmt_order = tmp;

			for (unsigned int i=0; i<pPhrase->fmt_count; i++)

			{

				fmt_order[i] = -1;

			}

			memset(&params[0], 0, sizeof(params));

		}

		const char *in_ptr = value;

		char *out_ptr = out_buf;

		unsigned int order_idx = 0;

		while (*in_ptr != '\0')

		{

			if (*in_ptr == '\\')

			{

				switch (*(in_ptr + 1))

				{

				case '\n':

					*out_ptr++ = '\n';

					break;

				case '\t':

					*out_ptr++ = '\t';

					break;

				case '\r':

					*out_ptr++ = '\r';

					break;

				case '{':

					*out_ptr++ = '{';

					break;

				default:

					/* Copy both bytes since we don't know what's going on */

					*out_ptr++ = *in_ptr++;

					*out_ptr++ = *in_ptr;

					break;

				}

				/* Skip past the last byte read */

				in_ptr++;

			} 

			else if (*in_ptr == '{' && fmt_list != NULL) 

			{

				/* Search for parameters if this is a formatted string */

				const char *scrap_in_point = in_ptr;

				const char *digit_start = ++in_ptr;

				unsigned int bytes;

				while (*in_ptr != '\0')

				{

					bytes = textparsers->GetUTF8CharBytes(in_ptr);

					if (bytes != 1)

					{

						goto scrap_point;

					}

					if (*in_ptr == '}')

					{

						/* Did we get an index? */

						if (in_ptr == digit_start)

						{

							goto scrap_point;

						}

						/* Is it valid? */

						long idx = strtol(digit_start, NULL, 10);

						if (idx < 1 || idx > (int)pPhrase->fmt_count)

						{

							goto scrap_point;

						}

						if (params[idx-1])

						{

							goto scrap_point;

						}

						/* We're safe to insert the string.  First mark order. */

						fmt_order[order_idx++] = (int)idx - 1;

						/* Now concatenate */

						out_ptr += sprintf(out_ptr, "%s", m_pStringTab->GetString(fmt_list[idx-1]));