2024-05-12 14:09:43 -07:00
|
|
|
// Copyright (c) FIRST and other WPILib contributors.
|
|
|
|
|
// Open Source Software; you can modify and/or share it under the terms of
|
|
|
|
|
// the WPILib BSD license file in the root directory of this project.
|
|
|
|
|
|
2025-11-07 19:56:21 -05:00
|
|
|
#include "wpi/datalog/DataLogBackgroundWriter.hpp"
|
2024-05-12 14:09:43 -07:00
|
|
|
|
|
|
|
|
#ifndef _WIN32
|
|
|
|
|
#include <unistd.h>
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#ifdef _WIN32
|
|
|
|
|
#ifndef WIN32_LEAN_AND_MEAN
|
|
|
|
|
#define WIN32_LEAN_AND_MEAN
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#include <windows.h> // NOLINT(build/include_order)
|
|
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#include <random>
|
2024-09-20 17:43:39 -07:00
|
|
|
#include <string>
|
|
|
|
|
#include <utility>
|
|
|
|
|
#include <vector>
|
2024-05-12 14:09:43 -07:00
|
|
|
|
|
|
|
|
#include <fmt/format.h>
|
|
|
|
|
|
2025-11-07 19:56:21 -05:00
|
|
|
#include "wpi/util/Logger.hpp"
|
|
|
|
|
#include "wpi/util/fs.hpp"
|
2024-05-12 14:09:43 -07:00
|
|
|
|
|
|
|
|
using namespace wpi::log;
|
|
|
|
|
|
|
|
|
|
static constexpr uintmax_t kMinFreeSpace = 5 * 1024 * 1024;
|
|
|
|
|
|
|
|
|
|
static std::string FormatBytesSize(uintmax_t value) {
|
|
|
|
|
static constexpr uintmax_t kKiB = 1024;
|
|
|
|
|
static constexpr uintmax_t kMiB = kKiB * 1024;
|
|
|
|
|
static constexpr uintmax_t kGiB = kMiB * 1024;
|
|
|
|
|
if (value >= kGiB) {
|
|
|
|
|
return fmt::format("{:.1f} GiB", static_cast<double>(value) / kGiB);
|
|
|
|
|
} else if (value >= kMiB) {
|
|
|
|
|
return fmt::format("{:.1f} MiB", static_cast<double>(value) / kMiB);
|
|
|
|
|
} else if (value >= kKiB) {
|
|
|
|
|
return fmt::format("{:.1f} KiB", static_cast<double>(value) / kKiB);
|
|
|
|
|
} else {
|
|
|
|
|
return fmt::format("{} B", value);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
DataLogBackgroundWriter::DataLogBackgroundWriter(std::string_view dir,
|
|
|
|
|
std::string_view filename,
|
|
|
|
|
double period,
|
|
|
|
|
std::string_view extraHeader)
|
|
|
|
|
: DataLogBackgroundWriter{s_defaultMessageLog, dir, filename, period,
|
|
|
|
|
extraHeader} {}
|
|
|
|
|
|
2025-11-07 20:00:05 -05:00
|
|
|
DataLogBackgroundWriter::DataLogBackgroundWriter(wpi::util::Logger& msglog,
|
2024-05-12 14:09:43 -07:00
|
|
|
std::string_view dir,
|
|
|
|
|
std::string_view filename,
|
|
|
|
|
double period,
|
|
|
|
|
std::string_view extraHeader)
|
|
|
|
|
: DataLog{msglog, extraHeader},
|
|
|
|
|
m_period{period},
|
|
|
|
|
m_newFilename{filename},
|
|
|
|
|
m_thread{[this, dir = std::string{dir}] { WriterThreadMain(dir); }} {}
|
|
|
|
|
|
|
|
|
|
DataLogBackgroundWriter::DataLogBackgroundWriter(
|
|
|
|
|
std::function<void(std::span<const uint8_t> data)> write, double period,
|
|
|
|
|
std::string_view extraHeader)
|
|
|
|
|
: DataLogBackgroundWriter{s_defaultMessageLog, std::move(write), period,
|
|
|
|
|
extraHeader} {}
|
|
|
|
|
|
|
|
|
|
DataLogBackgroundWriter::DataLogBackgroundWriter(
|
2025-11-07 20:00:05 -05:00
|
|
|
wpi::util::Logger& msglog,
|
2024-05-12 14:09:43 -07:00
|
|
|
std::function<void(std::span<const uint8_t> data)> write, double period,
|
|
|
|
|
std::string_view extraHeader)
|
|
|
|
|
: DataLog{msglog, extraHeader},
|
|
|
|
|
m_period{period},
|
|
|
|
|
m_thread{[this, write = std::move(write)] {
|
|
|
|
|
WriterThreadMain(std::move(write));
|
|
|
|
|
}} {}
|
|
|
|
|
|
|
|
|
|
DataLogBackgroundWriter::~DataLogBackgroundWriter() {
|
|
|
|
|
{
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
m_shutdown = true;
|
|
|
|
|
m_doFlush = true;
|
|
|
|
|
}
|
|
|
|
|
m_cond.notify_all();
|
|
|
|
|
m_thread.join();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::SetFilename(std::string_view filename) {
|
|
|
|
|
{
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
m_newFilename = filename;
|
|
|
|
|
}
|
|
|
|
|
m_cond.notify_all();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::Flush() {
|
|
|
|
|
{
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
m_doFlush = true;
|
|
|
|
|
}
|
|
|
|
|
m_cond.notify_all();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::Pause() {
|
|
|
|
|
DataLog::Pause();
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
m_state = kPaused;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::Resume() {
|
|
|
|
|
DataLog::Resume();
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
if (m_state == kPaused) {
|
|
|
|
|
m_state = kActive;
|
|
|
|
|
} else if (m_state == kStopped) {
|
|
|
|
|
m_state = kStart;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::Stop() {
|
|
|
|
|
DataLog::Stop();
|
|
|
|
|
{
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
m_state = kStopped;
|
|
|
|
|
m_newFilename.clear();
|
|
|
|
|
}
|
|
|
|
|
m_cond.notify_all();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void WriteToFile(fs::file_t f, std::span<const uint8_t> data,
|
2025-11-07 20:00:05 -05:00
|
|
|
std::string_view filename, wpi::util::Logger& msglog) {
|
2024-05-12 14:09:43 -07:00
|
|
|
do {
|
|
|
|
|
#ifdef _WIN32
|
|
|
|
|
DWORD ret;
|
|
|
|
|
if (!WriteFile(f, data.data(), data.size(), &ret, nullptr)) {
|
|
|
|
|
WPI_ERROR(msglog, "Error writing to log file '{}': {}", filename,
|
|
|
|
|
GetLastError());
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
#else
|
|
|
|
|
ssize_t ret = ::write(f, data.data(), data.size());
|
|
|
|
|
if (ret < 0) {
|
|
|
|
|
// If it's a recoverable error, swallow it and retry the write
|
|
|
|
|
if (errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Otherwise it's a non-recoverable error; quit trying
|
|
|
|
|
WPI_ERROR(msglog, "Error writing to log file '{}': {}", filename,
|
|
|
|
|
std::strerror(errno));
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
// The write may have written some or all of the data
|
|
|
|
|
data = data.subspan(ret);
|
|
|
|
|
} while (data.size() > 0);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static std::string MakeRandomFilename() {
|
|
|
|
|
// build random filename
|
|
|
|
|
static std::random_device dev;
|
|
|
|
|
static std::mt19937 rng(dev());
|
|
|
|
|
std::uniform_int_distribution<int> dist(0, 15);
|
|
|
|
|
const char* v = "0123456789abcdef";
|
|
|
|
|
std::string filename = "wpilog_";
|
|
|
|
|
for (int i = 0; i < 16; i++) {
|
|
|
|
|
filename += v[dist(rng)];
|
|
|
|
|
}
|
|
|
|
|
filename += ".wpilog";
|
|
|
|
|
return filename;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct DataLogBackgroundWriter::WriterThreadState {
|
2024-10-03 15:32:09 -04:00
|
|
|
explicit WriterThreadState(std::string_view dir)
|
|
|
|
|
: dirPath{dir.empty() ? "." : dir} {}
|
2024-05-12 14:09:43 -07:00
|
|
|
WriterThreadState(const WriterThreadState&) = delete;
|
|
|
|
|
WriterThreadState& operator=(const WriterThreadState&) = delete;
|
|
|
|
|
~WriterThreadState() { Close(); }
|
|
|
|
|
|
|
|
|
|
void Close() {
|
2025-01-30 22:01:35 -05:00
|
|
|
if (f != WPI_kInvalidFile) {
|
2024-05-12 14:09:43 -07:00
|
|
|
fs::CloseFile(f);
|
2025-01-30 22:01:35 -05:00
|
|
|
f = WPI_kInvalidFile;
|
2024-05-12 14:09:43 -07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void SetFilename(std::string_view fn) {
|
|
|
|
|
baseFilename = fn;
|
|
|
|
|
filename = fn;
|
|
|
|
|
path = dirPath / filename;
|
|
|
|
|
segmentCount = 1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void IncrementFilename() {
|
|
|
|
|
fs::path basePath{baseFilename};
|
|
|
|
|
filename = fmt::format("{}.{}{}", basePath.stem().string(), ++segmentCount,
|
|
|
|
|
basePath.extension().string());
|
|
|
|
|
path = dirPath / filename;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
fs::path dirPath;
|
|
|
|
|
std::string baseFilename;
|
|
|
|
|
std::string filename;
|
|
|
|
|
fs::path path;
|
2025-01-30 22:01:35 -05:00
|
|
|
fs::file_t f = WPI_kInvalidFile;
|
2024-05-12 14:09:43 -07:00
|
|
|
uintmax_t freeSpace = UINTMAX_MAX;
|
|
|
|
|
int segmentCount = 1;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::BufferHalfFull() {
|
|
|
|
|
Flush();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool DataLogBackgroundWriter::BufferFull() {
|
|
|
|
|
WPI_ERROR(m_msglog,
|
|
|
|
|
"outgoing buffers exceeded threshold, pausing logging--"
|
|
|
|
|
"consider flushing to disk more frequently (smaller period)");
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::StartLogFile(WriterThreadState& state) {
|
|
|
|
|
std::error_code ec;
|
|
|
|
|
|
|
|
|
|
if (state.filename.empty()) {
|
|
|
|
|
state.SetFilename(MakeRandomFilename());
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// get free space
|
|
|
|
|
auto freeSpaceInfo = fs::space(state.dirPath, ec);
|
|
|
|
|
if (!ec) {
|
|
|
|
|
state.freeSpace = freeSpaceInfo.available;
|
|
|
|
|
} else {
|
|
|
|
|
state.freeSpace = UINTMAX_MAX;
|
|
|
|
|
}
|
|
|
|
|
if (state.freeSpace < kMinFreeSpace) {
|
|
|
|
|
WPI_ERROR(m_msglog,
|
|
|
|
|
"Insufficient free space ({} available), no log being saved",
|
|
|
|
|
FormatBytesSize(state.freeSpace));
|
2024-06-10 22:22:50 -04:00
|
|
|
m_state = kStopped;
|
2024-05-12 14:09:43 -07:00
|
|
|
} else {
|
|
|
|
|
// try preferred filename, or randomize it a few times, before giving up
|
|
|
|
|
for (int i = 0; i < 5; ++i) {
|
|
|
|
|
// open file for append
|
|
|
|
|
#ifdef _WIN32
|
|
|
|
|
// WIN32 doesn't allow combination of CreateNew and Append
|
|
|
|
|
state.f =
|
|
|
|
|
fs::OpenFileForWrite(state.path, ec, fs::CD_CreateNew, fs::OF_None);
|
|
|
|
|
#else
|
|
|
|
|
state.f =
|
|
|
|
|
fs::OpenFileForWrite(state.path, ec, fs::CD_CreateNew, fs::OF_Append);
|
|
|
|
|
#endif
|
|
|
|
|
if (ec) {
|
|
|
|
|
WPI_ERROR(m_msglog, "Could not open log file '{}': {}",
|
|
|
|
|
state.path.string(), ec.message());
|
|
|
|
|
// try again with random filename
|
|
|
|
|
state.SetFilename(MakeRandomFilename());
|
|
|
|
|
} else {
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2025-01-30 22:01:35 -05:00
|
|
|
if (state.f == WPI_kInvalidFile) {
|
2024-05-12 14:09:43 -07:00
|
|
|
WPI_ERROR(m_msglog, "Could not open log file, no log being saved");
|
|
|
|
|
} else {
|
|
|
|
|
WPI_INFO(m_msglog, "Logging to '{}' ({} free space)", state.path.string(),
|
|
|
|
|
FormatBytesSize(state.freeSpace));
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// start file
|
2025-01-30 22:01:35 -05:00
|
|
|
if (state.f != WPI_kInvalidFile) {
|
2024-05-12 14:09:43 -07:00
|
|
|
StartFile();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::WriterThreadMain(std::string_view dir) {
|
|
|
|
|
std::chrono::duration<double> periodTime{m_period};
|
|
|
|
|
|
|
|
|
|
WriterThreadState state{dir};
|
|
|
|
|
{
|
|
|
|
|
std::scoped_lock lock{m_mutex};
|
|
|
|
|
state.SetFilename(m_newFilename);
|
|
|
|
|
m_newFilename.clear();
|
|
|
|
|
}
|
|
|
|
|
StartLogFile(state);
|
|
|
|
|
|
|
|
|
|
std::error_code ec;
|
|
|
|
|
std::vector<DataLog::Buffer> toWrite;
|
|
|
|
|
int freeSpaceCount = 0;
|
|
|
|
|
int checkExistCount = 0;
|
|
|
|
|
bool blocked = false;
|
|
|
|
|
uintmax_t written = 0;
|
|
|
|
|
|
|
|
|
|
std::unique_lock lock{m_mutex};
|
|
|
|
|
do {
|
|
|
|
|
bool doFlush = false;
|
|
|
|
|
auto timeoutTime = std::chrono::steady_clock::now() + periodTime;
|
|
|
|
|
if (m_cond.wait_until(lock, timeoutTime) == std::cv_status::timeout) {
|
|
|
|
|
doFlush = true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (m_state == kStopped) {
|
|
|
|
|
state.Close();
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool doStart = false;
|
|
|
|
|
|
|
|
|
|
// if file was deleted, recreate it with the same name
|
|
|
|
|
if (++checkExistCount >= 10) {
|
|
|
|
|
checkExistCount = 0;
|
|
|
|
|
lock.unlock();
|
|
|
|
|
bool exists = fs::exists(state.path, ec);
|
|
|
|
|
lock.lock();
|
|
|
|
|
if (!ec && !exists) {
|
|
|
|
|
state.Close();
|
|
|
|
|
state.IncrementFilename();
|
|
|
|
|
WPI_INFO(m_msglog, "Log file deleted, recreating as fresh log '{}'",
|
|
|
|
|
state.filename);
|
|
|
|
|
doStart = true;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// start new file if file exceeds 1.8 GB
|
|
|
|
|
if (written > 1800000000ull) {
|
|
|
|
|
state.Close();
|
|
|
|
|
state.IncrementFilename();
|
|
|
|
|
WPI_INFO(m_msglog, "Log file reached 1.8 GB, starting new file '{}'",
|
|
|
|
|
state.filename);
|
|
|
|
|
doStart = true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (m_state == kStart || doStart) {
|
|
|
|
|
lock.unlock();
|
|
|
|
|
DataLog::Stop();
|
|
|
|
|
StartLogFile(state);
|
|
|
|
|
lock.lock();
|
|
|
|
|
if (m_state == kStopped) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
m_state = kActive;
|
|
|
|
|
written = 0;
|
|
|
|
|
}
|
|
|
|
|
|
2025-01-30 22:01:35 -05:00
|
|
|
if (!m_newFilename.empty() && state.f != WPI_kInvalidFile) {
|
2024-05-12 14:09:43 -07:00
|
|
|
auto newFilename = std::move(m_newFilename);
|
|
|
|
|
m_newFilename.clear();
|
|
|
|
|
// rename
|
|
|
|
|
if (state.filename != newFilename) {
|
|
|
|
|
lock.unlock();
|
|
|
|
|
fs::rename(state.path, state.dirPath / newFilename, ec);
|
|
|
|
|
lock.lock();
|
|
|
|
|
}
|
|
|
|
|
if (ec) {
|
|
|
|
|
WPI_ERROR(m_msglog, "Could not rename log file from '{}' to '{}': {}",
|
|
|
|
|
state.filename, newFilename, ec.message());
|
|
|
|
|
} else {
|
|
|
|
|
WPI_INFO(m_msglog, "Renamed log file from '{}' to '{}'", state.filename,
|
|
|
|
|
newFilename);
|
|
|
|
|
}
|
|
|
|
|
state.SetFilename(newFilename);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (doFlush || m_doFlush) {
|
|
|
|
|
// flush to file
|
|
|
|
|
m_doFlush = false;
|
|
|
|
|
DataLog::FlushBufs(&toWrite);
|
|
|
|
|
if (toWrite.empty()) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2025-01-30 22:01:35 -05:00
|
|
|
if (state.f != WPI_kInvalidFile && !blocked) {
|
2024-05-12 14:09:43 -07:00
|
|
|
lock.unlock();
|
|
|
|
|
|
|
|
|
|
// update free space every 10 flushes (in case other things are writing)
|
|
|
|
|
if (++freeSpaceCount >= 10) {
|
|
|
|
|
freeSpaceCount = 0;
|
|
|
|
|
auto freeSpaceInfo = fs::space(state.dirPath, ec);
|
|
|
|
|
if (!ec) {
|
|
|
|
|
state.freeSpace = freeSpaceInfo.available;
|
|
|
|
|
} else {
|
|
|
|
|
state.freeSpace = UINTMAX_MAX;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// write buffers to file
|
|
|
|
|
for (auto&& buf : toWrite) {
|
|
|
|
|
// stop writing when we go below the minimum free space
|
|
|
|
|
state.freeSpace -= buf.GetData().size();
|
|
|
|
|
written += buf.GetData().size();
|
|
|
|
|
if (state.freeSpace < kMinFreeSpace) {
|
|
|
|
|
[[unlikely]] WPI_ERROR(
|
|
|
|
|
m_msglog,
|
|
|
|
|
"Stopped logging due to low free space ({} available)",
|
|
|
|
|
FormatBytesSize(state.freeSpace));
|
|
|
|
|
blocked = true;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
WriteToFile(state.f, buf.GetData(), state.filename, m_msglog);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// sync to storage
|
|
|
|
|
#if defined(__linux__)
|
|
|
|
|
::fdatasync(state.f);
|
|
|
|
|
#elif defined(__APPLE__)
|
|
|
|
|
::fsync(state.f);
|
|
|
|
|
#endif
|
|
|
|
|
lock.lock();
|
|
|
|
|
if (blocked) {
|
|
|
|
|
[[unlikely]] m_state = kPaused;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// release buffers back to free list
|
|
|
|
|
ReleaseBufs(&toWrite);
|
|
|
|
|
}
|
|
|
|
|
} while (!m_shutdown);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void DataLogBackgroundWriter::WriterThreadMain(
|
|
|
|
|
std::function<void(std::span<const uint8_t> data)> write) {
|
|
|
|
|
std::chrono::duration<double> periodTime{m_period};
|
|
|
|
|
|
|
|
|
|
StartFile();
|
|
|
|
|
|
|
|
|
|
std::vector<DataLog::Buffer> toWrite;
|
|
|
|
|
|
|
|
|
|
std::unique_lock lock{m_mutex};
|
|
|
|
|
do {
|
|
|
|
|
bool doFlush = false;
|
|
|
|
|
auto timeoutTime = std::chrono::steady_clock::now() + periodTime;
|
|
|
|
|
if (m_cond.wait_until(lock, timeoutTime) == std::cv_status::timeout) {
|
|
|
|
|
doFlush = true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (doFlush || m_doFlush) {
|
|
|
|
|
// flush to file
|
|
|
|
|
m_doFlush = false;
|
|
|
|
|
DataLog::FlushBufs(&toWrite);
|
|
|
|
|
if (toWrite.empty()) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
lock.unlock();
|
|
|
|
|
// write buffers
|
|
|
|
|
for (auto&& buf : toWrite) {
|
|
|
|
|
if (!buf.GetData().empty()) {
|
|
|
|
|
write(buf.GetData());
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
lock.lock();
|
|
|
|
|
|
|
|
|
|
// release buffers back to free list
|
|
|
|
|
ReleaseBufs(&toWrite);
|
|
|
|
|
}
|
|
|
|
|
} while (!m_shutdown);
|
|
|
|
|
|
|
|
|
|
write({}); // indicate EOF
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
extern "C" {
|
|
|
|
|
|
|
|
|
|
struct WPI_DataLog* WPI_DataLog_CreateBackgroundWriter(
|
Change C APIs to a unified string implementation (#6299)
Currently in the entire C API of WPILib we have ~8 different ways of handling strings. The C API actually isn't built for pure C callers (We don't actually have any of those). Instead, they're built for interop between languages like LabVIEW and C# which can talk to C API's directly.
For output parameters, the choice was fairly obvious. An output struct containing a const string pointer and a length makes the most sense. Its easy to use these from most other languages, and doesn't require special null termination handling. Freeing these is also easy, as if you ever receive one of these string structures, theres just a single function call to free it.
Input parameters are a bit more complex. To be used from pure C, and from LabVIEW, a null terminated string is the best in most cases. However, null terminated strings in general have a lot of downsides. Additionally, from LabVIEW there are other considerations around encoding that having a wrapper struct helps make a bit easier. From a language like C#, a wrapper struct is by far the easiest, as custom marshalling can make it trivial to marshal both UTF8 and UTF16 strings down.
The final consideration is its nice to have an identical concept for both input and output. It makes the rules fairly easy to understand.
WPILib will not have any APIs that manipulate a string allocated externally. This means WPI_String can be const, as across the boundary it is always const.
If a WPILib API takes a const WPI_String*, WPILib will not manipulate or attempt to free that string, and that string is treated as an input. It is up to the caller to handle that memory, WPILib will never hold onto that memory longer than the call.
If a WPILib API takes a WPI_String*, that string is an output. WPILib will allocate that API with WPI_AllocateString(), fill in the string, and return to the caller. When the caller is done with the string, they must free it with WPI_FreeString().
If an output struct contains a WPI_String member, that member is considered read only, and should not be explicitly freed. The caller should call the free function for that struct.
If an array of WPI_Strings are returned, each individual string is considered read only, and should not be explicitly freed. The free function for that array should be called by the caller.
If an input struct containing a WPI_String, or an input array of WPI_Strings is passed to WPILib, the individual strings will not be manipulated or freed by WPILib, and the caller owns and should free that memory.
Callbacks also follow these rules. The most common is a callback either getting passed a const WPI_String* or a struct containing a WPI_String. In both of these cases, the callback target should consider these strings read only, and not attempt to free them or manipulate them.
2024-05-13 05:35:14 -07:00
|
|
|
const struct WPI_String* dir, const struct WPI_String* filename,
|
|
|
|
|
double period, const struct WPI_String* extraHeader) {
|
|
|
|
|
return reinterpret_cast<WPI_DataLog*>(new DataLogBackgroundWriter{
|
2025-11-07 20:00:05 -05:00
|
|
|
wpi::util::to_string_view(dir), wpi::util::to_string_view(filename), period,
|
|
|
|
|
wpi::util::to_string_view(extraHeader)});
|
2024-05-12 14:09:43 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct WPI_DataLog* WPI_DataLog_CreateBackgroundWriter_Func(
|
|
|
|
|
void (*write)(void* ptr, const uint8_t* data, size_t len), void* ptr,
|
Change C APIs to a unified string implementation (#6299)
Currently in the entire C API of WPILib we have ~8 different ways of handling strings. The C API actually isn't built for pure C callers (We don't actually have any of those). Instead, they're built for interop between languages like LabVIEW and C# which can talk to C API's directly.
For output parameters, the choice was fairly obvious. An output struct containing a const string pointer and a length makes the most sense. Its easy to use these from most other languages, and doesn't require special null termination handling. Freeing these is also easy, as if you ever receive one of these string structures, theres just a single function call to free it.
Input parameters are a bit more complex. To be used from pure C, and from LabVIEW, a null terminated string is the best in most cases. However, null terminated strings in general have a lot of downsides. Additionally, from LabVIEW there are other considerations around encoding that having a wrapper struct helps make a bit easier. From a language like C#, a wrapper struct is by far the easiest, as custom marshalling can make it trivial to marshal both UTF8 and UTF16 strings down.
The final consideration is its nice to have an identical concept for both input and output. It makes the rules fairly easy to understand.
WPILib will not have any APIs that manipulate a string allocated externally. This means WPI_String can be const, as across the boundary it is always const.
If a WPILib API takes a const WPI_String*, WPILib will not manipulate or attempt to free that string, and that string is treated as an input. It is up to the caller to handle that memory, WPILib will never hold onto that memory longer than the call.
If a WPILib API takes a WPI_String*, that string is an output. WPILib will allocate that API with WPI_AllocateString(), fill in the string, and return to the caller. When the caller is done with the string, they must free it with WPI_FreeString().
If an output struct contains a WPI_String member, that member is considered read only, and should not be explicitly freed. The caller should call the free function for that struct.
If an array of WPI_Strings are returned, each individual string is considered read only, and should not be explicitly freed. The free function for that array should be called by the caller.
If an input struct containing a WPI_String, or an input array of WPI_Strings is passed to WPILib, the individual strings will not be manipulated or freed by WPILib, and the caller owns and should free that memory.
Callbacks also follow these rules. The most common is a callback either getting passed a const WPI_String* or a struct containing a WPI_String. In both of these cases, the callback target should consider these strings read only, and not attempt to free them or manipulate them.
2024-05-13 05:35:14 -07:00
|
|
|
double period, const struct WPI_String* extraHeader) {
|
2024-05-12 14:09:43 -07:00
|
|
|
return reinterpret_cast<WPI_DataLog*>(new DataLogBackgroundWriter{
|
|
|
|
|
[=](auto data) { write(ptr, data.data(), data.size()); }, period,
|
2025-11-07 20:00:05 -05:00
|
|
|
wpi::util::to_string_view(extraHeader)});
|
2024-05-12 14:09:43 -07:00
|
|
|
}
|
|
|
|
|
|
Change C APIs to a unified string implementation (#6299)
Currently in the entire C API of WPILib we have ~8 different ways of handling strings. The C API actually isn't built for pure C callers (We don't actually have any of those). Instead, they're built for interop between languages like LabVIEW and C# which can talk to C API's directly.
For output parameters, the choice was fairly obvious. An output struct containing a const string pointer and a length makes the most sense. Its easy to use these from most other languages, and doesn't require special null termination handling. Freeing these is also easy, as if you ever receive one of these string structures, theres just a single function call to free it.
Input parameters are a bit more complex. To be used from pure C, and from LabVIEW, a null terminated string is the best in most cases. However, null terminated strings in general have a lot of downsides. Additionally, from LabVIEW there are other considerations around encoding that having a wrapper struct helps make a bit easier. From a language like C#, a wrapper struct is by far the easiest, as custom marshalling can make it trivial to marshal both UTF8 and UTF16 strings down.
The final consideration is its nice to have an identical concept for both input and output. It makes the rules fairly easy to understand.
WPILib will not have any APIs that manipulate a string allocated externally. This means WPI_String can be const, as across the boundary it is always const.
If a WPILib API takes a const WPI_String*, WPILib will not manipulate or attempt to free that string, and that string is treated as an input. It is up to the caller to handle that memory, WPILib will never hold onto that memory longer than the call.
If a WPILib API takes a WPI_String*, that string is an output. WPILib will allocate that API with WPI_AllocateString(), fill in the string, and return to the caller. When the caller is done with the string, they must free it with WPI_FreeString().
If an output struct contains a WPI_String member, that member is considered read only, and should not be explicitly freed. The caller should call the free function for that struct.
If an array of WPI_Strings are returned, each individual string is considered read only, and should not be explicitly freed. The free function for that array should be called by the caller.
If an input struct containing a WPI_String, or an input array of WPI_Strings is passed to WPILib, the individual strings will not be manipulated or freed by WPILib, and the caller owns and should free that memory.
Callbacks also follow these rules. The most common is a callback either getting passed a const WPI_String* or a struct containing a WPI_String. In both of these cases, the callback target should consider these strings read only, and not attempt to free them or manipulate them.
2024-05-13 05:35:14 -07:00
|
|
|
void WPI_DataLog_SetBackgroundWriterFilename(
|
|
|
|
|
struct WPI_DataLog* datalog, const struct WPI_String* filename) {
|
|
|
|
|
reinterpret_cast<DataLogBackgroundWriter*>(datalog)->SetFilename(
|
2025-11-07 20:00:05 -05:00
|
|
|
wpi::util::to_string_view(filename));
|
2024-05-12 14:09:43 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
} // extern "C"
|