mirror of
https://github.com/Abdess/retroarch_system.git
synced 2026-10-21 11:19:11 -05:00
Abdessamad Derraz
eba5f1dd62
A changed element was rendered again, losing its alignment and the place of its attributes, and an element whose md5 list was only reordered (neogeo.zip, the same 24 digests) was rewritten too: Recalbox accepts any digest of the list (Bios.cpp:109-130), so the order is not content. A changed attribute is now replaced where it stands, one dropped goes with its space, one added goes before />, and each system's opening tag keeps its own indentation. An empty truth gives back the cached es_bios.xml byte for byte.
347 lines
14 KiB
Python
347 lines
14 KiB
Python
"""Exporter for Recalbox's es_bios.xml.
|
|
|
|
The file is validated by es_bios.xsd, which makes path, md5 and core
|
|
required on every bios element. An entry we cannot give all three to is not
|
|
written: Recalbox would reject the file whole.
|
|
|
|
mandatory and hashMatchMandatory are separate axes. Recalbox reads a missing
|
|
attribute as true for both, so each is written only when it is false, or
|
|
when the platform stated it explicitly.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import html
|
|
import re
|
|
import sys
|
|
from pathlib import Path
|
|
from xml.etree.ElementTree import Comment, ParseError
|
|
from xml.sax.saxutils import quoteattr
|
|
|
|
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
|
|
|
|
from common import parse_untrusted_xml
|
|
|
|
from .base_exporter import BaseExporter
|
|
from .baseline import NativeFile, NativeSystem, Report
|
|
|
|
SOURCE_URL = (
|
|
"https://gitlab.com/recalbox/recalbox/-/raw/master/board/recalbox/fsoverlay"
|
|
"/recalbox/share_init/system/.emulationstation/es_bios.xml"
|
|
)
|
|
SCHEMA_URL = (
|
|
"https://gitlab.com/recalbox/recalbox/-/raw/master/board/recalbox/fsoverlay"
|
|
"/recalbox/share_init/system/.emulationstation/es_bios.xsd"
|
|
)
|
|
|
|
|
|
_ATTRIBUTE = re.compile(r'([A-Za-z][A-Za-z0-9]*)\s*=\s*"([^"]*)"')
|
|
_SYSTEM_BLOCK = re.compile(r"<system\b.*?</system>", re.S)
|
|
_BIOS_ELEMENT = re.compile(r"[ \t]*<bios\b.*?/>", re.S)
|
|
|
|
|
|
def _stated(text: str) -> dict[str, str]:
|
|
"""What an element says, without its layout: attribute order, spacing,
|
|
line breaks inside a value and the case of a digest are not content."""
|
|
attrs = {k: " ".join(html.unescape(v).split()) for k, v in _ATTRIBUTE.findall(text)}
|
|
if "md5" in attrs:
|
|
# Recalbox accepts any digest of the list (Bios.cpp:109-130): the
|
|
# order is not content, and neogeo.zip's 24, reordered, were
|
|
# written again.
|
|
attrs["md5"] = ",".join(sorted({h.strip().lower() for h in attrs["md5"].split(",")}))
|
|
return attrs
|
|
|
|
|
|
_SPACED_ATTRIBUTE = re.compile(r'(\s+)([A-Za-z][A-Za-z0-9]*)(\s*=\s*")([^"]*)(")')
|
|
|
|
|
|
def patch_element(kept: str, mine: str) -> str:
|
|
"""Recalbox's element with only the attributes that changed rewritten.
|
|
|
|
Rendered again, a corrected element lost its alignment and the place
|
|
of its attributes. A changed value is replaced where it stands, an
|
|
attribute ours drops goes with the space before it, and one ours adds
|
|
goes before the closing />.
|
|
"""
|
|
target = dict(_ATTRIBUTE.findall(mine))
|
|
target_stated = _stated(mine)
|
|
present: set[str] = set()
|
|
|
|
def rewrite(match: re.Match[str]) -> str:
|
|
name = match.group(2)
|
|
present.add(name)
|
|
if name not in target:
|
|
return ""
|
|
if _stated(f'<x {name}="{match.group(4)}"/>') == {name: target_stated[name]}:
|
|
return match.group(0)
|
|
return match.group(1) + name + match.group(3) + target[name] + match.group(5)
|
|
|
|
patched = _SPACED_ATTRIBUTE.sub(rewrite, kept)
|
|
added = "".join(f' {name}="{value}"' for name, value in target.items() if name not in present)
|
|
if added:
|
|
close = re.search(r"\s*/>\s*$", patched)
|
|
if close is None:
|
|
return mine
|
|
patched = patched[:close.start()] + added + patched[close.start():]
|
|
return patched
|
|
|
|
|
|
def _original_openings(original: str) -> list[str]:
|
|
"""Each system block's opening tag as written, its indentation included:
|
|
cdi's is indented four spaces where the others take two (l.604)."""
|
|
openings = []
|
|
for block in _SYSTEM_BLOCK.finditer(original):
|
|
line_start = original.rfind("\n", 0, block.start()) + 1
|
|
openings.append(original[line_start:block.start()] + block.group(0).split(">", 1)[0] + ">")
|
|
return openings
|
|
|
|
|
|
def _original_elements(original: str) -> list[dict[str, str]]:
|
|
"""Each system block's <bios> elements as written, by path, in order."""
|
|
blocks = []
|
|
for block in _SYSTEM_BLOCK.finditer(original):
|
|
elements: dict[str, str] = {}
|
|
for match in _BIOS_ELEMENT.finditer(block.group(0)):
|
|
raw = match.group(0)
|
|
elements.setdefault(_stated(raw).get("path", ""), raw)
|
|
blocks.append(elements)
|
|
return blocks
|
|
|
|
|
|
class Exporter(BaseExporter):
|
|
"""Write Recalbox's es_bios.xml, corrected."""
|
|
|
|
@staticmethod
|
|
def platform_name() -> str:
|
|
return "recalbox"
|
|
|
|
@staticmethod
|
|
def native_filename() -> str:
|
|
return "es_bios.xml"
|
|
|
|
@staticmethod
|
|
def keys_by_system() -> bool:
|
|
return True
|
|
|
|
@staticmethod
|
|
def carries() -> frozenset[str]:
|
|
return frozenset({"md5", "required"})
|
|
|
|
@staticmethod
|
|
def native_sources() -> dict[str, str]:
|
|
return {"es_bios.xml": SOURCE_URL, "es_bios.xsd": SCHEMA_URL}
|
|
|
|
def _path(self, fe: NativeFile) -> str:
|
|
"""The path Recalbox reads, pipe-joined when it accepts several.
|
|
|
|
A path Recalbox already states is reproduced exactly: several of its
|
|
entries sit at the BIOS root with no directory at all, and prefixing
|
|
them with the system would point the frontend somewhere else. An
|
|
addition goes where the core reads it, relative to the BIOS directory
|
|
Recalbox gives its cores: DuckStation reads scph5000.bin at the root,
|
|
not under psx/.
|
|
"""
|
|
if fe.platform is not None:
|
|
path = str(fe.platform.get("destination") or fe.name)
|
|
alternatives = fe.platform.get("alt_paths") or []
|
|
if alternatives:
|
|
return "|".join([path, *[str(a) for a in alternatives]])
|
|
return path
|
|
return fe.destination or fe.name
|
|
|
|
def _bios_element(self, fe: NativeFile, native_id: str) -> str:
|
|
attrs = [f"path={quoteattr(self._path(fe))}"]
|
|
attrs.append(f'md5={quoteattr(",".join(fe.hashes("md5")))}')
|
|
attrs.append(f'core={quoteattr(",".join(fe.cores()))}')
|
|
|
|
if not fe.required:
|
|
attrs.append('mandatory="false"')
|
|
elif fe.native("mandatory_declared", None) is True:
|
|
attrs.append('mandatory="true"')
|
|
|
|
hash_match = fe.native("hash_match_mandatory", None)
|
|
if hash_match is False:
|
|
attrs.append('hashMatchMandatory="false"')
|
|
elif hash_match is True:
|
|
attrs.append('hashMatchMandatory="true"')
|
|
|
|
# The platform's own note only: ours is prose about our profiles, with
|
|
# file:line citations, and the format does not carry note corrections.
|
|
note = " ".join(str((fe.platform or {}).get("note") or "").split())
|
|
if note:
|
|
attrs.append(f"note={quoteattr(note)}")
|
|
|
|
return f" <bios {' '.join(attrs)} />"
|
|
|
|
def states(self, fe: NativeFile, field_name: str) -> bool:
|
|
"""An element it cannot rewrite is written back as Recalbox wrote it.
|
|
|
|
es_bios.xsd makes md5 required: an entry whose md5 the truth
|
|
contradicts without naming another keeps its element whole, and
|
|
counting the md5 cleared announced a change absent from the file.
|
|
"""
|
|
if fe.platform is not None and not self.writable(fe):
|
|
return False
|
|
return super().states(fe, field_name)
|
|
|
|
@classmethod
|
|
def writable(cls, fe: NativeFile, require: str = "") -> bool:
|
|
"""es_bios.xsd makes md5 and core required: without them, no element.
|
|
|
|
An entry Recalbox already ships has both, so this only ever gates
|
|
what we would be adding.
|
|
"""
|
|
if fe.platform is None and fe.foreign_system:
|
|
return False
|
|
return bool(fe.hashes("md5")) and bool(fe.cores())
|
|
|
|
def render(
|
|
self,
|
|
systems: dict[str, NativeSystem],
|
|
report: Report,
|
|
originals: dict[str, str],
|
|
scraped: dict | None = None,
|
|
) -> dict[str, str]:
|
|
lines = [
|
|
'<?xml version="1.0" encoding="UTF-8"?>',
|
|
'<biosList xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"'
|
|
' xsi:noNamespaceSchemaLocation="es_bios.xsd">',
|
|
]
|
|
# Recalbox's own file is the base, so an element our transcription
|
|
# does not carry (added or renamed after a scrape) stays as written,
|
|
# and its system order with it. A platform can hold several blocks
|
|
# (Pentagon and Scorpion are both zxspectrum), and an entry the
|
|
# maintainer commented out stays commented out.
|
|
theirs: list[tuple[str, str, list[dict[str, str] | str]]] = []
|
|
original = originals.get(self.native_filename(), "")
|
|
if original:
|
|
try:
|
|
root = parse_untrusted_xml(original, self.native_filename(), comments=True)
|
|
except (ParseError, ValueError) as exc:
|
|
# Rebuilt from our model alone, the export would drop every
|
|
# system we do not transcribe and the maintainer's comments.
|
|
raise ValueError(
|
|
f"Recalbox's {self.native_filename()} does not parse ({exc}); "
|
|
"fetch it again with --refresh-cache"
|
|
) from exc
|
|
if root is not None:
|
|
theirs = [
|
|
(s.get("platform", ""), s.get("fullname", ""), [
|
|
f"<!--{child.text}-->" if child.tag is Comment else dict(child.attrib)
|
|
for child in s
|
|
if child.tag is Comment or child.tag == "bios"
|
|
])
|
|
for s in root.iter("system")
|
|
]
|
|
|
|
ours: dict[str, tuple[NativeSystem, dict[str, tuple[NativeFile, str]]]] = {}
|
|
for system in systems.values():
|
|
writable = [fe for fe in system.files if self.writable(fe)]
|
|
if writable:
|
|
ours[system.native_id] = (system, {
|
|
self._path(fe): (fe, self._bios_element(fe, system.native_id))
|
|
for fe in writable
|
|
})
|
|
|
|
def element(attrs: dict[str, str]) -> str:
|
|
rendered = " ".join(f"{k}={quoteattr(v)}" for k, v in attrs.items())
|
|
return f" <bios {rendered} />"
|
|
|
|
def emit(platform: str, fullname: str, body: list[str], opening: str = "") -> None:
|
|
stated = _stated(opening) if opening else {}
|
|
if stated.get("platform") == platform and stated.get("fullname") == fullname:
|
|
lines.append(opening)
|
|
else:
|
|
lines.append(
|
|
f" <system fullname={quoteattr(fullname)} platform={quoteattr(platform)}>"
|
|
)
|
|
lines.extend(body)
|
|
lines.append(" </system>")
|
|
|
|
# Each block of theirs corrects the entries it holds. An element that
|
|
# states what Recalbox wrote keeps Recalbox's text: re-rendered, 222
|
|
# unchanged elements made 180 lines of diff the report never named.
|
|
written = _original_elements(original) if original else []
|
|
openings = _original_openings(original) if original else []
|
|
blocks: list[tuple[str, str, list[str]]] = []
|
|
for index, (platform, fullname, children) in enumerate(theirs):
|
|
elements = ours.get(platform, (None, {}))[1]
|
|
raw = written[index] if index < len(written) else {}
|
|
body: list[str] = []
|
|
for child in children:
|
|
if isinstance(child, str):
|
|
body.append(f" {child}")
|
|
continue
|
|
path = child.get("path", "")
|
|
mine = elements.pop(path, None)
|
|
kept = raw.get(" ".join(path.split()))
|
|
if kept is not None and (mine is None or _stated(mine[1]) == _stated(kept)):
|
|
body.append(kept)
|
|
elif kept is not None and mine is not None and _stated(
|
|
patched := patch_element(kept, mine[1])
|
|
) == _stated(mine[1]):
|
|
body.append(patched)
|
|
else:
|
|
body.append(mine[1] if mine else element(child))
|
|
blocks.append((platform, fullname, body))
|
|
|
|
# An addition joins the block named like its file, else the first
|
|
# block of its platform.
|
|
first: dict[str, int] = {}
|
|
named: dict[tuple[str, str], int] = {}
|
|
for index, (platform, fullname, _) in enumerate(blocks):
|
|
first.setdefault(platform, index)
|
|
named.setdefault((platform, fullname), index)
|
|
for platform in [p for p in ours if p in first]:
|
|
_, elements = ours.pop(platform)
|
|
for fe, rendered in elements.values():
|
|
key = (platform, str(fe.native("native_name", "")))
|
|
blocks[named.get(key, first[platform])][2].append(rendered)
|
|
|
|
for index, (platform, fullname, body) in enumerate(blocks):
|
|
emit(platform, fullname, body, openings[index] if index < len(openings) else "")
|
|
for platform in sorted(ours):
|
|
system, elements = ours[platform]
|
|
emit(platform, self.display_name(system), [r for _, r in elements.values()])
|
|
|
|
lines.append("</biosList>")
|
|
lines.append("")
|
|
return {self.native_filename(): "\n".join(lines)}
|
|
|
|
def validate(
|
|
self,
|
|
systems: dict[str, NativeSystem],
|
|
produced: dict[str, str],
|
|
) -> list[str]:
|
|
content = produced[self.native_filename()]
|
|
issues: list[str] = []
|
|
try:
|
|
root = parse_untrusted_xml(content, self.native_filename())
|
|
except (ParseError, ValueError) as exc:
|
|
return [f"the XML does not parse: {exc}"]
|
|
|
|
for element in root.iter("bios"):
|
|
for attribute in ("path", "md5", "core"):
|
|
if not element.get(attribute):
|
|
issues.append(
|
|
f"es_bios.xsd requires {attribute}: "
|
|
f"{element.get('path', '?')}"
|
|
)
|
|
for element in root.iter("system"):
|
|
if not list(element):
|
|
issues.append(f"empty system: {element.get('platform', '?')}")
|
|
for attribute in ("fullname", "platform"):
|
|
if not element.get(attribute):
|
|
issues.append(f"es_bios.xsd requires {attribute} on system")
|
|
|
|
exported = {
|
|
element.get("path", "").casefold() for element in root.iter("bios")
|
|
}
|
|
for system in systems.values():
|
|
for fe in system.files:
|
|
if not self.writable(fe):
|
|
continue
|
|
if self._path(fe).casefold() not in exported:
|
|
issues.append(f"absent: {system.native_id}/{fe.name}")
|
|
return issues
|