From 32e14551a90e85000e41b3f0445d34d58a1431e4 Mon Sep 17 00:00:00 2001
From: Joel Klinghed <the_jk@spawned.biz>
Date: Wed, 10 Sep 2025 22:12:22 +0200
Subject: Add unicode general category lookup

Generate the lookup tables from UnicodeData.txt, do to that,
add gen_ugc, which uses csv, buffers, line, io and other modules
to do the job.
---
 src/decompress_z.cc | 120 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 120 insertions(+)
 create mode 100644 src/decompress_z.cc

(limited to 'src/decompress_z.cc')

diff --git a/src/decompress_z.cc b/src/decompress_z.cc
new file mode 100644
index 0000000..f9f87ae
--- /dev/null
+++ b/src/decompress_z.cc
@@ -0,0 +1,120 @@
+#include "decompress.hh"
+
+#include "buffer.hh"
+
+#define ZLIB_CONST
+#include <zlib.h>
+
+#include <algorithm>
+#include <cstddef>
+#include <expected>
+#include <limits>
+#include <memory>
+#include <optional>
+#include <utility>
+
+namespace decompress {
+
+namespace {
+
+const size_t kBufferSizeZ = static_cast<size_t>(1024) * 1024;
+
+class DecompressReader : public io::Reader {
+ public:
+  DecompressReader(std::unique_ptr<io::Reader> reader, bool gzip)
+      : reader_(std::move(reader)), gzip_(gzip) {}
+
+  ~DecompressReader() override {
+    if (initialized_)
+      inflateEnd(&stream_);
+  }
+
+  std::expected<size_t, io::ReadError> read(void* dst, size_t max) override {
+    auto err = fill();
+    if (err.has_value())
+      return std::unexpected(err.value());
+
+    // NOLINTNEXTLINE(misc-include-cleaner)
+    stream_.next_out = reinterpret_cast<Bytef*>(dst);
+    stream_.avail_out = max;
+
+    if (!initialized_) {
+      if (in_eof_ && buffer_->empty())
+        return 0;
+
+      stream_.zalloc = Z_NULL;
+      stream_.zfree = Z_NULL;
+      stream_.opaque = Z_NULL;
+      if (inflateInit2(&stream_, gzip_ ? 16 : 0) != Z_OK) {
+        return std::unexpected(io::ReadError::Error);
+      }
+      initialized_ = true;
+    }
+
+    auto* const rptr = stream_.next_in;
+    auto ret = inflate(&stream_, in_eof_ ? Z_FINISH : Z_NO_FLUSH);
+    auto got = max - stream_.avail_out;
+    if (ret == Z_STREAM_END) {
+      inflateEnd(&stream_);
+      initialized_ = false;
+      buffer_->consume(stream_.next_in - rptr);
+    } else if (ret == Z_OK) {
+      if (!in_eof_)
+        buffer_->consume(stream_.next_in - rptr);
+    } else {
+      return std::unexpected(
+          ret == Z_DATA_ERROR
+          ? io::ReadError::InvalidData : io::ReadError::Error);
+    }
+    return got;
+  }
+
+  std::expected<size_t, io::ReadError> skip(size_t max) override {
+    auto tmp = std::make_unique_for_overwrite<char[]>(max);
+    return read(tmp.get(), max);
+  }
+
+ private:
+  std::optional<io::ReadError> fill() {
+    size_t avail;
+    auto* rptr = buffer_->rptr(avail);
+    if (!in_eof_ && avail < kBufferSizeZ / 2) {
+      auto* wptr = buffer_->wptr(avail);
+      auto got = reader_->read(wptr, avail);
+      if (got.has_value()) {
+        buffer_->commit(got.value());
+        if (got.value() == 0)
+          in_eof_ = true;
+      } else {
+        return got.error();
+      }
+      rptr = buffer_->rptr(avail);
+    }
+    // NOLINTNEXTLINE(misc-include-cleaner)
+    stream_.next_in = reinterpret_cast<z_const Bytef*>(rptr);
+    stream_.avail_in = std::min(
+        // NOLINTNEXTLINE(misc-include-cleaner)
+        static_cast<size_t>(std::numeric_limits<uInt>::max()), avail);
+    return std::nullopt;
+  }
+
+  std::unique_ptr<io::Reader> reader_;
+  bool const gzip_;
+  bool in_eof_{false};
+  std::unique_ptr<Buffer> buffer_{Buffer::fixed(kBufferSizeZ)};
+  bool initialized_{false};
+  z_stream stream_;
+};
+
+}  // namespace
+
+std::unique_ptr<io::Reader> zlib(std::unique_ptr<io::Reader> reader) {
+  return std::make_unique<DecompressReader>(std::move(reader), /* gzip = */ false);
+}
+
+std::unique_ptr<io::Reader> gzip(std::unique_ptr<io::Reader> reader) {
+  return std::make_unique<DecompressReader>(std::move(reader), /* gzip = */ true);
+}
+
+
+}  // namespace decompress
-- 
cgit v1.3