Added imgui for configuration

1 files changed, 1306 insertions, 0 deletions

diff --git a/vendor/zgui/libs/imgui_test_engine/imgui_te_utils.cpp b/vendor/zgui/libs/imgui_test_engine/imgui_te_utils.cpp
new file mode 100644
index 0000000..ad473a7
--- /dev/null
+++ b/vendor/zgui/libs/imgui_test_engine/imgui_te_utils.cpp
@@ -0,0 +1,1306 @@
+// dear imgui test engine
+// (helpers/utilities. do NOT use this as a general purpose library)
+
+#if defined(_MSC_VER) && !defined(_CRT_SECURE_NO_WARNINGS)
+#define _CRT_SECURE_NO_WARNINGS
+#endif
+
+#include "imgui_te_utils.h"
+#include "imgui.h"
+#include "imgui_internal.h"
+#define STR_IMPLEMENTATION
+#include "thirdparty/Str/Str.h"
+
+#if defined(_WIN32)
+#if !defined(_WINDOWS_)
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+#endif
+#include <shellapi.h>   // ShellExecuteA()
+#include <stdio.h>
+#else
+#include <errno.h>
+#include <unistd.h>
+#endif
+#ifndef _MSC_VER
+#include <sys/types.h>
+#include <sys/stat.h>   // stat()
+#endif
+
+#if defined(__linux) || defined(__linux__) || defined(__MACH__) || defined(__MSL__) || defined(__MINGW32__)
+#include <pthread.h>    // pthread_setname_np()
+#endif
+#include <chrono>       // high_resolution_clock::now()
+#include <thread>       // this_thread::sleep_for()
+
+//-----------------------------------------------------------------------------
+// Hashing Helpers
+//-----------------------------------------------------------------------------
+// - ImHashDecoratedPathParseLiteral() [internal]
+// - ImHashDecoratedPath()
+// - ImFindNextDecoratedPartInPath()
+//-----------------------------------------------------------------------------
+
+// - Parse literals encoded as "$$xxxx/" and incorporate into our hash based on type.
+// - $$ not passed by caller.
+static ImGuiID ImHashDecoratedPathParseLiteral(ImGuiID crc, const unsigned char* str, const unsigned char* str_end, const unsigned char** out_str_remaining)
+{
+    // Parse type (default to int)
+    ImGuiDataType type = ImGuiDataType_S32;
+    if (*str == '(')
+    {
+        // "$$(int)????" where ???? is s32 or u32
+        if (str + 5 < str_end && memcmp(str, "(int)", 5) == 0)
+        {
+            type = ImGuiDataType_S32;
+            str += 5;
+        }
+        // "$$(ptr)0x????" where ???? is ptr size
+        else if (str + 7 < str_end && memcmp(str, "(ptr)0x", 7) == 0)
+        {
+            type = ImGuiDataType_Pointer;
+            str += 7;
+        }
+    }
+
+    // Parse value
+    switch (type)
+    {
+    case ImGuiDataType_S32:
+    {
+        // e.g. "$$(int)123" for s32/u32/ImGuiID, same as PushID(int)
+        int v = 0;
+        {
+            int negative = 0;
+            if (str < str_end && *str == '-') { negative = 1; str++; }
+            if (str < str_end && *str == '+') { str++; }
+            for (char c = *str; str < str_end; c = *(++str))
+            {
+                if (c >= '0' && c <= '9') { v = (v * 10) + (c - '0'); }
+                else break;
+            }
+            if (negative)
+                v = -v;
+        }
+        crc = ~ImHashData(&v, sizeof(int), ~crc);
+        break;
+    }
+    case ImGuiDataType_Pointer:
+    {
+        // e.g. "$$(ptr)0x1234FFFF" for pointers, same as PushID(void*)
+        intptr_t v = 0;
+        {
+            for (char c = *str; str < str_end; c = *(++str))
+            {
+                if (c >= '0' && c <= '9')       { v = (v << 4) + (c - '0'); }
+                else if (c >= 'A' && c <= 'F')  { v = (v << 4) + 10 + (c - 'A'); }
+                else if (c >= 'a' && c <= 'f')  { v = (v << 4) + 10 + (c - 'a'); }
+                else break;
+            }
+        }
+        crc = ~ImHashData(&v, sizeof(void*), ~crc);
+        break;
+    }
+    }
+
+    // "$$xxxx" must always be either end of string, either leading to a next section e.g. "$$xxxx/"
+    IM_ASSERT(str == str_end || *str == '/');
+
+    *out_str_remaining = str;
+    return crc;
+}
+
+// Hash "hello/world" as if it was "helloworld"
+// To hash a forward slash we need to use "hello\\/world"
+//   IM_ASSERT(ImHashDecoratedPath("Hello/world")   == ImHashStr("Helloworld", 0));
+//   IM_ASSERT(ImHashDecoratedPath("Hello\\/world") == ImHashStr("Hello/world", 0));
+//   IM_ASSERT(ImHashDecoratedPath("$$1")           == (n = 1, ImHashData(&n, sizeof(int))));
+// Adapted from ImHash(). Not particularly fast!
+ImGuiID ImHashDecoratedPath(const char* str, const char* str_end, ImGuiID seed)
+{
+    static ImU32 crc32_lut[256] = { 0 };
+    if (!crc32_lut[1])
+    {
+        const ImU32 polynomial = 0xEDB88320;
+        for (ImU32 i = 0; i < 256; i++)
+        {
+            ImU32 crc = i;
+            for (ImU32 j = 0; j < 8; j++)
+                crc = (crc >> 1) ^ (ImU32(-int(crc & 1)) & polynomial);
+            crc32_lut[i] = crc;
+        }
+    }
+
+    // Prefixing the string with / ignore the seed
+    if (str != str_end && str[0] == '/')
+        seed = 0;
+
+    seed = ~seed;
+    ImU32 crc = seed;
+
+    // Focus for non-zero terminated string for consistency
+    if (str_end == NULL)
+        str_end = str + strlen(str);
+
+    bool inhibit_one = false;
+    bool new_section = true;
+    const unsigned char* current = (const unsigned char*)str;
+    while (current < (const unsigned char*)str_end)
+    {
+        const unsigned char c = *current++;
+
+        // Backslash to inhibit special behavior of following character
+        if (c == '\\' && !inhibit_one)
+        {
+            inhibit_one = true;
+            continue;
+        }
+
+        // Forward slashes are ignored unless prefixed with a backward slash
+        if (c == '/' && !inhibit_one)
+        {
+            inhibit_one = false;
+            new_section = true;
+            seed = crc; // Set seed to the new path
+            continue;
+        }
+
+        // $$ at the beginning of a section to encode literals.
+        // - Currently: "$$????" = hash of 1 as int
+        // - May add pointers and other types.
+        if (c == '$' && current[0] == '$' && !inhibit_one && new_section)
+        {
+            crc = ImHashDecoratedPathParseLiteral(crc, current + 1, (const unsigned char*)str_end, &current);
+            continue;
+        }
+
+        // Reset the hash when encountering ###
+        if (c == '#' && current[0] == '#' && current[1] == '#')
+            crc = seed;
+
+        // Hash byte
+        crc = (crc >> 8) ^ crc32_lut[(crc & 0xFF) ^ c];
+
+        inhibit_one = new_section = false;
+    }
+    return ~crc;
+}
+
+// Returns a next element of decorated hash path.
+//    "//hello/world/child" --> "world/child"
+//    "world/child"         --> "child"
+// This is a helper for code needing to do some parsing of individual nodes in a path.
+// Note: we need the (unsigned char*) stuff in order to keep code similar to ImHashDecoratedPath(). They are not really necessary in this function tho.
+const char* ImFindNextDecoratedPartInPath(const char* str, const char* str_end)
+{
+    const unsigned char* current = (const unsigned char*)str;
+    while (*current == '/')
+        current++;
+
+    bool inhibit_one = false;
+    while (true)
+    {
+        if (str_end != NULL && current == (const unsigned char*)str_end)
+            break;
+
+        const unsigned char c = *current++;
+        if (c == 0)
+            break;
+        if (c == '\\' && !inhibit_one)
+        {
+            inhibit_one = true;
+            continue;
+        }
+
+        // Forward slashes are ignored unless prefixed with a backward slash
+        if (c == '/' && !inhibit_one)
+            return (const char*)current;
+
+        inhibit_one = false;
+    }
+    return NULL;
+}
+
+//-----------------------------------------------------------------------------
+// File/Directory Helpers
+//-----------------------------------------------------------------------------
+// - ImFileExist()
+// - ImFileCreateDirectoryChain()
+// - ImFileFindInParents()
+// - ImFileLoadSourceBlurb()
+//-----------------------------------------------------------------------------
+
+#if _WIN32
+static const char IM_DIR_SEPARATOR = '\\';
+
+static void ImUtf8ToWideChar(const char* multi_byte, ImVector<wchar_t>* buf)
+{
+    const int wsize = ::MultiByteToWideChar(CP_UTF8, 0, multi_byte, -1, NULL, 0);
+    buf->resize(wsize);
+    ::MultiByteToWideChar(CP_UTF8, 0, multi_byte, -1, (wchar_t*)buf->Data, wsize);
+}
+#else
+static const char IM_DIR_SEPARATOR = '/';
+#endif
+
+bool ImFileExist(const char* filename)
+{
+    struct stat dir_stat;
+    int ret = stat(filename, &dir_stat);
+    return (ret == 0);
+}
+
+bool ImFileDelete(const char* filename)
+{
+#if _WIN32
+    ImVector<wchar_t> buf;
+    ImUtf8ToWideChar(filename, &buf);
+    return ::DeleteFileW(&buf[0]) == TRUE;
+#else
+    unlink(filename);
+#endif
+    return false;
+}
+
+// Create directories for specified path. Slashes will be replaced with platform directory separators.
+// e.g. ImFileCreateDirectoryChain("aaaa/bbbb/cccc.png")
+// will try to create "aaaa/" then "aaaa/bbbb/".
+bool ImFileCreateDirectoryChain(const char* path, const char* path_end)
+{
+    IM_ASSERT(path != NULL);
+    IM_ASSERT(path[0] != 0);
+
+    if (path_end == NULL)
+        path_end = path + strlen(path);
+
+    // Copy in a local, zero-terminated buffer
+    size_t path_len = (size_t)(path_end - path);
+    char* path_local = (char*)IM_ALLOC(path_len + 1);
+    memcpy(path_local, path, path_len);
+    path_local[path_len] = 0;
+
+#if defined(_WIN32)
+    ImVector<ImWchar> buf;
+#endif
+    // Modification of passed file_name allows us to avoid extra temporary memory allocation.
+    // strtok() pokes \0 into places where slashes are, we create a directory using directory_name and restore slash.
+    for (char* token = strtok(path_local, "\\/"); token != NULL; token = strtok(NULL, "\\/"))
+    {
+        // strtok() replaces slashes with NULLs. Overwrite removed slashes here with the type of slashes the OS needs (win32 functions need backslashes).
+        if (token != path_local)
+            *(token - 1) = IM_DIR_SEPARATOR;
+
+#if defined(_WIN32)
+        // Use ::CreateDirectoryW() because ::CreateDirectoryA() treat filenames in the local code-page instead of UTF-8.
+        const int filename_wsize = ImTextCountCharsFromUtf8(path_local, NULL) + 1;
+        buf.resize(filename_wsize);
+        ImTextStrFromUtf8(&buf[0], filename_wsize, path_local, NULL);
+        if (!::CreateDirectoryW((wchar_t*)&buf[0], NULL) && GetLastError() != ERROR_ALREADY_EXISTS)
+#else
+        if (mkdir(path_local, S_IRWXU) != 0 && errno != EEXIST)
+#endif
+        {
+            IM_FREE(path_local);
+            return false;
+        }
+    }
+    IM_FREE(path_local);
+    return true;
+}
+
+bool ImFileFindInParents(const char* sub_path, int max_parent_count, Str* output)
+{
+    IM_ASSERT(sub_path != NULL);
+    IM_ASSERT(output != NULL);
+    for (int parent_level = 0; parent_level < max_parent_count; parent_level++)
+    {
+        output->clear();
+        for (int j = 0; j < parent_level; j++)
+            output->append("../");
+        output->append(sub_path);
+        if (ImFileExist(output->c_str()))
+            return true;
+    }
+    output->clear();
+    return false;
+}
+
+bool ImFileLoadSourceBlurb(const char* file_name, int line_no_start, int line_no_end, ImGuiTextBuffer* out_buf)
+{
+    size_t file_size = 0;
+    char* file_begin = (char*)ImFileLoadToMemory(file_name, "rb", &file_size, 1);
+    if (file_begin == NULL)
+        return false;
+
+    char* file_end = file_begin + file_size;
+    int line_no = 0;
+    const char* test_src_begin = NULL;
+    const char* test_src_end = NULL;
+    for (const char* p = file_begin; p < file_end; )
+    {
+        line_no++;
+        const char* line_begin = p;
+        const char* line_end = ImStrchrRange(line_begin + 1, file_end, '\n');
+        if (line_end == NULL)
+            line_end = file_end;
+        if (line_no >= line_no_start && line_no <= line_no_end)
+        {
+            if (test_src_begin == NULL)
+                test_src_begin = line_begin;
+            test_src_end = ImMax(test_src_end, line_end);
+        }
+        p = line_end + 1;
+    }
+
+    if (test_src_begin != NULL)
+        out_buf->append(test_src_begin, test_src_end);
+    else
+        out_buf->clear();
+
+    ImGui::MemFree(file_begin);
+    return true;
+}
+
+//-----------------------------------------------------------------------------
+// Path Helpers
+//-----------------------------------------------------------------------------
+// - ImPathFindFilename()
+// - ImPathFindFileExt()
+// - ImPathFixSeparatorsForCurrentOS()
+//-----------------------------------------------------------------------------
+
+const char* ImPathFindFilename(const char* path, const char* path_end)
+{
+    IM_ASSERT(path != NULL);
+    if (!path_end)
+        path_end = path + strlen(path);
+    const char* p = path_end;
+    while (p > path)
+    {
+        if (p[-1] == '/' || p[-1] == '\\')
+            break;
+        p--;
+    }
+    return p;
+}
+
+// "folder/filename" -> return pointer to "" (end of string)
+// "folder/filename.png" -> return pointer to ".png"
+// "folder/filename.png.bak" -> return pointer to ".png.bak"
+const char* ImPathFindExtension(const char* path, const char* path_end)
+{
+    if (!path_end)
+        path_end = path + strlen(path);
+    const char* filename = ImPathFindFilename(path, path_end);
+    const char* p = filename;
+    while (p < path_end)
+    {
+        if (p[0] == '.')
+            break;
+        p++;
+    }
+    return p;
+}
+
+void ImPathFixSeparatorsForCurrentOS(char* buf)
+{
+#ifdef _WIN32
+    for (char* p = buf; *p != 0; p++)
+        if (*p == '/')
+            *p = '\\';
+#else
+    for (char* p = buf; *p != 0; p++)
+        if (*p == '\\')
+            *p = '/';
+#endif
+}
+
+//-----------------------------------------------------------------------------
+// String Helpers
+//-----------------------------------------------------------------------------
+
+static const char* ImStrStr(const char* haystack, size_t hlen, const char* needle, int nlen)
+{
+    const char* end = haystack + hlen;
+    const char* p = haystack;
+    while ((p = (const char*)memchr(p, *needle, end - p)) != NULL)
+    {
+        if (end - p < nlen)
+            return NULL;
+        if (memcmp(p, needle, nlen) == 0)
+            return p;
+        p++;
+    }
+    return NULL;
+}
+
+void ImStrReplace(Str* s, const char* find, const char* repl)
+{
+    IM_ASSERT(find != NULL && *find);
+    IM_ASSERT(repl != NULL);
+    int find_len = (int)strlen(find);
+    int repl_len = (int)strlen(repl);
+    int repl_diff = repl_len - find_len;
+
+    // Estimate required length of new buffer if string size increases.
+    int need_capacity = s->capacity();
+    int num_matches = INT_MAX;
+    if (repl_diff > 0)
+    {
+        num_matches = 0;
+        need_capacity = s->length() + 1;
+        for (char* p = s->c_str(), *end = s->c_str() + s->length(); p != NULL && p < end;)
+        {
+            p = (char*)ImStrStr(p, end - p, find, find_len);
+            if (p)
+            {
+                need_capacity += repl_diff;
+                p += find_len;
+                num_matches++;
+            }
+        }
+    }
+
+    if (num_matches == 0)
+        return;
+
+    const char* not_owned_data = s->owned() ? NULL : s->c_str();
+    if (!s->owned() || need_capacity > s->capacity())
+        s->reserve(need_capacity);
+    if (not_owned_data != NULL)
+        s->set(not_owned_data);
+
+    // Replace data.
+    for (char* p = s->c_str(), *end = s->c_str() + s->length(); p != NULL && p < end && num_matches--;)
+    {
+        p = (char*)ImStrStr(p, end - p, find, find_len);
+        if (p)
+        {
+            memmove(p + repl_len, p + find_len, end - p - find_len + 1);
+            memcpy(p, repl, repl_len);
+            p += repl_len;
+            end += repl_diff;
+        }
+    }
+}
+
+const char* ImStrchrRangeWithEscaping(const char* str, const char* str_end, char find_c)
+{
+    while (str < str_end)
+    {
+        const char c = *str;
+        if (c == '\\')
+        {
+            str += 2;
+            continue;
+        }
+        if (c == find_c)
+            return str;
+        str++;
+    }
+    return NULL;
+}
+
+// Suboptimal but ok for the data size we are dealing with (see commit on 2022/08/22 for a faster and more complicated version)
+void ImStrXmlEscape(Str* s)
+{
+    ImStrReplace(s, "&", "&amp;");
+    ImStrReplace(s, "<", "&lt;");
+    ImStrReplace(s, ">", "&gt;");
+    ImStrReplace(s, "\"", "&quot;");
+    ImStrReplace(s, "\'", "&apos;");
+}
+
+// Based on code from https://github.com/EddieBreeg/C_b64 by @EddieBreeg.
+int ImStrBase64Encode(const unsigned char* src, char* dst, int length)
+{
+    static const char* b64Table = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=";
+    int i, j, k, l, encoded_len = 0;
+
+    while (length > 0)
+    {
+        switch (length)
+        {
+        case 1:
+            i = src[0] >> 2;
+            j = (src[0] & 3) << 4;
+            k = 64;
+            l = 64;
+            break;
+        case 2:
+            i = src[0] >> 2;
+            j = ((src[0] & 3) << 4) | (src[1] >> 4);
+            k = (src[1] & 15) << 2;
+            l = 64;
+            break;
+        default:
+            i = src[0] >> 2;
+            j = ((src[0] & 3) << 4) | (src[1] >> 4);
+            k = ((src[1] & 0xf) << 2) | (src[2] >> 6 & 3);
+            l = src[2] & 0x3f;
+            break;
+        }
+        dst[0] = b64Table[i];
+        dst[1] = b64Table[j];
+        dst[2] = b64Table[k];
+        dst[3] = b64Table[l];
+        src += 3;
+        dst += 4;
+        length -= 3;
+        encoded_len += 4;
+    }
+    return encoded_len;
+}
+
+//-----------------------------------------------------------------------------
+// Parsing Helpers
+//-----------------------------------------------------------------------------
+// - ImParseSplitCommandLine()
+// - ImParseFindIniSection()
+//-----------------------------------------------------------------------------
+
+void    ImParseExtractArgcArgvFromCommandLine(int* out_argc, char const*** out_argv, const char* cmd_line)
+{
+    size_t cmd_line_len = strlen(cmd_line);
+
+    int n = 1;
+    {
+        const char* p = cmd_line;
+        while (*p != 0)
+        {
+            const char* arg = p;
+            while (*arg == ' ')
+                arg++;
+            const char* arg_end = strchr(arg, ' ');
+            if (arg_end == NULL)
+                p = arg_end = cmd_line + cmd_line_len;
+            else
+                p = arg_end + 1;
+            n++;
+        }
+    }
+
+    int argc = n;
+    char const** argv = (char const**)malloc(sizeof(char*) * ((size_t)argc + 1) + (cmd_line_len + 1));
+    IM_ASSERT(argv != NULL);
+    char* cmd_line_dup = (char*)argv + sizeof(char*) * ((size_t)argc + 1);
+    strcpy(cmd_line_dup, cmd_line);
+
+    {
+        argv[0] = "main.exe";
+        argv[argc] = NULL;
+
+        char* p = cmd_line_dup;
+        for (n = 1; n < argc; n++)
+        {
+            char* arg = p;
+            char* arg_end = strchr(arg, ' ');
+            if (arg_end == NULL)
+                p = arg_end = cmd_line_dup + cmd_line_len;
+            else
+                p = arg_end + 1;
+            argv[n] = arg;
+            arg_end[0] = 0;
+        }
+    }
+
+    *out_argc = argc;
+    *out_argv = argv;
+}
+
+bool    ImParseFindIniSection(const char* ini_config, const char* header, ImVector<char>* result)
+{
+    IM_ASSERT(ini_config != NULL);
+    IM_ASSERT(header != NULL);
+    IM_ASSERT(result != NULL);
+
+    size_t ini_len = strlen(ini_config);
+    size_t header_len = strlen(header);
+
+    IM_ASSERT(header_len > 0);
+
+    if (ini_len == 0)
+        return false;
+
+    const char* section_start = strstr(ini_config, header);
+    if (section_start == NULL)
+        return false;
+
+    const char* section_end = strstr(section_start + header_len, "\n[");
+    if (section_end == NULL)
+        section_end = section_start + ini_len;
+
+    // "\n[" matches next header start on all platforms, but it cuts new line marker in half on windows.
+    if (*(section_end - 1) == '\r')
+        --section_end;
+
+    size_t section_len = (size_t)(section_end - section_start);
+    result->resize((int)section_len + 1);
+    ImStrncpy(result->Data, section_start, section_len);
+
+    return true;
+}
+
+//-----------------------------------------------------------------------------
+// Time Helpers
+//-----------------------------------------------------------------------------
+// - ImTimeGetInMicroseconds()
+// - ImTimestampToISO8601()
+//-----------------------------------------------------------------------------
+
+uint64_t ImTimeGetInMicroseconds()
+{
+    // Trying std::chrono out of unfettered optimism that it may actually work..
+    using namespace std;
+    chrono::microseconds ms = chrono::duration_cast<chrono::microseconds>(chrono::high_resolution_clock::now().time_since_epoch());
+    return (uint64_t)ms.count();
+}
+
+void ImTimestampToISO8601(uint64_t timestamp, Str* out_date)
+{
+    time_t unix_time = (time_t)(timestamp / 1000000); // Convert to seconds.
+    tm* time = gmtime(&unix_time);
+    const char* time_format = "%Y-%m-%dT%H:%M:%S";
+    size_t size_req = strftime(out_date->c_str(), out_date->capacity(), time_format, time);
+    if (size_req >= (size_t)out_date->capacity())
+    {
+        out_date->reserve((int)size_req);
+        strftime(out_date->c_str(), out_date->capacity(), time_format, time);
+    }
+}
+
+//-----------------------------------------------------------------------------
+// Threading Helpers
+//-----------------------------------------------------------------------------
+// - ImThreadSleepInMilliseconds()
+// - ImThreadSetCurrentThreadDescription()
+//-----------------------------------------------------------------------------
+
+void ImThreadSleepInMilliseconds(int ms)
+{
+    using namespace std;
+    this_thread::sleep_for(chrono::milliseconds(ms));
+}
+
+#if defined(_MSC_VER)
+// Helper function for setting thread name on Win32
+// This is a separate function because __try cannot coexist with local objects that need destructors called on stack unwind
+static void ImThreadSetCurrentThreadDescriptionWin32OldStyle(const char* description)
+{
+    // Old-style Win32 thread name setting method
+    // See https://docs.microsoft.com/en-us/visualstudio/debugger/how-to-set-a-thread-name-in-native-code
+    const DWORD MS_VC_EXCEPTION = 0x406D1388;
+#pragma pack(push,8)
+    typedef struct tagTHREADNAME_INFO
+    {
+        DWORD dwType; // Must be 0x1000.
+        LPCSTR szName; // Pointer to name (in user addr space).
+        DWORD dwThreadID; // Thread ID (-1=caller thread).
+        DWORD dwFlags; // Reserved for future use, must be zero.
+    } THREADNAME_INFO;
+#pragma pack(pop)
+
+    THREADNAME_INFO info;
+    info.dwType = 0x1000;
+    info.szName = description;
+    info.dwThreadID = (DWORD)-1;
+    info.dwFlags = 0;
+#pragma warning(push)
+#pragma warning(disable: 6320 6322)
+    __try
+    {
+        RaiseException(MS_VC_EXCEPTION, 0, sizeof(info) / sizeof(ULONG_PTR), (ULONG_PTR*)&info);
+    }
+    __except (EXCEPTION_EXECUTE_HANDLER)
+    {
+    }
+#pragma warning(pop)
+}
+#endif // #ifdef _WIN32
+
+// Set the description (name) of the current thread for debugging purposes
+void ImThreadSetCurrentThreadDescription(const char* description)
+{
+#if defined(_MSC_VER) // Windows + Visual Studio
+    // New-style thread name setting
+    // Only supported from Win 10 version 1607/Server 2016 onwards, hence the need for dynamic linking
+
+    typedef HRESULT(WINAPI* SetThreadDescriptionFunc)(HANDLE hThread, PCWSTR lpThreadDescription);
+
+    SetThreadDescriptionFunc set_thread_description = (SetThreadDescriptionFunc)::GetProcAddress(GetModuleHandleA("Kernel32.dll"), "SetThreadDescription");
+    if (set_thread_description)
+    {
+        ImVector<ImWchar> buf;
+        const int description_wsize = ImTextCountCharsFromUtf8(description, NULL) + 1;
+        buf.resize(description_wsize);
+        ImTextStrFromUtf8(&buf[0], description_wsize, description, NULL);
+        set_thread_description(::GetCurrentThread(), (wchar_t*)&buf[0]);
+    }
+
+    // Also do the old-style method too even if the new-style one worked, as the two work in slightly different sets of circumstances
+    ImThreadSetCurrentThreadDescriptionWin32OldStyle(description);
+#elif defined(__linux) || defined(__linux__) || defined(__MINGW32__) // Linux or MingW
+    pthread_setname_np(pthread_self(), description);
+#elif defined(__MACH__) || defined(__MSL__) // OSX
+    pthread_setname_np(description);
+#else
+    // This is a nice-to-have rather than critical functionality, so fail silently if we don't support this platform
+#endif
+}
+
+//-----------------------------------------------------------------------------
+// Build info helpers
+//-----------------------------------------------------------------------------
+// - ImBuildGetCompilationInfo()
+// - ImBuildGetGitBranchName()
+//-----------------------------------------------------------------------------
+
+// Turn __DATE__ "Jan 10 2019" into "2019-01-10"
+static void ImBuildParseDateFromCompilerIntoYMD(const char* in_date, char* out_buf, size_t out_buf_size)
+{
+    char month_str[5];
+    int year, month, day;
+    sscanf(in_date, "%3s %d %d", month_str, &day, &year);
+    const char month_names[] = "JanFebMarAprMayJunJulAugSepOctNovDec";
+    const char* p = strstr(month_names, month_str);
+    month = p ? (int)(1 + (p - month_names) / 3) : 0;
+    ImFormatString(out_buf, out_buf_size, "%04d-%02d-%02d", year, month, day);
+}
+
+// Those strings are used to output easily identifiable markers in compare logs. We only need to support what we use for testing.
+// We can probably grab info in eaplatform.h/eacompiler.h etc. in EASTL
+const ImBuildInfo* ImBuildGetCompilationInfo()
+{
+    static ImBuildInfo build_info;
+
+    if (build_info.Type[0] == '\0')
+    {
+        // Build Type
+#if defined(DEBUG) || defined(_DEBUG)
+        build_info.Type = "Debug";
+#else
+        build_info.Type = "Release";
+#endif
+
+        // CPU
+#if defined(_M_X86) || defined(_M_IX86) || defined(__i386) || defined(__i386__) || defined(_X86_) || defined(_M_AMD64) || defined(_AMD64_) || defined(__x86_64__)
+        build_info.Cpu = (sizeof(size_t) == 4) ? "X86" : "X64";
+#elif defined(__aarch64__)
+        build_info.Cpu = "ARM64";
+#else
+#error
+        build_info.Cpu = (sizeof(size_t) == 4) ? "Unknown32" : "Unknown64";
+#endif
+
+        // Platform/OS
+#if defined(_WIN32)
+        build_info.OS = "Windows";
+#elif defined(__linux) || defined(__linux__)
+        build_info.OS = "Linux";
+#elif defined(__MACH__) || defined(__MSL__)
+        build_info.OS = "OSX";
+#elif defined(__ORBIS__)
+        build_info.OS = "PS4";
+#elif defined(_DURANGO)
+        build_info.OS = "XboxOne";
+#else
+        build_info.OS = "Unknown";
+#endif
+
+        // Compiler
+#if defined(_MSC_VER)
+        build_info.Compiler = "MSVC";
+#elif defined(__clang__)
+        build_info.Compiler = "Clang";
+#elif defined(__GNUC__)
+        build_info.Compiler = "GCC";
+#else
+        build_info.Compiler = "Unknown";
+#endif
+
+        // Date/Time
+        ImBuildParseDateFromCompilerIntoYMD(__DATE__, build_info.Date, IM_ARRAYSIZE(build_info.Date));
+        build_info.Time = __TIME__;
+    }
+
+    return &build_info;
+}
+
+bool ImBuildFindGitBranchName(const char* git_repo_path, Str* branch_name)
+{
+    IM_ASSERT(git_repo_path != NULL);
+    IM_ASSERT(branch_name != NULL);
+    Str256f head_path("%s/.git/HEAD", git_repo_path);
+    size_t head_size = 0;
+    bool result = false;
+    if (char* git_head = (char*)ImFileLoadToMemory(head_path.c_str(), "r", &head_size, 1))
+    {
+        const char prefix[] = "ref: refs/heads/";       // Branch name is prefixed with this in HEAD file.
+        const int prefix_length = IM_ARRAYSIZE(prefix) - 1;
+        strtok(git_head, "\r\n");                       // Trim new line
+        if (head_size > prefix_length && strncmp(git_head, prefix, prefix_length) == 0)
+        {
+            // "ref: refs/heads/master" -> "master"
+            branch_name->set(git_head + prefix_length);
+        }
+        else
+        {
+            // Should be git hash, keep first 8 characters (see #42)
+            branch_name->setf("%.8s", git_head);
+        }
+        result = true;
+        IM_FREE(git_head);
+    }
+    return result;
+}
+
+//-----------------------------------------------------------------------------
+// Operating System Helpers
+//-----------------------------------------------------------------------------
+// - ImOsCreateProcess()
+// - ImOsPOpen()
+// - ImOsPClose()
+// - ImOsOpenInShell()
+// - ImOsConsoleSetTextColor()
+// - ImOsIsDebuggerPresent()
+//-----------------------------------------------------------------------------
+
+bool    ImOsCreateProcess(const char* cmd_line)
+{
+#ifdef _WIN32
+    STARTUPINFOA siStartInfo;
+    PROCESS_INFORMATION piProcInfo;
+    ZeroMemory(&siStartInfo, sizeof(STARTUPINFOA));
+    char* cmd_line_copy = ImStrdup(cmd_line);
+    BOOL ret = ::CreateProcessA(NULL, cmd_line_copy, NULL, NULL, FALSE, 0, NULL, NULL, &siStartInfo, &piProcInfo);
+    free(cmd_line_copy);
+    ::CloseHandle(siStartInfo.hStdInput);
+    ::CloseHandle(siStartInfo.hStdOutput);
+    ::CloseHandle(siStartInfo.hStdError);
+    ::CloseHandle(piProcInfo.hProcess);
+    ::CloseHandle(piProcInfo.hThread);
+    return ret != 0;
+#else
+    IM_UNUSED(cmd_line);
+    return false;
+#endif
+}
+
+FILE*       ImOsPOpen(const char* cmd_line, const char* mode)
+{
+    IM_ASSERT(cmd_line != NULL && *cmd_line);
+    IM_ASSERT(mode != NULL && *mode);
+#if _WIN32
+    ImVector<wchar_t> w_cmd_line;
+    ImVector<wchar_t> w_mode;
+    ImUtf8ToWideChar(cmd_line, &w_cmd_line);
+    ImUtf8ToWideChar(mode, &w_mode);
+    w_mode.resize(w_mode.Size + 1);
+    wcscat(w_mode.Data, L"b");   // Windows requires 'b' mode while unixes do not support it and default to binary.
+    return _wpopen(w_cmd_line.Data, w_mode.Data);
+#else
+    return popen(cmd_line, mode);
+#endif
+}
+
+void        ImOsPClose(FILE* fp)
+{
+    IM_ASSERT(fp != NULL);
+#if _WIN32
+    _pclose(fp);
+#else
+    pclose(fp);
+#endif
+}
+
+void    ImOsOpenInShell(const char* path)
+{
+    Str256 command(path);
+#ifdef _WIN32
+    ImPathFixSeparatorsForCurrentOS(command.c_str());
+    ::ShellExecuteA(NULL, "open", command.c_str(), NULL, NULL, SW_SHOWDEFAULT);
+#else
+#if __APPLE__
+    const char* open_executable = "open";
+#else
+    const char* open_executable = "xdg-open";
+#endif
+    command.setf("%s \"%s\"", open_executable, path);
+    ImPathFixSeparatorsForCurrentOS(command.c_str());
+    system(command.c_str());
+#endif
+}
+
+void    ImOsConsoleSetTextColor(ImOsConsoleStream stream, ImOsConsoleTextColor color)
+{
+#ifdef _WIN32
+    HANDLE hConsole = 0;
+    switch (stream)
+    {
+    case ImOsConsoleStream_StandardOutput: hConsole = ::GetStdHandle(STD_OUTPUT_HANDLE); break;
+    case ImOsConsoleStream_StandardError:  hConsole = ::GetStdHandle(STD_ERROR_HANDLE);  break;
+    }
+    WORD wAttributes = FOREGROUND_RED | FOREGROUND_GREEN | FOREGROUND_BLUE;
+    switch (color)
+    {
+    case ImOsConsoleTextColor_Black:        wAttributes = 0x00; break;
+    case ImOsConsoleTextColor_White:        wAttributes = FOREGROUND_RED | FOREGROUND_GREEN | FOREGROUND_BLUE; break;
+    case ImOsConsoleTextColor_BrightWhite:  wAttributes = FOREGROUND_RED | FOREGROUND_GREEN | FOREGROUND_BLUE | FOREGROUND_INTENSITY; break;
+    case ImOsConsoleTextColor_BrightRed:    wAttributes = FOREGROUND_RED | FOREGROUND_INTENSITY; break;
+    case ImOsConsoleTextColor_BrightGreen:  wAttributes = FOREGROUND_GREEN | FOREGROUND_INTENSITY; break;
+    case ImOsConsoleTextColor_BrightBlue:   wAttributes = FOREGROUND_BLUE | FOREGROUND_INTENSITY; break;
+    case ImOsConsoleTextColor_BrightYellow: wAttributes = FOREGROUND_RED | FOREGROUND_GREEN | FOREGROUND_INTENSITY; break;
+    default: IM_ASSERT(0);
+    }
+    ::SetConsoleTextAttribute(hConsole, wAttributes);
+#elif defined(__linux) || defined(__linux__) || defined(__MACH__) || defined(__MSL__)
+    // FIXME: check system capabilities (with environment variable TERM)
+    FILE* handle = 0;
+    switch (stream)
+    {
+    case ImOsConsoleStream_StandardOutput: handle = stdout; break;
+    case ImOsConsoleStream_StandardError:  handle = stderr; break;
+    }
+
+    const char* modifier = "";
+    switch (color)
+    {
+    case ImOsConsoleTextColor_Black:        modifier = "\033[30m";   break;
+    case ImOsConsoleTextColor_White:        modifier = "\033[0m";    break;
+    case ImOsConsoleTextColor_BrightWhite:  modifier = "\033[1;37m"; break;
+    case ImOsConsoleTextColor_BrightRed:    modifier = "\033[1;31m"; break;
+    case ImOsConsoleTextColor_BrightGreen:  modifier = "\033[1;32m"; break;
+    case ImOsConsoleTextColor_BrightBlue:   modifier = "\033[1;34m"; break;
+    case ImOsConsoleTextColor_BrightYellow: modifier = "\033[1;33m"; break;
+    default: IM_ASSERT(0);
+    }
+
+    fprintf(handle, "%s", modifier);
+#endif
+}
+
+bool    ImOsIsDebuggerPresent()
+{
+#ifdef _WIN32
+    return ::IsDebuggerPresent() != 0;
+#elif defined(__linux__)
+    int debugger_pid = 0;
+    char buf[2048];                                 // TracerPid is located near the start of the file. If end of the buffer gets cut off thats fine.
+    FILE* fp = fopen("/proc/self/status", "rb");    // Can not use ImFileLoadToMemory because size detection of /proc/self/status would fail.
+    if (fp == NULL)
+        return false;
+    fread(buf, 1, IM_ARRAYSIZE(buf), fp);
+    fclose(fp);
+    buf[IM_ARRAYSIZE(buf) - 1] = 0;
+    if (char* tracer_pid = strstr(buf, "TracerPid:"))
+    {
+        tracer_pid += 10;   // Skip label
+        while (isspace(*tracer_pid))
+            tracer_pid++;
+        debugger_pid = atoi(tracer_pid);
+    }
+    return debugger_pid != 0;
+#else
+    // FIXME
+    return false;
+#endif
+}
+
+void    ImOsOutputDebugString(const char* message)
+{
+#ifdef _WIN32
+    OutputDebugStringA(message);
+#else
+    IM_UNUSED(message);
+#endif
+}
+
+//-----------------------------------------------------------------------------
+// Str.h + InputText bindings
+//-----------------------------------------------------------------------------
+
+struct InputTextCallbackStr_UserData
+{
+    Str*                    StrObj;
+    ImGuiInputTextCallback  ChainCallback;
+    void*                   ChainCallbackUserData;
+};
+
+static int InputTextCallbackStr(ImGuiInputTextCallbackData* data)
+{
+    InputTextCallbackStr_UserData* user_data = (InputTextCallbackStr_UserData*)data->UserData;
+    if (data->EventFlag == ImGuiInputTextFlags_CallbackResize)
+    {
+        // Resize string callback
+        // If for some reason we refuse the new length (BufTextLen) and/or capacity (BufSize) we need to set them back to what we want.
+        Str* str = user_data->StrObj;
+        IM_ASSERT(data->Buf == str->c_str());
+        str->reserve(data->BufTextLen + 1);
+        data->Buf = (char*)str->c_str();
+    }
+    else if (user_data->ChainCallback)
+    {
+        // Forward to user callback, if any
+        data->UserData = user_data->ChainCallbackUserData;
+        return user_data->ChainCallback(data);
+    }
+    return 0;
+}
+
+// Draw an extra colored frame over the previous item
+// Similar to DebugDrawItemRect() but use Max(1.0f, FrameBorderSize)
+void ImGui::ItemErrorFrame(ImU32 col)
+{
+    ImGuiContext& g = *GetCurrentContext();
+    ImDrawList* drawlist = GetWindowDrawList();
+    ImGuiStyle& style = GetStyle();
+    // FIXME: GetItemRectMin() / GetItemRectMax() will include label. NavRect is not probably defined :(
+    drawlist->AddRect(g.LastItemData.NavRect.Min, g.LastItemData.NavRect.Max, GetColorU32(col), style.FrameRounding, ImDrawFlags_None, ImMax(1.0f, style.FrameBorderSize));
+}
+
+bool ImGui::InputText(const char* label, Str* str, ImGuiInputTextFlags flags, ImGuiInputTextCallback callback, void* user_data)
+{
+    IM_ASSERT((flags & ImGuiInputTextFlags_CallbackResize) == 0);
+    flags |= ImGuiInputTextFlags_CallbackResize;
+
+    InputTextCallbackStr_UserData cb_user_data;
+    cb_user_data.StrObj = str;
+    cb_user_data.ChainCallback = callback;
+    cb_user_data.ChainCallbackUserData = user_data;
+    return InputText(label, (char*)str->c_str(), (size_t)str->capacity() + 1, flags, InputTextCallbackStr, &cb_user_data);
+}
+
+bool ImGui::InputTextWithHint(const char* label, const char* hint, Str* str, ImGuiInputTextFlags flags, ImGuiInputTextCallback callback, void* user_data)
+{
+    IM_ASSERT((flags & ImGuiInputTextFlags_CallbackResize) == 0);
+    flags |= ImGuiInputTextFlags_CallbackResize;
+
+    InputTextCallbackStr_UserData cb_user_data;
+    cb_user_data.StrObj = str;
+    cb_user_data.ChainCallback = callback;
+    cb_user_data.ChainCallbackUserData = user_data;
+    return InputTextWithHint(label, hint, (char*)str->c_str(), (size_t)str->capacity() + 1, flags, InputTextCallbackStr, &cb_user_data);
+}
+
+bool ImGui::InputTextMultiline(const char* label, Str* str, const ImVec2& size, ImGuiInputTextFlags flags, ImGuiInputTextCallback callback, void* user_data)
+{
+    IM_ASSERT((flags & ImGuiInputTextFlags_CallbackResize) == 0);
+    flags |= ImGuiInputTextFlags_CallbackResize;
+
+    InputTextCallbackStr_UserData cb_user_data;
+    cb_user_data.StrObj = str;
+    cb_user_data.ChainCallback = callback;
+    cb_user_data.ChainCallbackUserData = user_data;
+    return InputTextMultiline(label, (char*)str->c_str(), (size_t)str->capacity() + 1, size, flags, InputTextCallbackStr, &cb_user_data);
+}
+
+// anchor parameter indicates which split would retain it's constant size.
+// anchor = 0 - both splits resize when parent container size changes. Both value_1 and value_2 should be persistent.
+// anchor = -1 - top/left split would have a constant size. bottom/right split would resize when parent container size changes. value_1 should be persistent, value_2 will always be recalculated from value_1.
+// anchor = +1 - bottom/right split would have a constant size. top/left split would resize when parent container size changes. value_2 should be persistent, value_1 will always be recalculated from value_2.
+bool ImGui::Splitter(const char* id, float* value_1, float* value_2, int axis, int anchor, float min_size_0, float min_size_1)
+{
+    // FIXME-DOGFOODING: This needs further refining.
+    // FIXME-SCROLL: When resizing either we'd like to keep scroll focus on something (e.g. last clicked item for list, bottom for log)
+    // See https://github.com/ocornut/imgui/issues/319
+    ImGuiContext& g = *GImGui;
+    ImGuiStyle& style = g.Style;
+    ImGuiWindow* window = ImGui::GetCurrentWindow();
+    if (min_size_0 < 0)
+        min_size_0 = ImGui::GetFrameHeight();
+    if (min_size_1)
+        min_size_1 = ImGui::GetFrameHeight();
+
+    IM_ASSERT(axis == ImGuiAxis_X || axis == ImGuiAxis_Y);
+
+    float& v_1 = *value_1;
+    float& v_2 = *value_2;
+    ImRect splitter_bb;
+    const float avail = axis == ImGuiAxis_X ? ImGui::GetContentRegionAvail().x - style.ItemSpacing.x : ImGui::GetContentRegionAvail().y - style.ItemSpacing.y;
+    if (anchor < 0)
+    {
+        v_2 = ImMax(avail - v_1, min_size_1);   // First split is constant size.
+    }
+    else if (anchor > 0)
+    {
+        v_1 = ImMax(avail - v_2, min_size_0);   // Second split is constant size.
+    }
+    else
+    {
+        float r = v_1 / (v_1 + v_2);            // Both splits maintain same relative size to parent.
+        v_1 = IM_ROUND(avail * r) - 1;
+        v_2 = IM_ROUND(avail * (1.0f - r)) - 1;
+    }
+    if (axis == ImGuiAxis_X)
+    {
+        float x = window->DC.CursorPos.x + v_1 + IM_ROUND(style.ItemSpacing.x * 0.5f);
+        splitter_bb = ImRect(x - 1, window->WorkRect.Min.y, x + 1, window->WorkRect.Max.y);
+    }
+    else if (axis == ImGuiAxis_Y)
+    {
+        float y = window->DC.CursorPos.y + v_1 + IM_ROUND(style.ItemSpacing.y * 0.5f);
+        splitter_bb = ImRect(window->WorkRect.Min.x, y - 1, window->WorkRect.Max.x, y + 1);
+    }
+    return ImGui::SplitterBehavior(splitter_bb, ImGui::GetID(id), (ImGuiAxis)axis, &v_1, &v_2, min_size_0, min_size_1, 3.0f);
+}
+
+// FIXME-TESTS: Should eventually remove.
+ImFont* ImGui::FindFontByPrefix(const char* prefix)
+{
+    ImGuiContext& g = *GImGui;
+    for (ImFont* font : g.IO.Fonts->Fonts)
+        if (strncmp(font->ConfigData->Name, prefix, strlen(prefix)) == 0)
+            return font;
+    return NULL;
+}
+
+// Legacy version support
+#if IMGUI_VERSION_NUM < 18924
+const char* ImGui::TabBarGetTabName(ImGuiTabBar* tab_bar, ImGuiTabItem* tab)
+{
+    return tab_bar->GetTabName(tab);
+}
+#endif
+
+#if IMGUI_VERSION_NUM < 18927
+ImGuiID ImGui::TableGetInstanceID(ImGuiTable* table, int instance_no)
+{
+    // Changed in #6140
+    return table->ID + instance_no;
+}
+#endif
+
+ImGuiID TableGetHeaderID(ImGuiTable* table, const char* column, int instance_no)
+{
+    IM_ASSERT(table != NULL);
+    int column_n = -1;
+    for (int n = 0; n < table->Columns.size() && column_n < 0; n++)
+        if (strcmp(ImGui::TableGetColumnName(table, n), column) == 0)
+            column_n = n;
+    IM_ASSERT(column_n != -1);
+    return TableGetHeaderID(table, column_n, instance_no);
+}
+
+ImGuiID TableGetHeaderID(ImGuiTable* table, int column_n, int instance_no)
+{
+    IM_ASSERT(column_n >= 0 && column_n < table->ColumnsCount);
+    const ImGuiID table_instance_id = ImGui::TableGetInstanceID(table, instance_no);
+    const char* column_name = ImGui::TableGetColumnName(table, column_n);
+#if IMGUI_VERSION_NUM >= 18927
+    const int column_id_differencier = column_n;
+#else
+    const int column_id_differencier = instance_no * table->ColumnsCount + column_n;
+#endif
+    const int column_id = ImHashData(&column_id_differencier, sizeof(column_id_differencier), table_instance_id);
+    return ImHashData(column_name, strlen(column_name), column_id);
+}
+
+// FIXME: Could be moved to core as an internal function?
+void TableDiscardInstanceAndSettings(ImGuiID table_id)
+{
+    ImGuiContext& g = *GImGui;
+    IM_ASSERT(g.CurrentTable == NULL);
+    if (ImGuiTableSettings* settings = ImGui::TableSettingsFindByID(table_id))
+        settings->ID = 0;
+
+    if (ImGuiTable* table = ImGui::TableFindByID(table_id))
+        ImGui::TableRemove(table);
+    // FIXME-TABLE: We should be able to use TableResetSettings() instead of TableRemove()! Maybe less of a clean slate but would be good to check that it does the job
+    //ImGui::TableResetSettings(table);
+}
+
+// Helper to verify ImDrawData integrity of buffer count (broke before e.g. #6716)
+void DrawDataVerifyMatchingBufferCount(ImDrawData* draw_data)
+{
+    int total_vtx_count = 0;
+    int total_idx_count = 0;
+    for (ImDrawList* draw_list : draw_data->CmdLists)
+    {
+        total_vtx_count += draw_list->VtxBuffer.Size;
+        total_idx_count += draw_list->IdxBuffer.Size;
+    }
+    IM_UNUSED(total_vtx_count);
+    IM_UNUSED(total_idx_count);
+    IM_ASSERT(total_vtx_count == draw_data->TotalVtxCount);
+    IM_ASSERT(total_idx_count == draw_data->TotalIdxCount);
+}
+
+//-----------------------------------------------------------------------------
+// Simple CSV parser
+//-----------------------------------------------------------------------------
+
+void ImGuiCsvParser::Clear()
+{
+    Rows = Columns = 0;
+    if (_Data != NULL)
+        IM_FREE(_Data);
+    _Data = NULL;
+    _Index.clear();
+}
+
+bool ImGuiCsvParser::Load(const char* filename)
+{
+    size_t len = 0;
+    _Data = (char*)ImFileLoadToMemory(filename, "rb", &len, 1);
+    if (_Data == NULL)
+        return false;
+
+    int columns = 1;
+    if (Columns > 0)
+    {
+        columns = Columns;                                          // User-provided expected column count.
+    }
+    else
+    {
+        for (const char* c = _Data; *c != '\n' && *c != '\0'; c++)  // Count columns. Quoted columns with commas are not supported.
+            if (*c == ',')
+                columns++;
+    }
+
+    // Count rows. Extra new lines anywhere in the file are ignored.
+    int max_rows = 0;
+    for (const char* c = _Data, *end = c + len; c < end; c++)
+        if ((*c == '\n' && c[1] != '\r' && c[1] != '\n') || *c == '\0')
+            max_rows++;
+
+    if (columns == 0 || max_rows == 0)
+        return false;
+
+    // Create index
+    _Index.resize(columns * max_rows);
+
+    int col = 0;
+    char* col_data = _Data;
+    for (char* c = _Data; *c != '\0'; c++)
+    {
+        const bool is_comma = (*c == ',');
+        const bool is_eol = (*c == '\n' || *c == '\r');
+        const bool is_eof = (*c == '\0');
+        if (is_comma || is_eol || is_eof)
+        {
+            _Index[Rows * columns + col] = col_data;
+            col_data = c + 1;
+            if (is_comma)
+            {
+                col++;
+            }
+            else
+            {
+                if (col + 1 == columns)
+                    Rows++;
+                else
+                    fprintf(stderr, "%s: Unexpected number of columns on line %d, ignoring.\n", filename, Rows + 1); // FIXME
+                col = 0;
+            }
+            *c = 0;
+            if (is_eol)
+                while (c[1] == '\r' || c[1] == '\n')
+                    c++;
+        }
+    }
+
+    Columns = columns;
+    return true;
+}
+
+//-----------------------------------------------------------------------------