src/tint/reader/spirv/namer.cc

// Copyright 2020 The Tint Authors.
//
// 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 "src/tint/reader/spirv/namer.h"

#include <algorithm>
#include <sstream>
#include <unordered_set>

#include "src/tint/debug.h"

namespace tint::reader::spirv {

namespace {

const char* kWGSLReservedWords[] = {
    // Please keep this list sorted
    "array",      "as",          "asm",
    "bf16",       "binding",     "block",
    "bool",       "break",       "builtin",
    "case",       "cast",        "compute",
    "const",      "continue",    "default",
    "discard",    "do",          "else",
    "elseif",     "entry_point", "enum",
    "f16",        "f32",         "fallthrough",
    "false",      "fn",          "for",
    "fragment",   "i16",         "i32",
    "i64",        "i8",          "if",
    "image",      "import",      "in",
    "let",        "location",    "loop",
    "mat2x2",     "mat2x3",      "mat2x4",
    "mat3x2",     "mat3x3",      "mat3x4",
    "mat4x2",     "mat4x3",      "mat4x4",
    "offset",     "out",         "override",
    "premerge",   "private",     "ptr",
    "regardless", "return",      "set",
    "storage",    "struct",      "switch",
    "true",       "type",        "typedef",
    "u16",        "u32",         "u64",
    "u8",         "uniform",     "uniform_constant",
    "unless",     "using",       "var",
    "vec2",       "vec3",        "vec4",
    "vertex",     "void",        "while",
    "workgroup",
};

}  // namespace

Namer::Namer(const FailStream& fail_stream) : fail_stream_(fail_stream) {
    for (const auto* reserved : kWGSLReservedWords) {
        name_to_id_[std::string(reserved)] = 0;
    }
}

Namer::~Namer() = default;

std::string Namer::Sanitize(const std::string& suggested_name) {
    if (suggested_name.empty()) {
        return "empty";
    }
    // Otherwise, replace invalid characters by '_'.
    std::string result;
    std::string invalid_as_first_char = "_0123456789";
    std::string valid =
        "abcdefghijklmnopqrstuvwxyz"
        "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
        "_0123456789";
    // If the first character is invalid for starting a WGSL identifier, then
    // prefix the result with "x".
    if ((std::string::npos != invalid_as_first_char.find(suggested_name[0])) ||
        (std::string::npos == valid.find(suggested_name[0]))) {
        result = "x";
    }
    std::transform(
        suggested_name.begin(), suggested_name.end(), std::back_inserter(result),
        [&valid](const char c) { return (std::string::npos == valid.find(c)) ? '_' : c; });
    return result;
}

std::string Namer::GetMemberName(uint32_t struct_id, uint32_t member_index) const {
    std::string result;
    auto where = struct_member_names_.find(struct_id);
    if (where != struct_member_names_.end()) {
        auto& member_names = where->second;
        if (member_index < member_names.size()) {
            result = member_names[member_index];
        }
    }
    return result;
}

std::string Namer::FindUnusedDerivedName(const std::string& base_name) {
    // Ensure uniqueness among names.
    std::string derived_name;
    uint32_t& i = next_unusued_derived_name_id_[base_name];
    while (i != 0xffffffff) {
        std::stringstream new_name_stream;
        new_name_stream << base_name;
        if (i > 0) {
            new_name_stream << "_" << i;
        }
        derived_name = new_name_stream.str();
        if (!IsRegistered(derived_name)) {
            return derived_name;
        }
        i++;
    }
    TINT_ASSERT(Reader, false /* FindUnusedDerivedName() overflowed u32 */);
    return "<u32 overflow>";
}

std::string Namer::MakeDerivedName(const std::string& base_name) {
    auto result = FindUnusedDerivedName(base_name);
    const bool registered = RegisterWithoutId(result);
    TINT_ASSERT(Reader, registered);
    return result;
}

bool Namer::Register(uint32_t id, const std::string& name) {
    if (HasName(id)) {
        return Fail() << "internal error: ID " << id
                      << " already has registered name: " << id_to_name_[id];
    }
    if (!RegisterWithoutId(name)) {
        return false;
    }
    id_to_name_[id] = name;
    name_to_id_[name] = id;
    return true;
}

bool Namer::RegisterWithoutId(const std::string& name) {
    if (IsRegistered(name)) {
        return Fail() << "internal error: name already registered: " << name;
    }
    name_to_id_[name] = 0;
    return true;
}

bool Namer::SuggestSanitizedName(uint32_t id, const std::string& suggested_name) {
    if (HasName(id)) {
        return false;
    }

    return Register(id, FindUnusedDerivedName(Sanitize(suggested_name)));
}

bool Namer::SuggestSanitizedMemberName(uint32_t struct_id,
                                       uint32_t member_index,
                                       const std::string& suggested_name) {
    // Creates an empty vector the first time we visit this struct.
    auto& name_vector = struct_member_names_[struct_id];
    // Resizing will set new entries to the empty string.
    name_vector.resize(std::max(name_vector.size(), size_t(member_index + 1)));
    auto& entry = name_vector[member_index];
    if (entry.empty()) {
        entry = Sanitize(suggested_name);
        return true;
    }
    return false;
}

void Namer::ResolveMemberNamesForStruct(uint32_t struct_id, uint32_t num_members) {
    auto& name_vector = struct_member_names_[struct_id];
    // Resizing will set new entries to the empty string.
    // It would have been an error if the client had registered a name for
    // an out-of-bounds member index, so toss those away.
    name_vector.resize(num_members);

    std::unordered_set<std::string> used_names;

    // Returns a name, based on the suggestion, which does not equal
    // any name in the used_names set.
    auto disambiguate_name = [&used_names](const std::string& suggestion) -> std::string {
        if (used_names.find(suggestion) == used_names.end()) {
            // There is no collision.
            return suggestion;
        }

        uint32_t i = 1;
        std::string new_name;
        do {
            std::stringstream new_name_stream;
            new_name_stream << suggestion << "_" << i;
            new_name = new_name_stream.str();
            ++i;
        } while (used_names.find(new_name) != used_names.end());
        return new_name;
    };

    // First ensure uniqueness among names for which we have already taken
    // suggestions.
    for (auto& name : name_vector) {
        if (!name.empty()) {
            // This modifies the names in-place, i.e. update the name_vector
            // entries.
            name = disambiguate_name(name);
            used_names.insert(name);
        }
    }

    // Now ensure uniqueness among the rest.  Doing this in a second pass
    // allows us to preserve suggestions as much as possible.  Otherwise
    // a generated name such as 'field1' might collide with a user-suggested
    // name of 'field1' attached to a later member.
    uint32_t index = 0;
    for (auto& name : name_vector) {
        if (name.empty()) {
            std::stringstream suggestion;
            suggestion << "field" << index;
            // Again, modify the name-vector in-place.
            name = disambiguate_name(suggestion.str());
            used_names.insert(name);
        }
        index++;
    }
}

}  // namespace tint::reader::spirv