sots-engine/tests/mars_stream/test_save.cpp

// Real-save test: reads every *.sav in $SOTS_SAVES_DIR (owner's data, never
// in the repo) and checks that the walker parses clean, the typed shapes
// load, both round trips are byte-identical and the RNG blob is a valid
// MT19937 state.  Skips (exit 0) when the variable is unset.
//
// With SOTS_DUMP_DIR set, writes <name>.cpp.dump / <name>.cpp.summary there
// for tests/mars_stream/oracle/compare.py.
#include <algorithm>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <dirent.h>
#include <fstream>
#include <string>
#include <vector>

#include "mars/rng/mt19937.h"
#include "mars/stream/dump.h"
#include "mars/stream/probe.h"
#include "mars/stream/save.h"

using namespace mars::stream;

static int fails = 0;
#define CHECK(cond)                                                                     \
    do {                                                                                \
        if (!(cond)) {                                                                  \
            std::printf("FAIL %s:%d: %s\n", __FILE__, __LINE__, #cond);                 \
            ++fails;                                                                    \
        }                                                                               \
    } while (0)

static std::vector<std::string> list_saves(const std::string& dir) {
    std::vector<std::string> out;
    DIR* d = opendir(dir.c_str());
    if (!d) return out;
    while (dirent* e = readdir(d)) {
        std::string n = e->d_name;
        if (n.size() > 4 && n.compare(n.size() - 4, 4, ".sav") == 0) out.push_back(dir + "/" + n);
    }
    closedir(d);
    std::sort(out.begin(), out.end());
    return out;
}

static void check_save(const std::string& path, const char* dump_dir) {
    std::printf("== %s\n", path.c_str());
    SaveDocument doc = read_save_file(path);
    std::printf("   inflated %zu bytes, items %u, frames %u, resyncs %u, hint-failures %u, raw-bytes %u\n",
                doc.inflated.size(), doc.stats.items, doc.stats.frames, doc.stats.resyncs, doc.stats.hint_failures,
                doc.stats.raw_bytes);
    std::printf("   issues: %zu error, %zu warn, %zu info\n", doc.count(Issue::Error), doc.count(Issue::Warn),
                doc.count(Issue::Info));
    for (const Issue& i : doc.issues)
        if (i.level != Issue::Info) std::printf("     %s\n", format_issue(i).c_str());

    // --- the confirmed format parses clean --------------------------------------
    CHECK(doc.stats.resyncs == 0);
    CHECK(doc.stats.hint_failures == 0);
    CHECK(doc.count(Issue::Error) == 0);
    CHECK(doc.count(Issue::Warn) == 0);
    CHECK(doc.stats.raw_bytes == 2503);  // only the opaque RNG blob

    // --- typed shapes -------------------------------------------------------------
    const auto& s = doc.game.summary;
    const auto& sim = doc.game.sim;
    std::printf("   summary: game='%s' turn=%d numSys=%d players=%zu\n", s.gameName.c_str(), s.turn, s.numSys,
                s.players.size());
    CHECK(!s.gameName.empty());
    CHECK(s.turn >= 1);
    CHECK(s.numSys > 0);
    CHECK(!s.players.empty());
    CHECK(sim.gameName == s.gameName);
    CHECK(int(sim.systems.size()) == s.numSys);
    CHECK(sim.systems.size() == sim.systemIds.size());
    CHECK(sim.players.size() == sim.playerIds.size());
    CHECK(sim.fleets.size() == sim.fleetIds.size());
    CHECK(sim.species.size() == 7);
    CHECK(!sim.players.empty() && !sim.players[0].player.plryName.empty());
    CHECK(doc.game.createParams.name == s.gameName);
    CHECK(doc.game.createParams.nSys == s.numSys);
    CHECK(doc.game.createParams.mapP.planets.size() == size_t(s.numSys));
    for (const auto& se : sim.systems) CHECK(!se.sys.name.empty());
    for (const auto& fe : sim.fleets) CHECK(fe.flt.ships.size() >= 1);
    CHECK(!doc.game.cdTable.ids.empty());

    // Every ".TurnCommands_v5" block must be consumed by the prologue plus the 27
    // counted lists with nothing left over.  `extra` is the tail `ar.rest()` would
    // absorb, so an empty `extra` is the item-granular statement that the model
    // accounts for the whole frame -- on the 35-item empty blocks and on the
    // 123-item one alike.  It is what a wrong list count or a wrong element width
    // would break first.
    for (const auto& cd : doc.game.customData) {
        if (cd.which != shapes::CustomDataBlock::Which::TurnCmds) continue;
        CHECK(cd.turnCommands.extra.empty());
        CHECK(cd.turnCommands.playerID != 0);
        std::printf("   TurnCommands: player %d, rate %.4g, %zu build, %zu sysRate, %zu colonize, %zu move\n",
                    cd.turnCommands.playerID, double(cd.turnCommands.researchRate),
                    cd.turnCommands.buildOrders.size(), cd.turnCommands.systemRates.size(),
                    cd.turnCommands.colonizeOrders.size(), cd.turnCommands.fleetMoves.size());
    }

    // --- RNG blob: mt[624] + left, produced by our MT19937 from RSeed --------------
    CHECK(sim.rng.is_complex() && sim.rng.children.size() == 1);
    const Node& blob = sim.rng.children[0];
    CHECK(blob.kind == Kind::Raw && blob.raw.size() == 2503);
    mars::rng::MT19937 saved(1u);
    CHECK(saved.load_state(blob.raw.data(), blob.raw.size()));
    std::printf("   rng: left=%d (index %d), RSeed=%d\n", saved.left(), saved.index(), doc.game.createParams.rseed);
    CHECK(saved.left() >= 0 && saved.left() <= mars::rng::MT19937::N);
    {
        // the saved block must be reachable from seed(RSeed) by whole twists
        mars::rng::MT19937 gen(uint32_t(doc.game.createParams.rseed));
        int twists = -1;
        for (int k = 0; k < 16 && twists < 0; ++k) {
            if (std::memcmp(gen.state(), saved.state(), sizeof(uint32_t) * mars::rng::MT19937::N) == 0) twists = k;
            else
                for (int i = 0; i < mars::rng::MT19937::N; ++i) gen.next_u32();  // consume a block -> next twist
        }
        std::printf("   rng: state == seed(RSeed) after %d twist(s)\n", twists);
        CHECK(twists >= 0);
    }

    // --- coverage: what the shapes understand vs what a Node merely carries ---------
    // The round trip below is byte-identical either way, because an opaque Node is
    // copied verbatim.  This is the number that actually moves when a body gets typed.
    {
        CoverageArchive cov;
        doc.game.io(cov);
        size_t total = cov.typed + cov.opaque;
        double pct = total ? 100.0 * double(cov.typed) / double(total) : 0.0;
        std::printf("   coverage: %zu typed, %zu opaque (%.1f%% of %u stream items typed)\n", cov.typed,
                    cov.opaque, pct, doc.stats.items);
        std::vector<std::pair<size_t, std::string>> worst;
        for (const auto& kv : cov.opaque_by_tag) worst.emplace_back(kv.second, kv.first);
        std::sort(worst.rbegin(), worst.rend());
        std::printf("   still opaque:");
        for (size_t i = 0; i < worst.size() && i < 8; ++i)
            std::printf(" %s=%zu", worst[i].second.c_str(), worst[i].first);
        std::printf("\n");
        // Ratchet, not a target: typing a body must never silently regress.
        // With TurnCommands_v5 typed, the only items left carried as Nodes are the
        // two of the MT19937 block, which is deliberately opaque.
        CHECK(pct >= 99.99);
    }

    // --- round trips ------------------------------------------------------------------
    Bytes tree_bytes = write_tree(doc.tree);
    CHECK(tree_bytes == doc.inflated);
    Bytes typed_bytes = write_save(doc.game);
    if (typed_bytes != doc.inflated) {
        size_t i = 0, n = std::min(typed_bytes.size(), doc.inflated.size());
        while (i < n && typed_bytes[i] == doc.inflated[i]) ++i;
        std::printf("   typed round trip differs at 0x%zx (sizes %zu vs %zu)\n", i, typed_bytes.size(),
                    doc.inflated.size());
    }
    CHECK(typed_bytes == doc.inflated);
    std::printf("   round trip: tree %s, typed %s\n", tree_bytes == doc.inflated ? "identical" : "DIFFERS",
                typed_bytes == doc.inflated ? "identical" : "DIFFERS");

    // --- optional dump for the oracle comparison -------------------------------------
    if (dump_dir) {
        std::string base = path.substr(path.find_last_of('/') + 1);
        std::ofstream d(std::string(dump_dir) + "/" + base + ".cpp.dump");
        for (const std::string& l : dump_tree(doc.tree)) d << l << '\n';
        std::ofstream sm(std::string(dump_dir) + "/" + base + ".cpp.summary");
        sm << "summary: game=" << json_quote_cp1252(s.gameName) << " turn=" << s.turn << " numSys=" << s.numSys
           << " players=" << s.players.size() << '\n';
        sm << "sim: players=" << sim.players.size() << " systems=" << sim.systems.size()
           << " fleets=" << sim.fleets.size() << '\n';
    }
}

int main() {
    const char* dir = std::getenv("SOTS_SAVES_DIR");
    if (!dir || !*dir) {
        std::printf("test_save: SKIPPED (SOTS_SAVES_DIR not set)\n");
        return 0;
    }
    std::vector<std::string> saves = list_saves(dir);
    if (saves.empty()) {
        std::printf("test_save: SKIPPED (no *.sav in %s)\n", dir);
        return 0;
    }
    const char* dump_dir = std::getenv("SOTS_DUMP_DIR");
    for (const std::string& p : saves) {
        try {
            check_save(p, dump_dir && *dump_dir ? dump_dir : nullptr);
        } catch (const std::exception& e) {
            std::printf("FAIL %s: %s\n", p.c_str(), e.what());
            ++fails;
        }
    }
    std::printf("test_save: %s (%zu save(s), %d failures)\n", fails ? "FAILED" : "ok", saves.size(), fails);
    return fails ? 1 : 0;
}