nixpkgs_hammer.py

import json
import os
import subprocess
import sys
from typing import Any, Dict, List

from .nix import Attr

# https://github.com/rmcgibbo/nixpkgs-review-bot/issues/26
IGNORE_SUGGESTIONS_FOR_FILES = {
    "pkgs/misc/vim-plugins/generated.nix",
    "pkgs/misc/vim-plugins/overrides.nix",
    "pkgs/misc/tmux-plugins/default.nix",
}

# https://github.com/jtojnar/nixpkgs-hammering/issues/73
# https://github.com/jtojnar/nixpkgs-hammering/issues/77#issuecomment-786193493
# https://github.com/jtojnar/nixpkgs-hammering/pull/78#pullrequestreview-599072677
ATTRS_THAT_BREAK_NIXPKGS_HAMMER = {
    "acl",
    "attr",
    "bash",
    "binutils-unwrapped",
    "bzip2",
    "coreutils",
    "coreutils-full",
    "coreutils-prefixed",
    "datadog-agent",
    "diffutils",
    "findutils",
    "gawkInteractive",
    "gcc-unwrapped",
    "glibc",
    "gnugrep",
    "gnupatch",
    "gnused",
    "gnutar",
    "gzip",
    "holochain-go",
    "javaPackages.junit_4_12",
    "javaPackages.mavenHello_1_0",
    "javaPackages.mavenHello_1_1",
    "libgccjit",
    "zfsbackup",
}


def nixpkgs_hammer(attrs: List[Attr], modified_files: List[str] = None) -> List[Attr]:
    """Run nixpkgs-hammer on each attr, saves the results into the
    'check_report' field on the attr.

    Only save checks if the 'location' of the bug flagged by the check is within
    the set of files modified by the pr, which should be passed in `modified_files`.
    """

    cmd = [
        "nixpkgs-hammer",
        "-f",
        ".",
        "--json",
        "--exclude",
        "attribute-ordering",
        "--exclude",
        "explicit-phases",
        # Discussion here: https://github.com/jtojnar/nixpkgs-hammering/pull/38#issuecomment-778381992
        "--exclude",
        "attribute-typo",
    ] + [
        a.name
        for a in attrs
        if (a.name not in ATTRS_THAT_BREAK_NIXPKGS_HAMMER)
        and (not a.skipped)
        and (not a.blacklisted)
        and (not a.broken)
    ]

    print("$" + " ".join(cmd), file=sys.stderr)

    try:
        proc = subprocess.run(
            cmd,
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
            check=True,
            text=True,
        )
    except subprocess.CalledProcessError as e:
        print(os.path.abspath(os.getcwd()), file=sys.stderr)
        print(e.stderr, file=sys.stderr)
        print(e.stderr, file=sys.stderr)
        print(e.returncode, file=sys.stderr)
        raise
    hammer_report = json.loads(proc.stdout)

    for name, data in hammer_report.items():
        for msg in data:
            for location in msg.get("locations", []):
                if "file" in location and isinstance(location["file"], str):
                    location["file"] = os.path.relpath(location["file"], ".")

    attrs_by_name = {a.name: a for a in attrs}
    for name, data in hammer_report.items():
        for msg in (m for m in data if is_acceptable_hammer_message(m, modified_files)):
            attrs_by_name[name].check_report.append(stringify_message(**msg))

    return attrs


def stringify_location(file: str, line: int, column: int):
    with open(file, "r") as opened_file:
        all_lines = opened_file.read().splitlines()
        line_contents = all_lines[line - 1]
        line_spaces = " " * len(str(line))
        pointer = " " * (column - 1) + "^"

        location_lines = [
            "Near " + file + ":" + str(line) + ":" + str(column) + ":",
            "```",
            line_spaces + " |",
            str(line) + " | " + line_contents,
            line_spaces + " | " + pointer,
            "```",
            "",
        ]

    return "\n".join(location_lines)


def stringify_message(
    name: str,
    msg: str,
    locations: List[Dict[str, Any]] = [],
    cond: bool = True,
    link: bool = True,
    severity: str = "warning",
) -> str:
    if link:
        linked_name = f'<a href="https://github.com/jtojnar/nixpkgs-hammering/blob/master/explanations/{name}.md">{name}</a>'
    else:
        linked_name = name

    message_lines = [
        f"{severity}: {linked_name}",
        "",
        msg,
    ] + list(map(lambda loc: stringify_location(**loc), locations))

    return "\n".join(message_lines)


def is_acceptable_hammer_message(
    msg: Dict[str, Any], modified_files: List[str] = None
) -> bool:
    predicates = [
        lambda: is_modified_location(msg, modified_files=modified_files),
        lambda: not is_ignored_file(msg),
        lambda: not is_ignored_golang_buildFlagsArray_msg(msg),
        lambda: msg["name"] not in ("no-build-output",),
    ]
    return all(pred() for pred in predicates)


def is_modified_location(msg: Dict[str, Any], modified_files: List[str] = None) -> bool:
    if modified_files is None:
        return True

    for loc in msg.get("locations", []):
        if loc["file"] in set(modified_files):
            return True
    return False


def is_ignored_file(msg: Dict[str, Any]) -> bool:
    return any(
        loc["file"] in IGNORE_SUGGESTIONS_FOR_FILES for loc in msg.get("locations", [])
    )


def is_ignored_golang_buildFlagsArray_msg(msg: Dict[str, Any]) -> bool:
    for loc in msg["locations"]:
        is_golang = False
        with open(loc["file"]) as f:
            content = f.read()
            is_golang = "buildGoModule" in content or "buildGoPackage" in content

        if is_golang and msg["name"] == "no-flags-array":
            return True

    return False