lib/llvm-ast.ts

// Copyright (c) 2021, Compiler Explorer Authors
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
//
//     * Redistributions of source code must retain the above copyright notice,
//       this list of conditions and the following disclaimer.
//     * Redistributions in binary form must reproduce the above copyright
//       notice, this list of conditions and the following disclaimer in the
//       documentation and/or other materials provided with the distribution.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
// POSSIBILITY OF SUCH DAMAGE.

import type {CompilationResult} from '../types/compilation/compilation.interfaces.js';
import type {ResultLine} from '../types/resultline/resultline.interfaces.js';

import type {PropertyGetter} from './properties.interfaces.js';

type Point = {
    line: number | null;
    col: number | null;
};

export class LlvmAstParser {
    maxAstLines: number;

    // Almost every line of AST includes a span of related source lines:
    // In different forms like <line:a:b, line:c:d>
    static readonly locTypes = {
        NONE: 'none', // No location specified
        POINT: 'point', // A single location: beginning of a token
        SPAN: 'span', // Two locations: first token to last token (beginning)
    } as const;

    constructor(compilerProps: PropertyGetter) {
        this.maxAstLines = 500;
        if (compilerProps) {
            this.maxAstLines = compilerProps('maxLinesOfAst', this.maxAstLines);
        }
    }

    // Accepts "line:a:b" and "col:b"
    parsePoint(ptLine: string, lastLineNo: number | null): Point {
        const lineRegex = /line:(\d+):/;
        const colRegex = /(?:col|\d):(\d+)(?::|$)/;
        const lineMatch = ptLine.match(lineRegex);
        const colMatch = ptLine.match(colRegex);
        const line = lineMatch ? Number(lineMatch[1]) : lastLineNo;
        const col = colMatch ? Number(colMatch[1]) : null; // Does not happen for well-formed strings
        return {line, col};
    }

    // Accepts "<X, X>" and "<X>", where
    // X can be "col:a" or "line:a:b"
    // lastLineNo - the line number of the previous node,
    // reused when only a column specified.
    parseSpan(
        line: string,
        lastLineNo: number | null,
    ):
        | {type: typeof LlvmAstParser.locTypes.SPAN; begin: Point; end: Point}
        | {type: typeof LlvmAstParser.locTypes.POINT; loc: Point}
        | {type: typeof LlvmAstParser.locTypes.NONE} {
        const spanRegex = /<((?:line|col)[\d ,:ceilno]+)>/;
        const m = line.match(spanRegex);
        if (m) {
            const span = m[1];
            const beginEnd = span.split(',');
            if (beginEnd.length === 2) {
                const begin = this.parsePoint(beginEnd[0], lastLineNo);
                const end = this.parsePoint(beginEnd[1], begin.line);
                return {type: LlvmAstParser.locTypes.SPAN, begin, end};
            } else {
                return {type: LlvmAstParser.locTypes.POINT, loc: this.parsePoint(span, lastLineNo)};
            }
        }
        return {type: LlvmAstParser.locTypes.NONE};
    }

    // Link the AST lines with spans of source locations (lines+columns)
    parseAndSetSourceLines(astDump: ResultLine[]) {
        let lfrom: any = {line: null, loc: null},
            lto: any = {line: null, loc: null};
        for (const line of astDump) {
            const span = this.parseSpan(line.text, lfrom.line);
            switch (span.type) {
                case LlvmAstParser.locTypes.NONE: {
                    break;
                }
                case LlvmAstParser.locTypes.POINT: {
                    lfrom = span.loc;
                    lto = span.loc;
                    break;
                }
                case LlvmAstParser.locTypes.SPAN: {
                    lfrom = span.begin;
                    lto = span.end;
                    break;
                }
            }
            if (span.type !== LlvmAstParser.locTypes.NONE) {
                // TODO: ResultLineSource doesn't have to/from
                (line.source as any) = {from: lfrom, to: lto};
            }
        }
    }

    processAst(result: CompilationResult) {
        const output = result.stdout;

        // Top level decls start with |- or `-
        const topLevelRegex = /^([`|])-/;

        // Refers to the user's source file rather than a system header
        const sourceRegex = /<source>/g;

        const slocRegex = /<<invalid sloc>>/;

        // <<invalid sloc>, /app/hell.hpp:5:1>
        const userSource = /<<invalid sloc>, \/app\/.*:\d+:\d+>/;

        // </usr/include/x86_64-linux-gnu/bits/types.h:31:1, col:23>
        // <line:229:1, /usr/include/x86_64-linux-gnu/sys/cdefs.h:61:27> /usr/include/time.h:229:12
        // </usr/include/x86_64-linux-gnu/sys/cdefs.h:293:44, /usr/include/time.h:258:27> 1
        // </opt/compiler-explorer/gcc-11.2.0/lib/gcc/x86_64-linux-gnu/11.2.0/etc...
        const systemSource = /(<\/usr\/|,\s\/usr\/|<\/opt\/)/;

        // Refers to whatever the most recent file specified was
        const lineRegex = /<(col|line):/;

        let mostRecentIsSource = false;

        // Remove all AST nodes which aren't directly from the user's source code
        for (let i = 0; i < output.length; ++i) {
            if (topLevelRegex.test(output[i].text)) {
                if (lineRegex.test(output[i].text) && mostRecentIsSource) {
                    // do nothing
                } else if (sourceRegex.test(output[i].text)) {
                    mostRecentIsSource = true;
                } else {
                    // This is a system header or implicit definition,
                    // remove everything up to the next top level decl
                    // Top level decls with invalid sloc as the file don't change the most recent file
                    if (systemSource.test(output[i].text)) {
                        // skip ast from this source
                    } else if (userSource.test(output[i].text)) {
                        continue;
                    } else if (!slocRegex.test(output[i].text)) {
                        mostRecentIsSource = false;
                        continue;
                    }

                    let spliceMax = i + 1;
                    while (output[spliceMax] && !topLevelRegex.test(output[spliceMax].text)) {
                        spliceMax++;
                    }
                    output.splice(i, spliceMax - i);
                    --i;
                }
            }
            // Filter out the symbol addresses
            const addressRegex = /^([^A-Za-z]*[A-Za-z]+) 0x[\da-z]+/gm;
            output[i].text = output[i].text.replace(addressRegex, '$1');

            // Filter out <invalid sloc> and <<invalid sloc>>
            const slocRegex2 = / ?<?<invalid sloc>>?/g;
            output[i].text = output[i].text.replace(slocRegex2, '');

            // Unify file references
            output[i].text = output[i].text.replace(sourceRegex, 'line');
        }
        this.parseAndSetSourceLines(output);
        return output;
    }
}