blender/intern/cycles/util/util_path.cpp
Sergey Sharybin 0d01cf4488 Cycles: Extra tweaks to performance of header expansion
Two main things here:

1. Replace all unsafe for #line directive characters into a single loop,
   avoiding multiple iterations and multiple temporary strings created.

2. Don't merge token char by char but calculate start and end point and
   then copy all substring at once.

This gives about 15% speedup of source processing time. At this point
(with all previous commits from today) we've shrinked down compiled
sources size from 108 MB down to ~5.5 MB and lowered processing time
from 4.5 sec down to 0.047 sec on my laptop running Linux (this was a
constant time which Blender will always spent first time loading kernel,
even if we've got compiled clbin).
2017-08-03 08:07:06 +02:00

992 lines
22 KiB
C++

/*
* Copyright 2011-2013 Blender Foundation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "util/util_debug.h"
#include "util/util_md5.h"
#include "util/util_path.h"
#include "util/util_string.h"
#include <OpenImageIO/filesystem.h>
#include <OpenImageIO/strutil.h>
#include <OpenImageIO/sysutil.h>
OIIO_NAMESPACE_USING
#include <stdio.h>
#include <sys/stat.h>
#if defined(_WIN32)
# define DIR_SEP '\\'
# define DIR_SEP_ALT '/'
# include <direct.h>
#else
# define DIR_SEP '/'
# include <dirent.h>
# include <pwd.h>
# include <unistd.h>
# include <sys/types.h>
#endif
#ifdef HAVE_SHLWAPI_H
# include <shlwapi.h>
#endif
#include "util/util_map.h"
#include "util/util_windows.h"
CCL_NAMESPACE_BEGIN
#ifdef _WIN32
# if defined(_MSC_VER) || defined(__MINGW64__)
typedef struct _stat64 path_stat_t;
# elif defined(__MINGW32__)
typedef struct _stati64 path_stat_t;
# else
typedef struct _stat path_stat_t;
# endif
# ifndef S_ISDIR
# define S_ISDIR(x) (((x) & _S_IFDIR) == _S_IFDIR)
# endif
#else
typedef struct stat path_stat_t;
#endif
static string cached_path = "";
static string cached_user_path = "";
static string cached_xdg_cache_path = "";
namespace {
#ifdef _WIN32
class directory_iterator {
public:
class path_info {
public:
path_info(const string& path,
const WIN32_FIND_DATAW& find_data)
: path_(path),
find_data_(find_data)
{
}
string path() {
return path_join(path_, string_from_wstring(find_data_.cFileName));
}
protected:
const string& path_;
const WIN32_FIND_DATAW& find_data_;
};
directory_iterator()
: path_info_("", find_data_),
h_find_(INVALID_HANDLE_VALUE)
{
}
explicit directory_iterator(const string& path)
: path_(path),
path_info_(path, find_data_)
{
string wildcard = path;
if(wildcard[wildcard.size() - 1] != DIR_SEP) {
wildcard += DIR_SEP;
}
wildcard += "*";
h_find_ = FindFirstFileW(string_to_wstring(wildcard).c_str(),
&find_data_);
if(h_find_ != INVALID_HANDLE_VALUE) {
skip_dots();
}
}
~directory_iterator()
{
if(h_find_ != INVALID_HANDLE_VALUE) {
FindClose(h_find_);
}
}
directory_iterator& operator++()
{
step();
return *this;
}
path_info* operator-> ()
{
return &path_info_;
}
bool operator!=(const directory_iterator& other)
{
return h_find_ != other.h_find_;
}
protected:
bool step()
{
if(do_step()) {
return skip_dots();
}
return false;
}
bool do_step()
{
if(h_find_ != INVALID_HANDLE_VALUE) {
bool result = FindNextFileW(h_find_, &find_data_) == TRUE;
if(!result) {
FindClose(h_find_);
h_find_ = INVALID_HANDLE_VALUE;
}
return result;
}
return false;
}
bool skip_dots()
{
while(wcscmp(find_data_.cFileName, L".") == 0 ||
wcscmp(find_data_.cFileName, L"..") == 0)
{
if(!do_step()) {
return false;
}
}
return true;
}
string path_;
path_info path_info_;
WIN32_FIND_DATAW find_data_;
HANDLE h_find_;
};
#else /* _WIN32 */
class directory_iterator {
public:
class path_info {
public:
explicit path_info(const string& path)
: path_(path),
entry_(NULL)
{
}
string path() {
return path_join(path_, entry_->d_name);
}
void current_entry_set(const struct dirent *entry)
{
entry_ = entry;
}
protected:
const string& path_;
const struct dirent *entry_;
};
directory_iterator()
: path_info_(""),
name_list_(NULL),
num_entries_(-1),
cur_entry_(-1)
{
}
explicit directory_iterator(const string& path)
: path_(path),
path_info_(path_),
cur_entry_(0)
{
num_entries_ = scandir(path.c_str(),
&name_list_,
NULL,
alphasort);
if(num_entries_ < 0) {
perror("scandir");
}
else {
skip_dots();
}
}
~directory_iterator()
{
destroy_name_list();
}
directory_iterator& operator++()
{
step();
return *this;
}
path_info* operator-> ()
{
path_info_.current_entry_set(name_list_[cur_entry_]);
return &path_info_;
}
bool operator!=(const directory_iterator& other)
{
return name_list_ != other.name_list_;
}
protected:
bool step()
{
if(do_step()) {
return skip_dots();
}
return false;
}
bool do_step()
{
++cur_entry_;
if(cur_entry_ >= num_entries_) {
destroy_name_list();
return false;
}
return true;
}
/* Skip . and .. folders. */
bool skip_dots()
{
while(strcmp(name_list_[cur_entry_]->d_name, ".") == 0 ||
strcmp(name_list_[cur_entry_]->d_name, "..") == 0)
{
if(!step()) {
return false;
}
}
return true;
}
void destroy_name_list()
{
if(name_list_ == NULL) {
return;
}
for(int i = 0; i < num_entries_; ++i) {
free(name_list_[i]);
}
free(name_list_);
name_list_ = NULL;
}
string path_;
path_info path_info_;
struct dirent **name_list_;
int num_entries_, cur_entry_;
};
#endif /* _WIN32 */
size_t find_last_slash(const string& path)
{
for(size_t i = 0; i < path.size(); ++i) {
size_t index = path.size() - 1 - i;
#ifdef _WIN32
if(path[index] == DIR_SEP || path[index] == DIR_SEP_ALT)
#else
if(path[index] == DIR_SEP)
#endif
{
return index;
}
}
return string::npos;
}
} /* namespace */
static char *path_specials(const string& sub)
{
static bool env_init = false;
static char *env_shader_path;
static char *env_source_path;
if(!env_init) {
env_shader_path = getenv("CYCLES_SHADER_PATH");
/* NOTE: It is KERNEL in env variable for compatibility reasons. */
env_source_path = getenv("CYCLES_KERNEL_PATH");
env_init = true;
}
if(env_shader_path != NULL && sub == "shader") {
return env_shader_path;
}
else if(env_shader_path != NULL && sub == "source") {
return env_source_path;
}
return NULL;
}
#if defined(__linux__) || defined(__APPLE__)
static string path_xdg_cache_get()
{
const char *home = getenv("XDG_CACHE_HOME");
if(home) {
return string(home);
}
else {
home = getenv("HOME");
if(home == NULL) {
home = getpwuid(getuid())->pw_dir;
}
return path_join(string(home), ".cache");
}
}
#endif
void path_init(const string& path, const string& user_path)
{
cached_path = path;
cached_user_path = user_path;
#ifdef _MSC_VER
// workaround for https://svn.boost.org/trac/boost/ticket/6320
// indirectly init boost codec here since it's not thread safe, and can
// cause crashes when it happens in multithreaded image load
OIIO::Filesystem::exists(path);
#endif
}
string path_get(const string& sub)
{
char *special = path_specials(sub);
if(special != NULL)
return special;
if(cached_path == "")
cached_path = path_dirname(Sysutil::this_program_path());
return path_join(cached_path, sub);
}
string path_user_get(const string& sub)
{
if(cached_user_path == "")
cached_user_path = path_dirname(Sysutil::this_program_path());
return path_join(cached_user_path, sub);
}
string path_cache_get(const string& sub)
{
#if defined(__linux__) || defined(__APPLE__)
if(cached_xdg_cache_path == "") {
cached_xdg_cache_path = path_xdg_cache_get();
}
string result = path_join(cached_xdg_cache_path, "cycles");
return path_join(result, sub);
#else
/* TODO(sergey): What that should be on Windows? */
return path_user_get(path_join("cache", sub));
#endif
}
#if defined(__linux__) || defined(__APPLE__)
string path_xdg_home_get(const string& sub = "");
#endif
string path_filename(const string& path)
{
size_t index = find_last_slash(path);
if(index != string::npos) {
/* Corner cases to match boost behavior. */
#ifndef _WIN32
if(index == 0 && path.size() == 1) {
return path;
}
#endif
if(index == path.size() - 1) {
#ifdef _WIN32
if(index == 2) {
return string(1, DIR_SEP);
}
#endif
return ".";
}
return path.substr(index + 1, path.size() - index - 1);
}
return path;
}
string path_dirname(const string& path)
{
size_t index = find_last_slash(path);
if(index != string::npos) {
#ifndef _WIN32
if(index == 0 && path.size() > 1) {
return string(1, DIR_SEP);
}
#endif
return path.substr(0, index);
}
return "";
}
string path_join(const string& dir, const string& file)
{
if(dir.size() == 0) {
return file;
}
if(file.size() == 0) {
return dir;
}
string result = dir;
#ifndef _WIN32
if(result[result.size() - 1] != DIR_SEP &&
file[0] != DIR_SEP)
#else
if(result[result.size() - 1] != DIR_SEP &&
result[result.size() - 1] != DIR_SEP_ALT &&
file[0] != DIR_SEP &&
file[0] != DIR_SEP_ALT)
#endif
{
result += DIR_SEP;
}
result += file;
return result;
}
string path_escape(const string& path)
{
string result = path;
string_replace(result, " ", "\\ ");
return result;
}
bool path_is_relative(const string& path)
{
#ifdef _WIN32
# ifdef HAVE_SHLWAPI_H
return PathIsRelative(path.c_str());
# else /* HAVE_SHLWAPI_H */
if(path.size() >= 3) {
return !(((path[0] >= 'a' && path[0] <= 'z') ||
(path[0] >= 'A' && path[0] <= 'Z')) &&
path[1] == ':' && path[2] == DIR_SEP);
}
return true;
# endif /* HAVE_SHLWAPI_H */
#else /* _WIN32 */
if(path.size() == 0) {
return 1;
}
return path[0] != DIR_SEP;
#endif /* _WIN32 */
}
#ifdef _WIN32
/* Add a slash if the UNC path points to a share. */
static string path_unc_add_slash_to_share(const string& path)
{
size_t slash_after_server = path.find(DIR_SEP, 2);
if(slash_after_server != string::npos) {
size_t slash_after_share = path.find(DIR_SEP,
slash_after_server + 1);
if(slash_after_share == string::npos) {
return path + DIR_SEP;
}
}
return path;
}
/* Convert:
* \\?\UNC\server\share\folder\... to \\server\share\folder\...
* \\?\C:\ to C:\ and \\?\C:\folder\... to C:\folder\...
*/
static string path_unc_to_short(const string& path)
{
size_t len = path.size();
if((len > 3) &&
(path[0] == DIR_SEP) &&
(path[1] == DIR_SEP) &&
(path[2] == '?') &&
((path[3] == DIR_SEP) || (path[3] == DIR_SEP_ALT)))
{
if((len > 5) && (path[5] == ':')) {
return path.substr(4, len - 4);
}
else if((len > 7) &&
(path.substr(4, 3) == "UNC") &&
((path[7] == DIR_SEP) || (path[7] == DIR_SEP_ALT)))
{
return "\\\\" + path.substr(8, len - 8);
}
}
return path;
}
static string path_cleanup_unc(const string& path)
{
string result = path_unc_to_short(path);
if(path.size() > 2) {
/* It's possible path is now a non-UNC. */
if(result[0] == DIR_SEP && result[1] == DIR_SEP) {
return path_unc_add_slash_to_share(result);
}
}
return result;
}
/* Make path compatible for stat() functions. */
static string path_make_compatible(const string& path)
{
string result = path;
/* In Windows stat() doesn't recognize dir ending on a slash. */
if(result.size() > 3 && result[result.size() - 1] == DIR_SEP) {
result.resize(result.size() - 1);
}
/* Clean up UNC path. */
if((path.size() >= 3) && (path[0] == DIR_SEP) && (path[1] == DIR_SEP)) {
result = path_cleanup_unc(result);
}
/* Make sure volume-only path ends up wit ha directory separator. */
if(result.size() == 2 && result[1] == ':') {
result += DIR_SEP;
}
return result;
}
static int path_wstat(const wstring& path_wc, path_stat_t *st)
{
#if defined(_MSC_VER) || defined(__MINGW64__)
return _wstat64(path_wc.c_str(), st);
#elif defined(__MINGW32__)
return _wstati64(path_wc.c_str(), st);
#else
return _wstat(path_wc.c_str(), st);
#endif
}
static int path_stat(const string& path, path_stat_t *st)
{
wstring path_wc = string_to_wstring(path);
return path_wstat(path_wc, st);
}
#else /* _WIN32 */
static int path_stat(const string& path, path_stat_t *st)
{
return stat(path.c_str(), st);
}
#endif /* _WIN32 */
size_t path_file_size(const string& path)
{
path_stat_t st;
if(path_stat(path, &st) != 0) {
return -1;
}
return st.st_size;
}
bool path_exists(const string& path)
{
#ifdef _WIN32
string fixed_path = path_make_compatible(path);
wstring path_wc = string_to_wstring(fixed_path);
path_stat_t st;
if(path_wstat(path_wc, &st) != 0) {
return false;
}
return st.st_mode != 0;
#else /* _WIN32 */
struct stat st;
if(stat(path.c_str(), &st) != 0) {
return 0;
}
return st.st_mode != 0;
#endif /* _WIN32 */
}
bool path_is_directory(const string& path)
{
path_stat_t st;
if(path_stat(path, &st) != 0) {
return false;
}
return S_ISDIR(st.st_mode);
}
static void path_files_md5_hash_recursive(MD5Hash& hash, const string& dir)
{
if(path_exists(dir)) {
directory_iterator it(dir), it_end;
for(; it != it_end; ++it) {
if(path_is_directory(it->path())) {
path_files_md5_hash_recursive(hash, it->path());
}
else {
string filepath = it->path();
hash.append((const uint8_t*)filepath.c_str(), filepath.size());
hash.append_file(filepath);
}
}
}
}
string path_files_md5_hash(const string& dir)
{
/* computes md5 hash of all files in the directory */
MD5Hash hash;
path_files_md5_hash_recursive(hash, dir);
return hash.get_hex();
}
static bool create_directories_recursivey(const string& path)
{
if(path_is_directory(path)) {
/* Directory already exists, nothing to do. */
return true;
}
if(path_exists(path)) {
/* File exists and it's not a directory. */
return false;
}
string parent = path_dirname(path);
if(parent.size() > 0 && parent != path) {
if(!create_directories_recursivey(parent)) {
return false;
}
}
#ifdef _WIN32
wstring path_wc = string_to_wstring(path);
return _wmkdir(path_wc.c_str()) == 0;
#else
return mkdir(path.c_str(), 0777) == 0;
#endif
}
void path_create_directories(const string& filepath)
{
string path = path_dirname(filepath);
create_directories_recursivey(path);
}
bool path_write_binary(const string& path, const vector<uint8_t>& binary)
{
path_create_directories(path);
/* write binary file from memory */
FILE *f = path_fopen(path, "wb");
if(!f)
return false;
if(binary.size() > 0)
fwrite(&binary[0], sizeof(uint8_t), binary.size(), f);
fclose(f);
return true;
}
bool path_write_text(const string& path, string& text)
{
vector<uint8_t> binary(text.length(), 0);
std::copy(text.begin(), text.end(), binary.begin());
return path_write_binary(path, binary);
}
bool path_read_binary(const string& path, vector<uint8_t>& binary)
{
/* read binary file into memory */
FILE *f = path_fopen(path, "rb");
if(!f) {
binary.resize(0);
return false;
}
binary.resize(path_file_size(path));
if(binary.size() == 0) {
fclose(f);
return false;
}
if(fread(&binary[0], sizeof(uint8_t), binary.size(), f) != binary.size()) {
fclose(f);
return false;
}
fclose(f);
return true;
}
bool path_read_text(const string& path, string& text)
{
vector<uint8_t> binary;
if(!path_exists(path) || !path_read_binary(path, binary))
return false;
const char *str = (const char*)&binary[0];
size_t size = binary.size();
text = string(str, size);
return true;
}
uint64_t path_modified_time(const string& path)
{
path_stat_t st;
if(path_stat(path, &st) != 0) {
return 0;
}
return st.st_mtime;
}
bool path_remove(const string& path)
{
return remove(path.c_str()) == 0;
}
struct SourceReplaceState {
typedef map<string, string> ProcessedMapping;
/* Base director for all relative include headers. */
string base;
/* Result of processed files. */
ProcessedMapping processed_files;
/* Set of files which are considered "precompiled" and which are replaced
* with and empty string on a subsequent occurrence in include statement.
*/
set<string> precompiled_headers;
};
static string path_source_replace_includes_recursive(
const string& source,
const string& source_filepath,
SourceReplaceState *state);
static string line_directive(const SourceReplaceState& state,
const string& path,
const int line)
{
string unescaped_path = path;
/* First we make path relative. */
if(string_startswith(unescaped_path, state.base.c_str())) {
const string base_file = path_filename(state.base);
const size_t base_len = state.base.length();
unescaped_path = base_file +
unescaped_path.substr(base_len,
unescaped_path.length() - base_len);
}
/* Second, we replace all unsafe characters. */
const size_t length = unescaped_path.length();
string escaped_path = "";
for(size_t i = 0; i < length; ++i) {
const char ch = unescaped_path[i];
if(strchr("\"\'\?\\", ch) != NULL) {
escaped_path += "\\";
}
escaped_path += ch;
}
/* TODO(sergey): Check whether using std::to_string combined with several
* concatenation operations is any faster.
*/
return string_printf("#line %d \"%s\"", line, escaped_path.c_str());
}
static string path_source_handle_preprocessor(
const string& preprocessor_line,
const string& source_filepath,
const size_t line_number,
SourceReplaceState *state)
{
string result = preprocessor_line;
string token = string_strip(
preprocessor_line.substr(1, preprocessor_line.size() - 1));
if(string_startswith(token, "include")) {
token = string_strip(token.substr(7, token.size() - 7));
if(token[0] == '"') {
const size_t n_start = 1;
const size_t n_end = token.find("\"", n_start);
const string filename = token.substr(n_start, n_end - n_start);
const bool is_precompiled = string_endswith(token, "// PRECOMPILED");
string filepath = path_join(state->base, filename);
if(!path_exists(filepath)) {
filepath = path_join(path_dirname(source_filepath),
filename);
}
if(is_precompiled) {
state->precompiled_headers.insert(filepath);
}
string text;
if(path_read_text(filepath, text)) {
text = path_source_replace_includes_recursive(
text, filepath, state);
/* Use line directives for better error messages. */
result = line_directive(*state, filepath, 1) + "\n"
+ text + "\n"
+ line_directive(*state, source_filepath, line_number + 1);
}
}
}
return result;
}
/* Our own little c preprocessor that replaces #includes with the file
* contents, to work around issue of OpenCL drivers not supporting
* include paths with spaces in them.
*/
static string path_source_replace_includes_recursive(
const string& source,
const string& source_filepath,
SourceReplaceState *state)
{
/* Try to re-use processed file without spending time on replacing all
* include directives again.
*/
SourceReplaceState::ProcessedMapping::iterator replaced_file =
state->processed_files.find(source_filepath);
if(replaced_file != state->processed_files.end()) {
if(state->precompiled_headers.find(source_filepath) !=
state->precompiled_headers.end()) {
return "";
}
return replaced_file->second;
}
/* Perform full file processing. */
string result = "";
const size_t source_length = source.length();
size_t index = 0;
/* Information about where we are in the source. */
size_t line_number = 0, column_number = 1;
/* Currently gathered non-preprocessor token.
* Store as start/length rather than token itself to avoid overhead of
* memory re-allocations on each character concatenation.
*/
size_t token_start = 0, token_length = 0;
/* Denotes whether we're inside of preprocessor line, together with
* preprocessor line itself.
*
* TODO(sergey): Investigate whether using token start/end position
* gives measurable speedup.
*/
bool inside_preprocessor = false;
string preprocessor_line = "";
/* Actual loop over the whole source. */
while(index < source_length) {
const char ch = source[index];
if(ch == '\n') {
if(inside_preprocessor) {
result += path_source_handle_preprocessor(preprocessor_line,
source_filepath,
line_number,
state);
/* Start gathering net part of the token. */
token_start = index;
token_length = 0;
}
inside_preprocessor = false;
preprocessor_line = "";
column_number = 0;
++line_number;
}
else if(ch == '#' && column_number == 1 && !inside_preprocessor) {
/* Append all possible non-preprocessor token to the result. */
if(token_length != 0) {
result.append(source, token_start, token_length);
token_start = index;
token_length = 0;
}
inside_preprocessor = true;
}
if(inside_preprocessor) {
preprocessor_line += ch;
}
else {
++token_length;
}
++index;
++column_number;
}
/* Append possible tokens which happened before special events handled
* above.
*/
if(token_length != 0) {
result.append(source, token_start, token_length);
}
if(inside_preprocessor) {
result += path_source_handle_preprocessor(preprocessor_line,
source_filepath,
line_number,
state);
}
/* Store result for further reuse. */
state->processed_files[source_filepath] = result;
return result;
}
string path_source_replace_includes(const string& source,
const string& path,
const string& source_filename)
{
SourceReplaceState state;
state.base = path;
return path_source_replace_includes_recursive(
source,
path_join(path, source_filename),
&state);
}
FILE *path_fopen(const string& path, const string& mode)
{
#ifdef _WIN32
wstring path_wc = string_to_wstring(path);
wstring mode_wc = string_to_wstring(mode);
return _wfopen(path_wc.c_str(), mode_wc.c_str());
#else
return fopen(path.c_str(), mode.c_str());
#endif
}
void path_cache_clear_except(const string& name, const set<string>& except)
{
string dir = path_user_get("cache");
if(path_exists(dir)) {
directory_iterator it(dir), it_end;
for(; it != it_end; ++it) {
string filename = path_filename(it->path());
if(string_startswith(filename, name.c_str()))
if(except.find(filename) == except.end())
path_remove(it->path());
}
}
}
CCL_NAMESPACE_END