/* $Id: USBIdDatabaseGenerator.cpp 59733 2016-02-19 01:42:18Z vboxsync $ */
/** @file
 * USB device vendor and product ID database - generator.
 */

/*
 * Copyright (C) 2015-2016 Oracle Corporation
 *
 * This file is part of VirtualBox Open Source Edition (OSE), as
 * available from http://www.virtualbox.org. This file is free software;
 * you can redistribute it and/or modify it under the terms of the GNU
 * General Public License (GPL) as published by the Free Software
 * Foundation, in version 2 as it comes in the "COPYING" file of the
 * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
 * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
 */


/*********************************************************************************************************************************
*   Header Files                                                                                                                 *
*********************************************************************************************************************************/
#include <stdio.h>

#include <fstream>
#include <iostream>
#include <iomanip>
#include <algorithm>
#include <map>
#include <string>
#include <vector>

#include <iprt/initterm.h>
#include <iprt/message.h>
#include <iprt/string.h>
#include <iprt/stream.h>

#include "../include/USBIdDatabase.h"


/*
 * Include the string table generator.
 */
#define BLDPROG_STRTAB_MAX_STRLEN  (USB_ID_DATABASE_MAX_STRING - 1)
#ifdef USB_ID_DATABASE_WITH_COMPRESSION
# define BLDPROG_STRTAB_WITH_COMPRESSION
#else
# undef  BLDPROG_STRTAB_WITH_COMPRESSION
#endif
#define BLDPROG_STRTAB_WITH_CAMEL_WORDS
#undef  BLDPROG_STRTAB_PURE_ASCII
#include <iprt/bldprog-strtab-template.cpp.h>



/*********************************************************************************************************************************
*   Global Variables                                                                                                             *
*********************************************************************************************************************************/
/** For verbose output.   */
static bool g_fVerbose = false;
/** Output prefix for informational output. */
#define INFO_PREF "USBIdDatabaseGenerator: Info: "


using namespace std;

static const char * const header =
    "/** @file\n"
    " * USB device vendor and product ID database - Autogenerated from <stupid C++ cannot do %s>\n"
    " */\n"
    "\n"
    "/*\n"
    " * Copyright (C) 2015-2016 Oracle Corporation\n"
    " *\n"
    " * This file is part of VirtualBox Open Source Edition(OSE), as\n"
    " * available from http ://www.virtualbox.org. This file is free software;\n"
    " * you can redistribute it and / or modify it under the terms of the GNU\n"
    " * General Public License(GPL) as published by the Free Software\n"
    " * Foundation, in version 2 as it comes in the \"COPYING\" file of the\n"
    " * VirtualBox OSE distribution.VirtualBox OSE is distributed in the\n"
    " * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.\n"
    " */"
    "\n"
    "\n"
    "#include \"USBIdDatabase.h\"\n"
    "\n";
static const char * const product_header =
    "/**\n"
    " * USB devices aliases array.\n"
    " * Format: VendorId, ProductId, Vendor Name, Product Name\n"
    " * The source of the list is http://www.linux-usb.org/usb.ids\n"
    " */\n"
    "USBIDDBPROD const USBIdDatabase::s_aProducts[] =\n"
    "{\n";
const char * const product_part2 =
    "};\n"
    "\n"
    "\nconst RTBLDPROGSTRREF USBIdDatabase::s_aProductNames[] =\n"
    "{\n";
const char * const product_footer =
    "};\n"
    "\n"
    "const size_t USBIdDatabase::s_cProducts = RT_ELEMENTS(USBIdDatabase::s_aProducts);\n";

const char * const vendor_header =
    "\nUSBIDDBVENDOR const USBIdDatabase::s_aVendors[] =\n"
    "{\n";
const char * const vendor_part2 =
    "};\n"
    "\n"
    "\nconst RTBLDPROGSTRREF USBIdDatabase::s_aVendorNames[] =\n"
    "{\n";
const char * const vendor_footer =
    "};\n"
    "\n"
    "const size_t USBIdDatabase::s_cVendors = RT_ELEMENTS(USBIdDatabase::s_aVendors);\n";

const char * const start_block = "# Vendors, devices and interfaces. Please keep sorted.";
const char * const end_block = "# List of known device classes, subclasses and protocols";


/*********************************************************************************************************************************
*   Defined Constants And Macros                                                                                                 *
*********************************************************************************************************************************/
// error codes (complements RTEXITCODE_XXX).
#define ERROR_OPEN_FILE         (12)
#define ERROR_IN_PARSE_LINE     (13)
#define ERROR_DUPLICATE_ENTRY   (14)
#define ERROR_WRONG_FILE_FORMAT (15)
#define ERROR_TOO_MANY_PRODUCTS (16)


/*********************************************************************************************************************************
*   Structures and Typedefs                                                                                                      *
*********************************************************************************************************************************/
struct VendorRecord
{
    size_t vendorID;
    size_t iProduct;
    size_t cProducts;
    std::string str;
    BLDPROGSTRING StrRef;
};

struct ProductRecord
{
    size_t key;
    size_t vendorID;
    size_t productID;
    std::string str;
    BLDPROGSTRING StrRef;
};


namespace State
{
    typedef int Value;
    enum
    {
        lookForStartBlock,
        lookForEndBlock,
        finished
    };
}

typedef vector<ProductRecord> ProductsSet;
typedef vector<VendorRecord> VendorsSet;


/*********************************************************************************************************************************
*   Global Variables                                                                                                             *
*********************************************************************************************************************************/
ProductsSet g_products;
VendorsSet g_vendors;

/** The size of all the raw strings, including terminators. */
static size_t g_cbRawStrings = 0;



bool operator < (const ProductRecord& lh, const ProductRecord& rh)
{
    return lh.key < rh.key;
}

bool operator < (const VendorRecord& lh, const VendorRecord& rh)
{
    return lh.vendorID < rh.vendorID;
}

bool operator == (const ProductRecord& lh, const ProductRecord& rh)
{
    return lh.key == rh.key;
}

bool operator == (const VendorRecord& lh, const VendorRecord& rh)
{
    return lh.vendorID == rh.vendorID;
}


/*
 * Input file parsing.
 */
int ParseAlias(const string& src, size_t& id, string& desc)
{
    unsigned int i = 0;
    if (sscanf(src.c_str(), "%x", &i) != 1)
        return ERROR_IN_PARSE_LINE;

    /* skip the number and following whitespace. */
    size_t offNext = src.find_first_of(" \t", 1);
    offNext = src.find_first_not_of(" \t", offNext);
    if (offNext != string::npos)
    {
        size_t cchLength = src.length() - offNext;
        if (cchLength <= USB_ID_DATABASE_MAX_STRING)
        {
            id = i;
            desc = src.substr(offNext);

            /* Check the string encoding. */
            int rc = RTStrValidateEncoding(desc.c_str());
            if (RT_SUCCESS(rc))
            {
                g_cbRawStrings += desc.length() + 1;
                return RTEXITCODE_SUCCESS;
            }

            RTMsgError("Invalid encoding: '%s' (rc=%Rrc)", desc.c_str(), rc);
        }
        else
            RTMsgError("String to long: %zu", cchLength);
    }
    else
        RTMsgError("Error parsing '%s'", src.c_str());
    return ERROR_IN_PARSE_LINE;
}

bool IsCommentOrEmptyLine(const string& str)
{
    size_t index = str.find_first_not_of(" \t");// skip left spaces
    return index == string::npos || str[index] == '#';
}

bool getline(PRTSTREAM instream, string& resString)
{
    const size_t szBuf = 4096;
    char buf[szBuf] = { 0 };

    int rc = RTStrmGetLine(instream, buf, szBuf);
    if (RT_SUCCESS(rc))
    {
        resString = buf;
        return true;
    }

    if (rc != VERR_EOF)
        RTMsgWarning("RTStrmGetLine failed: %Rrc", rc);
    return false;
}

int ParseUsbIds(PRTSTREAM instream)
{
    State::Value state = State::lookForStartBlock;
    string line;
    int res = 0;
    VendorRecord vendor = { 0, 0, 0, "" };

    while (state != State::finished && getline(instream, line))
    {
        switch (state)
        {
            case State::lookForStartBlock:
            {
                if (line.find(start_block) != string::npos)
                    state = State::lookForEndBlock;
                break;
            }
            case State::lookForEndBlock:
            {
                if (line.find(end_block) != string::npos)
                    state = State::finished;
                else
                {
                    if (!IsCommentOrEmptyLine(line))
                    {
                        if (line[0] == '\t')
                        {
                            // Parse Product line
                            // first line should be vendor
                            if (vendor.vendorID == 0)
                                return RTMsgErrorExit((RTEXITCODE)ERROR_WRONG_FILE_FORMAT,
                                                      "Wrong file format. Product before vendor: '%s'", line.c_str());
                            ProductRecord product = { 0, vendor.vendorID, 0, "" };
                            if (ParseAlias(line.substr(1), product.productID, product.str) != 0)
                                return RTMsgErrorExit((RTEXITCODE)ERROR_IN_PARSE_LINE,
                                                      "Error in parsing product line: '", line.c_str());
                            product.key = RT_MAKE_U32(product.productID, product.vendorID);
                            Assert(product.vendorID == vendor.vendorID);
                            g_products.push_back(product);
                        }
                        else
                        {
                            // Parse vendor line
                            if (ParseAlias(line, vendor.vendorID, vendor.str) != 0)
                                return RTMsgErrorExit((RTEXITCODE)ERROR_IN_PARSE_LINE,
                                                      "Error in parsing vendor line: '", line.c_str());
                            g_vendors.push_back(vendor);
                        }
                    }
                }
                break;
            }
        }
    }
    if (state == State::lookForStartBlock)
        return RTMsgErrorExit((RTEXITCODE)ERROR_WRONG_FILE_FORMAT,
                              "wrong format of input file. Start line is not found.");
    return 0;
}


static int usage(FILE *pOut, const char *argv0)
{
    fprintf(pOut, "Usage: %s [linux.org usb list file] [custom usb list file] [-o output file]\n", argv0);
    return RTEXITCODE_SYNTAX;
}


int main(int argc, char *argv[])
{
    /*
     * Initialize IPRT and convert argv to UTF-8.
     */
    int rc = RTR3InitExe(argc, &argv, 0);
    if (RT_FAILURE(rc))
        return RTMsgInitFailure(rc);

    /*
     * Parse arguments and read input files.
     */
    if (argc < 4)
    {
        usage(stderr, argv[0]);
        return RTMsgErrorExit(RTEXITCODE_SYNTAX, "Insufficient arguments.");
    }
    g_products.reserve(20000);
    g_vendors.reserve(3500);

    const char *pszOutFile = NULL;
    for (int i = 1; i < argc; i++)
    {
        if (strcmp(argv[i], "-o") == 0)
        {
            pszOutFile = argv[++i];
            continue;
        }
        if (   strcmp(argv[i], "-h") == 0
            || strcmp(argv[i], "-?") == 0
            || strcmp(argv[i], "--help") == 0)
        {
            usage(stdout, argv[0]);
            return RTEXITCODE_SUCCESS;
        }

        PRTSTREAM pInStrm;
        rc = RTStrmOpen(argv[i], "r", &pInStrm);
        if (RT_FAILURE(rc))
            return RTMsgErrorExit((RTEXITCODE)ERROR_OPEN_FILE,
                                  "Failed to open file '%s' for reading: %Rrc", argv[i], rc);

        rc = ParseUsbIds(pInStrm);
        RTStrmClose(pInStrm);
        if (rc != 0)
        {
            RTMsgError("Failed parsing USB devices file '%s'", argv[i]);
            return rc;
        }
    }

    /*
     * Due to USBIDDBVENDOR::iProduct, there is currently a max of 64KB products.
     * (Not a problem as we've only have less that 54K products currently.)
     */
    if (g_products.size() > _64K)
        return RTMsgErrorExit((RTEXITCODE)ERROR_TOO_MANY_PRODUCTS,
                              "More than 64K products is not supported: %u products", g_products.size());

    /*
     * Sort the IDs and fill in the iProduct and cProduct members.
     */
    sort(g_products.begin(), g_products.end());
    sort(g_vendors.begin(), g_vendors.end());

    size_t iProduct = 0;
    for (size_t iVendor = 0; iVendor < g_vendors.size(); iVendor++)
    {
        size_t const idVendor = g_vendors[iVendor].vendorID;
        g_vendors[iVendor].iProduct = iProduct;
        if (   iProduct < g_products.size()
            && g_products[iProduct].vendorID <= idVendor)
        {
            if (g_products[iProduct].vendorID == idVendor)
                do
                    iProduct++;
                while (   iProduct < g_products.size()
                       && g_products[iProduct].vendorID == idVendor);
            else
                return RTMsgErrorExit((RTEXITCODE)ERROR_IN_PARSE_LINE, "product without vendor after sorting. impossible!");
        }
        g_vendors[iVendor].cProducts = iProduct - g_vendors[iVendor].iProduct;
    }

    /*
     * Verify that all IDs are unique.
     */
    ProductsSet::iterator ita = adjacent_find(g_products.begin(), g_products.end());
    if (ita != g_products.end())
        return RTMsgErrorExit((RTEXITCODE)ERROR_DUPLICATE_ENTRY, "Duplicate alias detected: idProduct=%#06x", ita->productID);

    /*
     * Build the string table.
     * Do string compression and create the string table.
     */
    BLDPROGSTRTAB StrTab;
    if (!BldProgStrTab_Init(&StrTab, g_products.size() + g_vendors.size()))
        return RTMsgErrorExit(RTEXITCODE_FAILURE, "Out of memory!");

    for (ProductsSet::iterator it = g_products.begin(); it != g_products.end(); ++it)
    {
        it->StrRef.pszString = (char *)it->str.c_str();
        BldProgStrTab_AddString(&StrTab, &it->StrRef);
    }
    for (VendorsSet::iterator it = g_vendors.begin(); it != g_vendors.end(); ++it)
    {
        it->StrRef.pszString = (char *)it->str.c_str();
        BldProgStrTab_AddString(&StrTab, &it->StrRef);
    }

    if (!BldProgStrTab_CompileIt(&StrTab, g_fVerbose))
        return RTMsgErrorExit(RTEXITCODE_FAILURE, "BldProgStrTab_CompileIt failed!\n");

    /*
     * Print stats.
     */
    size_t const cbVendorEntry  = sizeof(USBIdDatabase::s_aVendors[0]) + sizeof(USBIdDatabase::s_aVendorNames[0]);
    size_t const cbProductEntry = sizeof(USBIdDatabase::s_aProducts[0]) + sizeof(USBIdDatabase::s_aProductNames[0]);

    size_t cbOldRaw = (g_products.size() + g_vendors.size()) * sizeof(const char *) * 2 + g_cbRawStrings;
    size_t cbRaw    = g_vendors.size() * cbVendorEntry + g_products.size() * cbProductEntry + g_cbRawStrings;
    size_t cbActual = g_vendors.size() * cbVendorEntry + g_products.size() * cbProductEntry + StrTab.cchStrTab;
#ifdef USB_ID_DATABASE_WITH_COMPRESSION
    cbActual += sizeof(StrTab.aCompDict);
#endif
    cout << INFO_PREF "Total " << dec << cbActual << " bytes";
    if (cbActual < cbRaw)
        cout << " saving " << dec << ((cbRaw - cbActual) * 100 / cbRaw) << "% (" << (cbRaw - cbActual) << " bytes)";
    else
        cout << " wasting " << dec << (cbActual - cbRaw) << " bytes";
    cout << "; old version " << cbOldRaw << " bytes + relocs ("
         << ((cbOldRaw - cbActual) * 100 / cbOldRaw) << "% save)." << endl;


    /*
     * Produce the source file.
     */
    if (!pszOutFile)
        return RTMsgErrorExit((RTEXITCODE)ERROR_OPEN_FILE, "Output file is not specified.");

    FILE *pOut = fopen(pszOutFile, "w");
    if (!pOut)
        return RTMsgErrorExit((RTEXITCODE)ERROR_OPEN_FILE, "Error opening '%s' for writing", pszOutFile);

    fputs(header, pOut);
    BldProgStrTab_WriteStringTable(&StrTab, pOut, "", "USBIdDatabase::s_", "StrTab");

    fputs(product_header, pOut);
    for (ProductsSet::iterator itp = g_products.begin(); itp != g_products.end(); ++itp)
        fprintf(pOut, "    { 0x%04x },\n", itp->productID);

    fputs(product_part2, pOut);
    for (ProductsSet::iterator itp = g_products.begin(); itp != g_products.end(); ++itp)
        fprintf(pOut, "{ 0x%06x, 0x%02x },\n", itp->StrRef.offStrTab, itp->StrRef.cchString);
    fputs(product_footer, pOut);

    fputs(vendor_header, pOut);
    for (VendorsSet::iterator itv = g_vendors.begin(); itv != g_vendors.end(); ++itv)
        fprintf(pOut, "    { 0x%04x, 0x%04x, 0x%04x },\n", itv->vendorID, itv->iProduct, itv->cProducts);
    fputs(vendor_part2, pOut);
    for (VendorsSet::iterator itv = g_vendors.begin(); itv != g_vendors.end(); ++itv)
        fprintf(pOut, "{ 0x%06x, 0x%02x },\n", itv->StrRef.offStrTab, itv->StrRef.cchString);
    fputs(vendor_footer, pOut);

    if (ferror(pOut))
        return RTMsgErrorExit((RTEXITCODE)ERROR_OPEN_FILE, "Error writing '%s'!", pszOutFile);
    if (fclose(pOut) != 0)
        return RTMsgErrorExit((RTEXITCODE)ERROR_OPEN_FILE, "Error closing '%s'!", pszOutFile);

    return RTEXITCODE_SUCCESS;
}