Move Serializer into its own file.

diff --git a/meson.build b/meson.build
index 2e68459c44428315809731dfa00dec92ac9fa5d1..7c6755bcd695e01af23b410d05fc8599e9d63713 100644 (file)
--- a/meson.build
+++ b/meson.build
@@ -12,7 +12,7 @@ if not uringdep.found()
        add_project_arguments('-DWITHOUT_URING', language: 'cpp')
 endif
 
-executable('plocate', ['plocate.cpp', 'io_uring_engine.cpp', 'turbopfor.cpp', 'parse_trigrams.cpp'],
+executable('plocate', ['plocate.cpp', 'io_uring_engine.cpp', 'turbopfor.cpp', 'parse_trigrams.cpp', 'serializer.cpp'],
        dependencies: [uringdep, zstddep, threaddep],
        install: true,
        install_mode: ['rwxr-sr-x', 'root', 'mlocate'])

diff --git a/options.h b/options.h
new file mode 100644 (file)
index 0000000..683c8d3
--- /dev/null
+++ b/options.h
@@ -0,0 +1,16 @@
+#ifndef _OPTIONS_H
+#define _OPTIONS_H
+
+#include <stdint.h>
+
+extern const char *dbpath;
+extern bool ignore_case;
+extern bool only_count;
+extern bool print_nul;
+extern bool use_debug;
+extern bool patterns_are_regex;
+extern bool use_extended_regex;
+extern int64_t limit_matches;
+extern int64_t limit_left;  // Not strictly an option.
+
+#endif  // !defined(_OPTIONS_H)

diff --git a/plocate.cpp b/plocate.cpp
index 926bbf1977831562217829cc07e7aaddf56ca4f8..2f3129326662bd52512f9ed50ba011acf627627f 100644 (file)
--- a/plocate.cpp
+++ b/plocate.cpp
@@ -2,34 +2,36 @@
 #include "dprintf.h"
 #include "io_uring_engine.h"
 #include "parse_trigrams.h"
+#include "serializer.h"
 #include "turbopfor.h"
 #include "unique_sort.h"
 
 #include <algorithm>
-#include <atomic>
 #include <assert.h>
+#include <atomic>
 #include <chrono>
 #include <condition_variable>
+#include <deque>
 #include <fcntl.h>
 #include <fnmatch.h>
 #include <functional>
 #include <getopt.h>
 #include <inttypes.h>
-#include <iosfwd>
 #include <iterator>
 #include <limits>
+#include <locale.h>
 #include <map>
 #include <memory>
 #include <mutex>
-#include <queue>
 #include <regex.h>
-#include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
+#include <stdint.h>
 #include <string.h>
 #include <string>
 #include <string_view>
 #include <thread>
+#include <tuple>
 #include <unistd.h>
 #include <unordered_map>
 #include <unordered_set>
@@ -57,83 +59,6 @@ ZSTD_DDict *ddict = nullptr;
 
 regex_t compile_regex(const string &needle);
 
-void apply_limit()
-{
-       if (--limit_left > 0) {
-               return;
-       }
-       dprintf("Done in %.1f ms, found %" PRId64 " matches.\n",
-               1e3 * duration<float>(steady_clock::now() - start).count(), limit_matches);
-       if (only_count) {
-               printf("%" PRId64 "\n", limit_matches);
-       }
-       exit(0);
-}
-
-class ResultReceiver {
-public:
-       virtual ~ResultReceiver() = default;
-       virtual void print(uint64_t seq, uint64_t skip, const string msg) = 0;
-};
-
-class Serializer : public ResultReceiver {
-public:
-       ~Serializer() { assert(limit_left <= 0 || pending.empty()); }
-       void print(uint64_t seq, uint64_t skip, const string msg) override;
-
-private:
-       uint64_t next_seq = 0;
-       struct Element {
-               uint64_t seq, skip;
-               string msg;
-
-               bool operator<(const Element &other) const
-               {
-                       return seq > other.seq;
-               }
-       };
-       priority_queue<Element> pending;
-};
-
-void Serializer::print(uint64_t seq, uint64_t skip, const string msg)
-{
-       if (only_count) {
-               if (!msg.empty()) {
-                       apply_limit();
-               }
-               return;
-       }
-
-       if (next_seq != seq) {
-               pending.push(Element{ seq, skip, move(msg) });
-               return;
-       }
-
-       if (!msg.empty()) {
-               if (print_nul) {
-                       printf("%s%c", msg.c_str(), 0);
-               } else {
-                       printf("%s\n", msg.c_str());
-               }
-               apply_limit();
-       }
-       next_seq += skip;
-
-       // See if any delayed prints can now be dealt with.
-       while (!pending.empty() && pending.top().seq == next_seq) {
-               if (!pending.top().msg.empty()) {
-                       if (print_nul) {
-                               printf("%s%c", pending.top().msg.c_str(), 0);
-                       } else {
-                               printf("%s\n", pending.top().msg.c_str());
-                       }
-                       apply_limit();
-               }
-               next_seq += pending.top().skip;
-               pending.pop();
-       }
-}
-
 struct Needle {
        enum { STRSTR,
               REGEX,
@@ -399,7 +324,7 @@ size_t scan_docids(const vector<Needle> &needles, const vector<uint32_t> &docids
 {
        Serializer docids_in_order;
        AccessRXCache access_rx_cache(engine);
-       atomic<uint64_t> matched{0};
+       atomic<uint64_t> matched{ 0 };
        for (size_t i = 0; i < docids.size(); ++i) {
                uint32_t docid = docids[i];
                corpus.get_compressed_filename_block(docid, [i, &matched, &needles, &access_rx_cache, &docids_in_order](string_view compressed) {
@@ -427,7 +352,8 @@ struct WorkerThread {
 
 class WorkerThreadReceiver : public ResultReceiver {
 public:
-       WorkerThreadReceiver(WorkerThread *wt) : wt(wt) {}
+       WorkerThreadReceiver(WorkerThread *wt)
+               : wt(wt) {}
 
        void print(uint64_t seq, uint64_t skip, const string msg) override
        {

diff --git a/serializer.cpp b/serializer.cpp
new file mode 100644 (file)
index 0000000..9277125
--- /dev/null
+++ b/serializer.cpp
@@ -0,0 +1,67 @@
+#include <inttypes.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <chrono>
+#include <memory>
+#include <utility>
+
+#include "dprintf.h"
+#include "serializer.h"
+
+using namespace std;
+using namespace std::chrono;
+
+extern steady_clock::time_point start;
+
+void apply_limit()
+{
+       if (--limit_left > 0) {
+               return;
+       }
+       dprintf("Done in %.1f ms, found %" PRId64 " matches.\n",
+               1e3 * duration<float>(steady_clock::now() - start).count(), limit_matches);
+       if (only_count) {
+               printf("%" PRId64 "\n", limit_matches);
+       }
+       exit(0);
+}
+
+void Serializer::print(uint64_t seq, uint64_t skip, const string msg)
+{
+       if (only_count) {
+               if (!msg.empty()) {
+                       apply_limit();
+               }
+               return;
+       }
+
+       if (next_seq != seq) {
+               pending.push(Element{ seq, skip, move(msg) });
+               return;
+       }
+
+       if (!msg.empty()) {
+               if (print_nul) {
+                       printf("%s%c", msg.c_str(), 0);
+               } else {
+                       printf("%s\n", msg.c_str());
+               }
+               apply_limit();
+       }
+       next_seq += skip;
+
+       // See if any delayed prints can now be dealt with.
+       while (!pending.empty() && pending.top().seq == next_seq) {
+               if (!pending.top().msg.empty()) {
+                       if (print_nul) {
+                               printf("%s%c", pending.top().msg.c_str(), 0);
+                       } else {
+                               printf("%s\n", pending.top().msg.c_str());
+                       }
+                       apply_limit();
+               }
+               next_seq += pending.top().skip;
+               pending.pop();
+       }
+}
+

diff --git a/serializer.h b/serializer.h
new file mode 100644 (file)
index 0000000..580e91d
--- /dev/null
+++ b/serializer.h
@@ -0,0 +1,36 @@
+#ifndef _SERIALIZER_H
+#define _SERIALIZER_H 1
+
+#include <assert.h>
+#include <bits/stdint-uintn.h>
+#include <queue>
+#include <string>
+
+#include "options.h"
+
+class ResultReceiver {
+public:
+       virtual ~ResultReceiver() = default;
+       virtual void print(uint64_t seq, uint64_t skip, const std::string msg) = 0;
+};
+
+class Serializer : public ResultReceiver {
+public:
+       ~Serializer() { assert(limit_left <= 0 || pending.empty()); }
+       void print(uint64_t seq, uint64_t skip, const std::string msg) override;
+
+private:
+       uint64_t next_seq = 0;
+       struct Element {
+               uint64_t seq, skip;
+               std::string msg;
+
+               bool operator<(const Element &other) const
+               {
+                       return seq > other.seq;
+               }
+       };
+       std::priority_queue<Element> pending;
+};
+
+#endif  // !defined(_SERIALIZER_H)