Complete Yocto mirror with license table for TQMa6UL (2038-compliance)

- 264 license table entries with exact download URLs (224/264 resolved)
- Complete sources/ directory with all BitBake recipes
- Build configuration: tqma6ul-multi-mba6ulx, spaetzle (musl)
- Full traceability for Softwarefreigabeantrag
- GCC 13.4.0, Linux 6.6.102, U-Boot 2023.04, musl 1.2.4
- License distribution: GPL-2.0 (24), MIT (23), GPL-2.0+ (18), BSD-3 (16)
This commit is contained in:
Siggi (OpenClaw Agent)
2026-03-01 20:58:18 +00:00
commit 16accb6b24
15086 changed files with 1292356 additions and 0 deletions

View File

@@ -0,0 +1,883 @@
# pyshlex.py - PLY compatible lexer for pysh.
#
# Copyright 2007 Patrick Mezard
#
# This software may be used and distributed according to the terms
# of the GNU General Public License, incorporated herein by reference.
# TODO:
# - review all "char in 'abc'" snippets: the empty string can be matched
# - test line continuations within quoted/expansion strings
# - eof is buggy wrt sublexers
# - the lexer cannot really work in pull mode as it would be required to run
# PLY in pull mode. It was designed to work incrementally and it would not be
# that hard to enable pull mode.
import re
from ply import lex
from bb.pysh.sherrors import *
class NeedMore(Exception):
pass
def is_blank(c):
return c in (' ', '\t')
_RE_DIGITS = re.compile(r'^\d+$')
def are_digits(s):
return _RE_DIGITS.search(s) is not None
_OPERATORS = dict([
('&&', 'AND_IF'),
('||', 'OR_IF'),
(';;', 'DSEMI'),
('<<', 'DLESS'),
('>>', 'DGREAT'),
('<&', 'LESSAND'),
('>&', 'GREATAND'),
('<>', 'LESSGREAT'),
('<<-', 'DLESSDASH'),
('>|', 'CLOBBER'),
('&', 'AMP'),
(';', 'COMMA'),
('<', 'LESS'),
('>', 'GREATER'),
('(', 'LPARENS'),
(')', 'RPARENS'),
])
#Make a function to silence pychecker "Local variable shadows global"
def make_partial_ops():
partials = {}
for k in _OPERATORS:
for i in range(1, len(k)+1):
partials[k[:i]] = None
return partials
_PARTIAL_OPERATORS = make_partial_ops()
def is_partial_op(s):
"""Return True if s matches a non-empty subpart of an operator starting
at its first character.
"""
return s in _PARTIAL_OPERATORS
def is_op(s):
"""If s matches an operator, returns the operator identifier. Return None
otherwise.
"""
return _OPERATORS.get(s)
_RESERVEDS = dict([
('if', 'If'),
('then', 'Then'),
('else', 'Else'),
('elif', 'Elif'),
('fi', 'Fi'),
('do', 'Do'),
('done', 'Done'),
('case', 'Case'),
('esac', 'Esac'),
('while', 'While'),
('until', 'Until'),
('for', 'For'),
('{', 'Lbrace'),
('}', 'Rbrace'),
('!', 'Bang'),
('in', 'In'),
('|', 'PIPE'),
])
def get_reserved(s):
return _RESERVEDS.get(s)
_RE_NAME = re.compile(r'^[0-9a-zA-Z_]+$')
def is_name(s):
return _RE_NAME.search(s) is not None
def find_chars(seq, chars):
for i,v in enumerate(seq):
if v in chars:
return i,v
return -1, None
class WordLexer:
"""WordLexer parse quoted or expansion expressions and return an expression
tree. The input string can be any well formed sequence beginning with quoting
or expansion character. Embedded expressions are handled recursively. The
resulting tree is made of lists and strings. Lists represent quoted or
expansion expressions. Each list first element is the opening separator,
the last one the closing separator. In-between can be any number of strings
or lists for sub-expressions. Non quoted/expansion expression can written as
strings or as lists with empty strings as starting and ending delimiters.
"""
NAME_CHARSET = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_'
NAME_CHARSET = dict(zip(NAME_CHARSET, NAME_CHARSET))
SPECIAL_CHARSET = '@*#?-$!0'
#Characters which can be escaped depends on the current delimiters
ESCAPABLE = {
'`': set(['$', '\\', '`']),
'"': set(['$', '\\', '`', '"']),
"'": set(),
}
def __init__(self, heredoc = False):
# _buffer is the unprocessed input characters buffer
self._buffer = []
# _stack is empty or contains a quoted list being processed
# (this is the DFS path to the quoted expression being evaluated).
self._stack = []
self._escapable = None
# True when parsing unquoted here documents
self._heredoc = heredoc
def add(self, data, eof=False):
"""Feed the lexer with more data. If the quoted expression can be
delimited, return a tuple (expr, remaining) containing the expression
tree and the unconsumed data.
Otherwise, raise NeedMore.
"""
self._buffer += list(data)
self._parse(eof)
result = self._stack[0]
remaining = ''.join(self._buffer)
self._stack = []
self._buffer = []
return result, remaining
def _is_escapable(self, c, delim=None):
if delim is None:
if self._heredoc:
# Backslashes works as if they were double quoted in unquoted
# here-documents
delim = '"'
else:
if len(self._stack)<=1:
return True
delim = self._stack[-2][0]
escapables = self.ESCAPABLE.get(delim, None)
return escapables is None or c in escapables
def _parse_squote(self, buf, result, eof):
if not buf:
raise NeedMore()
try:
pos = buf.index("'")
except ValueError:
raise NeedMore()
result[-1] += ''.join(buf[:pos])
result += ["'"]
return pos+1, True
def _parse_bquote(self, buf, result, eof):
if not buf:
raise NeedMore()
if buf[0]=='\n':
#Remove line continuations
result[:] = ['', '', '']
elif self._is_escapable(buf[0]):
result[-1] += buf[0]
result += ['']
else:
#Keep as such
result[:] = ['', '\\'+buf[0], '']
return 1, True
def _parse_dquote(self, buf, result, eof):
if not buf:
raise NeedMore()
pos, sep = find_chars(buf, '$\\`"')
if pos==-1:
raise NeedMore()
result[-1] += ''.join(buf[:pos])
if sep=='"':
result += ['"']
return pos+1, True
else:
#Keep everything until the separator and defer processing
return pos, False
def _parse_command(self, buf, result, eof):
if not buf:
raise NeedMore()
chars = '$\\`"\''
if result[0] == '$(':
chars += ')'
pos, sep = find_chars(buf, chars)
if pos == -1:
raise NeedMore()
result[-1] += ''.join(buf[:pos])
if (result[0]=='$(' and sep==')') or (result[0]=='`' and sep=='`'):
result += [sep]
return pos+1, True
else:
return pos, False
def _parse_parameter(self, buf, result, eof):
if not buf:
raise NeedMore()
pos, sep = find_chars(buf, '$\\`"\'}')
if pos==-1:
raise NeedMore()
result[-1] += ''.join(buf[:pos])
if sep=='}':
result += [sep]
return pos+1, True
else:
return pos, False
def _parse_dollar(self, buf, result, eof):
sep = result[0]
if sep=='$':
if not buf:
#TODO: handle empty $
raise NeedMore()
if buf[0]=='(':
if len(buf)==1:
raise NeedMore()
if buf[1]=='(':
result[0] = '$(('
buf[:2] = []
else:
result[0] = '$('
buf[:1] = []
elif buf[0]=='{':
result[0] = '${'
buf[:1] = []
else:
if buf[0] in self.SPECIAL_CHARSET:
result[-1] = buf[0]
read = 1
else:
for read,c in enumerate(buf):
if c not in self.NAME_CHARSET:
break
else:
if not eof:
raise NeedMore()
read += 1
result[-1] += ''.join(buf[0:read])
if not result[-1]:
result[:] = ['', result[0], '']
else:
result += ['']
return read,True
sep = result[0]
if sep=='$(':
parsefunc = self._parse_command
elif sep=='${':
parsefunc = self._parse_parameter
else:
raise NotImplementedError(sep)
pos, closed = parsefunc(buf, result, eof)
return pos, closed
def _parse(self, eof):
buf = self._buffer
stack = self._stack
recurse = False
while 1:
if not stack or recurse:
if not buf:
raise NeedMore()
if buf[0] not in ('"\\`$\''):
raise ShellSyntaxError('Invalid quoted string sequence')
stack.append([buf[0], ''])
buf[:1] = []
recurse = False
result = stack[-1]
if result[0]=="'":
parsefunc = self._parse_squote
elif result[0]=='\\':
parsefunc = self._parse_bquote
elif result[0]=='"':
parsefunc = self._parse_dquote
elif result[0]=='`':
parsefunc = self._parse_command
elif result[0][0]=='$':
parsefunc = self._parse_dollar
else:
raise NotImplementedError()
read, closed = parsefunc(buf, result, eof)
buf[:read] = []
if closed:
if len(stack)>1:
#Merge in parent expression
parsed = stack.pop()
stack[-1] += [parsed]
stack[-1] += ['']
else:
break
else:
recurse = True
def normalize_wordtree(wtree):
"""Fold back every literal sequence (delimited with empty strings) into
parent sequence.
"""
def normalize(wtree):
result = []
for part in wtree[1:-1]:
if isinstance(part, list):
part = normalize(part)
if part[0]=='':
#Move the part content back at current level
result += part[1:-1]
continue
elif not part:
#Remove empty strings
continue
result.append(part)
if not result:
result = ['']
return [wtree[0]] + result + [wtree[-1]]
return normalize(wtree)
def make_wordtree(token, here_document=False):
"""Parse a delimited token and return a tree similar to the ones returned by
WordLexer. token may contain any combinations of expansion/quoted fields and
non-ones.
"""
tree = ['']
remaining = token
delimiters = '\\$`'
if not here_document:
delimiters += '\'"'
while 1:
pos, sep = find_chars(remaining, delimiters)
if pos==-1:
tree += [remaining, '']
return normalize_wordtree(tree)
tree.append(remaining[:pos])
remaining = remaining[pos:]
try:
result, remaining = WordLexer(heredoc = here_document).add(remaining, True)
except NeedMore:
raise ShellSyntaxError('Invalid token "%s"')
tree.append(result)
def wordtree_as_string(wtree):
"""Rewrite an expression tree generated by make_wordtree as string."""
def visit(node, output):
for child in node:
if isinstance(child, list):
visit(child, output)
else:
output.append(child)
output = []
visit(wtree, output)
return ''.join(output)
def unquote_wordtree(wtree):
"""Fold the word tree while removing quotes everywhere. Other expansion
sequences are joined as such.
"""
def unquote(wtree):
unquoted = []
if wtree[0] in ('', "'", '"', '\\'):
wtree = wtree[1:-1]
for part in wtree:
if isinstance(part, list):
part = unquote(part)
unquoted.append(part)
return ''.join(unquoted)
return unquote(wtree)
class HereDocLexer:
"""HereDocLexer delimits whatever comes from the here-document starting newline
not included to the closing delimiter line included.
"""
def __init__(self, op, delim):
assert op in ('<<', '<<-')
if not delim:
raise ShellSyntaxError('invalid here document delimiter %s' % str(delim))
self._op = op
self._delim = delim
self._buffer = []
self._token = []
def add(self, data, eof):
"""If the here-document was delimited, return a tuple (content, remaining).
Raise NeedMore() otherwise.
"""
self._buffer += list(data)
self._parse(eof)
token = ''.join(self._token)
remaining = ''.join(self._buffer)
self._token, self._remaining = [], []
return token, remaining
def _parse(self, eof):
while 1:
#Look for first unescaped newline. Quotes may be ignored
escaped = False
for i,c in enumerate(self._buffer):
if escaped:
escaped = False
elif c=='\\':
escaped = True
elif c=='\n':
break
else:
i = -1
if i==-1 or self._buffer[i]!='\n':
if not eof:
raise NeedMore()
#No more data, maybe the last line is closing delimiter
line = ''.join(self._buffer)
eol = ''
self._buffer[:] = []
else:
line = ''.join(self._buffer[:i])
eol = self._buffer[i]
self._buffer[:i+1] = []
if self._op=='<<-':
line = line.lstrip('\t')
if line==self._delim:
break
self._token += [line, eol]
if i==-1:
break
class Token:
#TODO: check this is still in use
OPERATOR = 'OPERATOR'
WORD = 'WORD'
def __init__(self):
self.value = ''
self.type = None
def __getitem__(self, key):
#Behave like a two elements tuple
if key==0:
return self.type
if key==1:
return self.value
raise IndexError(key)
class HereDoc:
def __init__(self, op, name=None):
self.op = op
self.name = name
self.pendings = []
TK_COMMA = 'COMMA'
TK_AMPERSAND = 'AMP'
TK_OP = 'OP'
TK_TOKEN = 'TOKEN'
TK_COMMENT = 'COMMENT'
TK_NEWLINE = 'NEWLINE'
TK_IONUMBER = 'IO_NUMBER'
TK_ASSIGNMENT = 'ASSIGNMENT_WORD'
TK_HERENAME = 'HERENAME'
class Lexer:
"""Main lexer.
Call add() until the script AST is returned.
"""
# Here-document handling makes the whole thing more complex because they basically
# force tokens to be reordered: here-content must come right after the operator
# and the here-document name, while some other tokens might be following the
# here-document expression on the same line.
#
# So, here-doc states are basically:
# *self._state==ST_NORMAL
# - self._heredoc.op is None: no here-document
# - self._heredoc.op is not None but name is: here-document operator matched,
# waiting for the document name/delimiter
# - self._heredoc.op and name are not None: here-document is ready, following
# tokens are being stored and will be pushed again when the document is
# completely parsed.
# *self._state==ST_HEREDOC
# - The here-document is being delimited by self._herelexer. Once it is done
# the content is pushed in front of the pending token list then all these
# tokens are pushed once again.
ST_NORMAL = 'ST_NORMAL'
ST_OP = 'ST_OP'
ST_BACKSLASH = 'ST_BACKSLASH'
ST_QUOTED = 'ST_QUOTED'
ST_COMMENT = 'ST_COMMENT'
ST_HEREDOC = 'ST_HEREDOC'
#Match end of backquote strings
RE_BACKQUOTE_END = re.compile(r'(?<!\\)(`)')
def __init__(self, parent_state = None):
self._input = []
self._pos = 0
self._token = ''
self._type = TK_TOKEN
self._state = self.ST_NORMAL
self._parent_state = parent_state
self._wordlexer = None
self._heredoc = HereDoc(None)
self._herelexer = None
### Following attributes are not used for delimiting token and can safely
### be changed after here-document detection (see _push_toke)
# Count the number of tokens following a 'For' reserved word. Needed to
# return an 'In' reserved word if it comes in third place.
self._for_count = None
def add(self, data, eof=False):
"""Feed the lexer with data.
When eof is set to True, returns unconsumed data or raise if the lexer
is in the middle of a delimiting operation.
Raise NeedMore otherwise.
"""
self._input += list(data)
self._parse(eof)
self._input[:self._pos] = []
return ''.join(self._input)
def _parse(self, eof):
while self._state:
if self._pos>=len(self._input):
if not eof:
raise NeedMore()
elif self._state not in (self.ST_OP, self.ST_QUOTED, self.ST_HEREDOC):
#Delimit the current token and leave cleanly
self._push_token('')
break
else:
#Let the sublexer handle the eof themselves
pass
if self._state==self.ST_NORMAL:
self._parse_normal()
elif self._state==self.ST_COMMENT:
self._parse_comment()
elif self._state==self.ST_OP:
self._parse_op(eof)
elif self._state==self.ST_QUOTED:
self._parse_quoted(eof)
elif self._state==self.ST_HEREDOC:
self._parse_heredoc(eof)
else:
assert False, "Unknown state " + str(self._state)
if self._heredoc.op is not None:
raise ShellSyntaxError('missing here-document delimiter')
def _parse_normal(self):
c = self._input[self._pos]
if c=='\n':
self._push_token(c)
self._token = c
self._type = TK_NEWLINE
self._push_token('')
self._pos += 1
elif c in ('\\', '\'', '"', '`', '$'):
self._state = self.ST_QUOTED
elif is_partial_op(c):
self._push_token(c)
self._type = TK_OP
self._token += c
self._pos += 1
self._state = self.ST_OP
elif is_blank(c):
self._push_token(c)
#Discard blanks
self._pos += 1
elif self._token:
self._token += c
self._pos += 1
elif c=='#':
self._state = self.ST_COMMENT
self._type = TK_COMMENT
self._pos += 1
else:
self._pos += 1
self._token += c
def _parse_op(self, eof):
assert self._token
while 1:
if self._pos>=len(self._input):
if not eof:
raise NeedMore()
c = ''
else:
c = self._input[self._pos]
op = self._token + c
if c and is_partial_op(op):
#Still parsing an operator
self._token = op
self._pos += 1
else:
#End of operator
self._push_token(c)
self._state = self.ST_NORMAL
break
def _parse_comment(self):
while 1:
if self._pos>=len(self._input):
raise NeedMore()
c = self._input[self._pos]
if c=='\n':
#End of comment, do not consume the end of line
self._state = self.ST_NORMAL
break
else:
self._token += c
self._pos += 1
def _parse_quoted(self, eof):
"""Precondition: the starting backquote/dollar is still in the input queue."""
if not self._wordlexer:
self._wordlexer = WordLexer()
if self._pos<len(self._input):
#Transfer input queue character into the subparser
input = self._input[self._pos:]
self._pos += len(input)
wtree, remaining = self._wordlexer.add(input, eof)
self._wordlexer = None
self._token += wordtree_as_string(wtree)
#Put unparsed character back in the input queue
if remaining:
self._input[self._pos:self._pos] = list(remaining)
self._state = self.ST_NORMAL
def _parse_heredoc(self, eof):
assert not self._token
if self._herelexer is None:
self._herelexer = HereDocLexer(self._heredoc.op, self._heredoc.name)
if self._pos<len(self._input):
#Transfer input queue character into the subparser
input = self._input[self._pos:]
self._pos += len(input)
self._token, remaining = self._herelexer.add(input, eof)
#Reset here-document state
self._herelexer = None
heredoc, self._heredoc = self._heredoc, HereDoc(None)
if remaining:
self._input[self._pos:self._pos] = list(remaining)
self._state = self.ST_NORMAL
#Push pending tokens
heredoc.pendings[:0] = [(self._token, self._type, heredoc.name)]
for token, type, delim in heredoc.pendings:
self._token = token
self._type = type
self._push_token(delim)
def _push_token(self, delim):
if not self._token:
return 0
if self._heredoc.op is not None:
if self._heredoc.name is None:
#Here-document name
if self._type!=TK_TOKEN:
raise ShellSyntaxError("expecting here-document name, got '%s'" % self._token)
self._heredoc.name = unquote_wordtree(make_wordtree(self._token))
self._type = TK_HERENAME
else:
#Capture all tokens until the newline starting the here-document
if self._type==TK_NEWLINE:
assert self._state==self.ST_NORMAL
self._state = self.ST_HEREDOC
self._heredoc.pendings.append((self._token, self._type, delim))
self._token = ''
self._type = TK_TOKEN
return 1
# BEWARE: do not change parser state from here to the end of the function:
# when parsing between an here-document operator to the end of the line
# tokens are stored in self._heredoc.pendings. Therefore, they will not
# reach the section below.
#Check operators
if self._type==TK_OP:
#False positive because of partial op matching
op = is_op(self._token)
if not op:
self._type = TK_TOKEN
else:
#Map to the specific operator
self._type = op
if self._token in ('<<', '<<-'):
#Done here rather than in _parse_op because there is no need
#to change the parser state since we are still waiting for
#the here-document name
if self._heredoc.op is not None:
raise ShellSyntaxError("syntax error near token '%s'" % self._token)
assert self._heredoc.op is None
self._heredoc.op = self._token
if self._type==TK_TOKEN:
if '=' in self._token and not delim:
if self._token.startswith('='):
#Token is a WORD... a TOKEN that is.
pass
else:
prev = self._token[:self._token.find('=')]
if is_name(prev):
self._type = TK_ASSIGNMENT
else:
#Just a token (unspecified)
pass
else:
reserved = get_reserved(self._token)
if reserved is not None:
if reserved=='In' and self._for_count!=2:
#Sorry, not a reserved word after all
pass
else:
self._type = reserved
if reserved in ('For', 'Case'):
self._for_count = 0
elif are_digits(self._token) and delim in ('<', '>'):
#Detect IO_NUMBER
self._type = TK_IONUMBER
elif self._token==';':
self._type = TK_COMMA
elif self._token=='&':
self._type = TK_AMPERSAND
elif self._type==TK_COMMENT:
#Comments are not part of sh grammar, ignore them
self._token = ''
self._type = TK_TOKEN
return 0
if self._for_count is not None:
#Track token count in 'For' expression to detect 'In' reserved words.
#Can only be in third position, no need to go beyond
self._for_count += 1
if self._for_count==3:
self._for_count = None
self.on_token((self._token, self._type))
self._token = ''
self._type = TK_TOKEN
return 1
def on_token(self, token):
raise NotImplementedError
tokens = [
TK_TOKEN,
# To silence yacc unused token warnings
# TK_COMMENT,
TK_NEWLINE,
TK_IONUMBER,
TK_ASSIGNMENT,
TK_HERENAME,
]
#Add specific operators
tokens += _OPERATORS.values()
#Add reserved words
tokens += _RESERVEDS.values()
class PLYLexer(Lexer):
"""Bridge Lexer and PLY lexer interface."""
def __init__(self):
Lexer.__init__(self)
self._tokens = []
self._current = 0
self.lineno = 0
def on_token(self, token):
value, type = token
self.lineno = 0
t = lex.LexToken()
t.value = value
t.type = type
t.lexer = self
t.lexpos = 0
t.lineno = 0
self._tokens.append(t)
def is_empty(self):
return not bool(self._tokens)
#PLY compliant interface
def token(self):
if self._current>=len(self._tokens):
return None
t = self._tokens[self._current]
self._current += 1
return t
def get_tokens(s):
"""Parse the input string and return a tuple (tokens, unprocessed) where
tokens is a list of parsed tokens and unprocessed is the part of the input
string left untouched by the lexer.
"""
lexer = PLYLexer()
untouched = lexer.add(s, True)
tokens = []
while 1:
token = lexer.token()
if token is None:
break
tokens.append(token)
tokens = [(t.value, t.type) for t in tokens]
return tokens, untouched

View File

@@ -0,0 +1,783 @@
# pyshyacc.py - PLY grammar definition for pysh
#
# Copyright 2007 Patrick Mezard
#
# This software may be used and distributed according to the terms
# of the GNU General Public License, incorporated herein by reference.
"""PLY grammar file.
"""
import os.path
import sys
import bb.pysh.pyshlex as pyshlex
tokens = pyshlex.tokens
from ply import yacc
import bb.pysh.sherrors as sherrors
class IORedirect:
def __init__(self, op, filename, io_number=None):
self.op = op
self.filename = filename
self.io_number = io_number
class HereDocument:
def __init__(self, op, name, content, io_number=None):
self.op = op
self.name = name
self.content = content
self.io_number = io_number
def make_io_redirect(p):
"""Make an IORedirect instance from the input 'io_redirect' production."""
name, io_number, io_target = p
assert name=='io_redirect'
if io_target[0]=='io_file':
io_type, io_op, io_file = io_target
return IORedirect(io_op, io_file, io_number)
elif io_target[0]=='io_here':
io_type, io_op, io_name, io_content = io_target
return HereDocument(io_op, io_name, io_content, io_number)
else:
assert False, "Invalid IO redirection token %s" % repr(io_type)
class SimpleCommand:
"""
assigns contains (name, value) pairs.
"""
def __init__(self, words, redirs, assigns):
self.words = list(words)
self.redirs = list(redirs)
self.assigns = list(assigns)
class Pipeline:
def __init__(self, commands, reverse_status=False):
self.commands = list(commands)
assert self.commands #Grammar forbids this
self.reverse_status = reverse_status
class AndOr:
def __init__(self, op, left, right):
self.op = str(op)
self.left = left
self.right = right
class ForLoop:
def __init__(self, name, items, cmds):
self.name = str(name)
self.items = list(items)
self.cmds = list(cmds)
class WhileLoop:
def __init__(self, condition, cmds):
self.condition = list(condition)
self.cmds = list(cmds)
class UntilLoop:
def __init__(self, condition, cmds):
self.condition = list(condition)
self.cmds = list(cmds)
class FunDef:
def __init__(self, name, body):
self.name = str(name)
self.body = body
class BraceGroup:
def __init__(self, cmds):
self.cmds = list(cmds)
class IfCond:
def __init__(self, cond, if_cmds, else_cmds):
self.cond = list(cond)
self.if_cmds = if_cmds
self.else_cmds = else_cmds
class Case:
def __init__(self, name, items):
self.name = name
self.items = items
class SubShell:
def __init__(self, cmds):
self.cmds = cmds
class RedirectList:
def __init__(self, cmd, redirs):
self.cmd = cmd
self.redirs = list(redirs)
def get_production(productions, ptype):
"""productions must be a list of production tuples like (name, obj) where
name is the production string identifier.
Return the first production named 'ptype'. Raise KeyError if None can be
found.
"""
for production in productions:
if production is not None and production[0]==ptype:
return production
raise KeyError(ptype)
#-------------------------------------------------------------------------------
# PLY grammar definition
#-------------------------------------------------------------------------------
def p_multiple_commands(p):
"""multiple_commands : newline_sequence
| complete_command
| multiple_commands complete_command"""
if len(p)==2:
if p[1] is not None:
p[0] = [p[1]]
else:
p[0] = []
else:
p[0] = p[1] + [p[2]]
def p_complete_command(p):
"""complete_command : list separator
| list"""
if len(p)==3 and p[2] and p[2][1] == '&':
p[0] = ('async', p[1])
else:
p[0] = p[1]
def p_list(p):
"""list : list separator_op and_or
| and_or"""
if len(p)==2:
p[0] = [p[1]]
else:
#if p[2]!=';':
# raise NotImplementedError('AND-OR list asynchronous execution is not implemented')
p[0] = p[1] + [p[3]]
def p_and_or(p):
"""and_or : pipeline
| and_or AND_IF linebreak pipeline
| and_or OR_IF linebreak pipeline"""
if len(p)==2:
p[0] = p[1]
else:
p[0] = ('and_or', AndOr(p[2], p[1], p[4]))
def p_maybe_bang_word(p):
"""maybe_bang_word : Bang"""
p[0] = ('maybe_bang_word', p[1])
def p_pipeline(p):
"""pipeline : pipe_sequence
| bang_word pipe_sequence"""
if len(p)==3:
p[0] = ('pipeline', Pipeline(p[2][1:], True))
else:
p[0] = ('pipeline', Pipeline(p[1][1:]))
def p_pipe_sequence(p):
"""pipe_sequence : command
| pipe_sequence PIPE linebreak command"""
if len(p)==2:
p[0] = ['pipe_sequence', p[1]]
else:
p[0] = p[1] + [p[4]]
def p_command(p):
"""command : simple_command
| compound_command
| compound_command redirect_list
| function_definition"""
if p[1][0] in ( 'simple_command',
'for_clause',
'while_clause',
'until_clause',
'case_clause',
'if_clause',
'function_definition',
'subshell',
'brace_group',):
if len(p) == 2:
p[0] = p[1]
else:
p[0] = ('redirect_list', RedirectList(p[1], p[2][1:]))
else:
raise NotImplementedError('%s command is not implemented' % repr(p[1][0]))
def p_compound_command(p):
"""compound_command : brace_group
| subshell
| for_clause
| case_clause
| if_clause
| while_clause
| until_clause"""
p[0] = p[1]
def p_subshell(p):
"""subshell : LPARENS compound_list RPARENS"""
p[0] = ('subshell', SubShell(p[2][1:]))
def p_compound_list(p):
"""compound_list : term
| newline_list term
| term separator
| newline_list term separator"""
productions = p[1:]
try:
sep = get_production(productions, 'separator')
if sep[1]!=';':
raise NotImplementedError()
except KeyError:
pass
term = get_production(productions, 'term')
p[0] = ['compound_list'] + term[1:]
def p_term(p):
"""term : term separator and_or
| and_or"""
if len(p)==2:
p[0] = ['term', p[1]]
else:
if p[2] is not None and p[2][1] == '&':
p[0] = ['term', ('async', p[1][1:])] + [p[3]]
else:
p[0] = p[1] + [p[3]]
def p_maybe_for_word(p):
# Rearrange 'For' priority wrt TOKEN. See p_for_word
"""maybe_for_word : For"""
p[0] = ('maybe_for_word', p[1])
def p_for_clause(p):
"""for_clause : for_word name linebreak do_group
| for_word name linebreak in sequential_sep do_group
| for_word name linebreak in wordlist sequential_sep do_group"""
productions = p[1:]
do_group = get_production(productions, 'do_group')
try:
items = get_production(productions, 'in')[1:]
except KeyError:
raise NotImplementedError('"in" omission is not implemented')
try:
items = get_production(productions, 'wordlist')[1:]
except KeyError:
items = []
name = p[2]
p[0] = ('for_clause', ForLoop(name, items, do_group[1:]))
def p_name(p):
"""name : token""" #Was NAME instead of token
p[0] = p[1]
def p_in(p):
"""in : In"""
p[0] = ('in', p[1])
def p_wordlist(p):
"""wordlist : wordlist token
| token"""
if len(p)==2:
p[0] = ['wordlist', ('TOKEN', p[1])]
else:
p[0] = p[1] + [('TOKEN', p[2])]
def p_case_clause(p):
"""case_clause : Case token linebreak in linebreak case_list Esac
| Case token linebreak in linebreak case_list_ns Esac
| Case token linebreak in linebreak Esac"""
if len(p) < 8:
items = []
else:
items = p[6][1:]
name = p[2]
p[0] = ('case_clause', Case(name, [c[1] for c in items]))
def p_case_list_ns(p):
"""case_list_ns : case_list case_item_ns
| case_item_ns"""
p_case_list(p)
def p_case_list(p):
"""case_list : case_list case_item
| case_item"""
if len(p)==2:
p[0] = ['case_list', p[1]]
else:
p[0] = p[1] + [p[2]]
def p_case_item_ns(p):
"""case_item_ns : pattern RPARENS linebreak
| pattern RPARENS compound_list linebreak
| LPARENS pattern RPARENS linebreak
| LPARENS pattern RPARENS compound_list linebreak"""
p_case_item(p)
def p_case_item(p):
"""case_item : pattern RPARENS linebreak DSEMI linebreak
| pattern RPARENS compound_list DSEMI linebreak
| LPARENS pattern RPARENS linebreak DSEMI linebreak
| LPARENS pattern RPARENS compound_list DSEMI linebreak"""
if len(p) < 7:
name = p[1][1:]
else:
name = p[2][1:]
try:
cmds = get_production(p[1:], "compound_list")[1:]
except KeyError:
cmds = []
p[0] = ('case_item', (name, cmds))
def p_pattern(p):
"""pattern : token
| pattern PIPE token"""
if len(p)==2:
p[0] = ['pattern', ('TOKEN', p[1])]
else:
p[0] = p[1] + [('TOKEN', p[2])]
def p_maybe_if_word(p):
# Rearrange 'If' priority wrt TOKEN. See p_if_word
"""maybe_if_word : If"""
p[0] = ('maybe_if_word', p[1])
def p_maybe_then_word(p):
# Rearrange 'Then' priority wrt TOKEN. See p_then_word
"""maybe_then_word : Then"""
p[0] = ('maybe_then_word', p[1])
def p_if_clause(p):
"""if_clause : if_word compound_list then_word compound_list else_part Fi
| if_word compound_list then_word compound_list Fi"""
else_part = []
if len(p)==7:
else_part = p[5]
p[0] = ('if_clause', IfCond(p[2][1:], p[4][1:], else_part))
def p_else_part(p):
"""else_part : Elif compound_list then_word compound_list else_part
| Elif compound_list then_word compound_list
| Else compound_list"""
if len(p)==3:
p[0] = p[2][1:]
else:
else_part = []
if len(p)==6:
else_part = p[5]
p[0] = ('elif', IfCond(p[2][1:], p[4][1:], else_part))
def p_while_clause(p):
"""while_clause : While compound_list do_group"""
p[0] = ('while_clause', WhileLoop(p[2][1:], p[3][1:]))
def p_maybe_until_word(p):
# Rearrange 'Until' priority wrt TOKEN. See p_until_word
"""maybe_until_word : Until"""
p[0] = ('maybe_until_word', p[1])
def p_until_clause(p):
"""until_clause : until_word compound_list do_group"""
p[0] = ('until_clause', UntilLoop(p[2][1:], p[3][1:]))
def p_function_definition(p):
"""function_definition : fname LPARENS RPARENS linebreak function_body"""
p[0] = ('function_definition', FunDef(p[1], p[5]))
def p_function_body(p):
"""function_body : compound_command
| compound_command redirect_list"""
if len(p)!=2:
raise NotImplementedError('functions redirections lists are not implemented')
p[0] = p[1]
def p_fname(p):
"""fname : TOKEN""" #Was NAME instead of token
p[0] = p[1]
def p_brace_group(p):
"""brace_group : Lbrace compound_list Rbrace"""
p[0] = ('brace_group', BraceGroup(p[2][1:]))
def p_maybe_done_word(p):
#See p_assignment_word for details.
"""maybe_done_word : Done"""
p[0] = ('maybe_done_word', p[1])
def p_maybe_do_word(p):
"""maybe_do_word : Do"""
p[0] = ('maybe_do_word', p[1])
def p_do_group(p):
"""do_group : do_word compound_list done_word"""
#Do group contains a list of AndOr
p[0] = ['do_group'] + p[2][1:]
def p_simple_command(p):
"""simple_command : cmd_prefix cmd_word cmd_suffix
| cmd_prefix cmd_word
| cmd_prefix
| cmd_name cmd_suffix
| cmd_name"""
words, redirs, assigns = [], [], []
for e in p[1:]:
name = e[0]
if name in ('cmd_prefix', 'cmd_suffix'):
for sube in e[1:]:
subname = sube[0]
if subname=='io_redirect':
redirs.append(make_io_redirect(sube))
elif subname=='ASSIGNMENT_WORD':
assigns.append(sube)
else:
words.append(sube)
elif name in ('cmd_word', 'cmd_name'):
words.append(e)
cmd = SimpleCommand(words, redirs, assigns)
p[0] = ('simple_command', cmd)
def p_cmd_name(p):
"""cmd_name : TOKEN"""
p[0] = ('cmd_name', p[1])
def p_cmd_word(p):
"""cmd_word : token"""
p[0] = ('cmd_word', p[1])
def p_maybe_assignment_word(p):
#See p_assignment_word for details.
"""maybe_assignment_word : ASSIGNMENT_WORD"""
p[0] = ('maybe_assignment_word', p[1])
def p_cmd_prefix(p):
"""cmd_prefix : io_redirect
| cmd_prefix io_redirect
| assignment_word
| cmd_prefix assignment_word"""
try:
prefix = get_production(p[1:], 'cmd_prefix')
except KeyError:
prefix = ['cmd_prefix']
try:
value = get_production(p[1:], 'assignment_word')[1]
value = ('ASSIGNMENT_WORD', value.split('=', 1))
except KeyError:
value = get_production(p[1:], 'io_redirect')
p[0] = prefix + [value]
def p_cmd_suffix(p):
"""cmd_suffix : io_redirect
| cmd_suffix io_redirect
| token
| cmd_suffix token
| maybe_for_word
| cmd_suffix maybe_for_word
| maybe_done_word
| cmd_suffix maybe_done_word
| maybe_do_word
| cmd_suffix maybe_do_word
| maybe_until_word
| cmd_suffix maybe_until_word
| maybe_assignment_word
| cmd_suffix maybe_assignment_word
| maybe_if_word
| cmd_suffix maybe_if_word
| maybe_then_word
| cmd_suffix maybe_then_word
| maybe_bang_word
| cmd_suffix maybe_bang_word"""
try:
suffix = get_production(p[1:], 'cmd_suffix')
token = p[2]
except KeyError:
suffix = ['cmd_suffix']
token = p[1]
if isinstance(token, tuple):
if token[0]=='io_redirect':
p[0] = suffix + [token]
else:
#Convert maybe_* to TOKEN if necessary
p[0] = suffix + [('TOKEN', token[1])]
else:
p[0] = suffix + [('TOKEN', token)]
def p_redirect_list(p):
"""redirect_list : io_redirect
| redirect_list io_redirect"""
if len(p) == 2:
p[0] = ['redirect_list', make_io_redirect(p[1])]
else:
p[0] = p[1] + [make_io_redirect(p[2])]
def p_io_redirect(p):
"""io_redirect : io_file
| IO_NUMBER io_file
| io_here
| IO_NUMBER io_here"""
if len(p)==3:
p[0] = ('io_redirect', p[1], p[2])
else:
p[0] = ('io_redirect', None, p[1])
def p_io_file(p):
#Return the tuple (operator, filename)
"""io_file : LESS filename
| LESSAND filename
| GREATER filename
| GREATAND filename
| DGREAT filename
| LESSGREAT filename
| CLOBBER filename"""
#Extract the filename from the file
p[0] = ('io_file', p[1], p[2][1])
def p_filename(p):
#Return the filename
"""filename : TOKEN"""
p[0] = ('filename', p[1])
def p_io_here(p):
"""io_here : DLESS here_end
| DLESSDASH here_end"""
p[0] = ('io_here', p[1], p[2][1], p[2][2])
def p_here_end(p):
"""here_end : HERENAME TOKEN"""
p[0] = ('here_document', p[1], p[2])
def p_newline_sequence(p):
# Nothing in the grammar can handle leading NEWLINE productions, so add
# this one with the lowest possible priority relatively to newline_list.
"""newline_sequence : newline_list"""
p[0] = None
def p_newline_list(p):
"""newline_list : NEWLINE
| newline_list NEWLINE"""
p[0] = None
def p_linebreak(p):
"""linebreak : newline_list
| empty"""
p[0] = None
def p_separator_op(p):
"""separator_op : COMMA
| COMMA COMMA
| AMP"""
p[0] = p[1]
def p_separator(p):
"""separator : separator_op linebreak
| newline_list"""
if len(p)==2:
#Ignore newlines
p[0] = None
else:
#Keep the separator operator
p[0] = ('separator', p[1])
def p_sequential_sep(p):
"""sequential_sep : COMMA linebreak
| newline_list"""
p[0] = None
# Low priority TOKEN => for_word conversion.
# Let maybe_for_word be used as a token when necessary in higher priority
# rules.
def p_for_word(p):
"""for_word : maybe_for_word"""
p[0] = p[1]
def p_if_word(p):
"""if_word : maybe_if_word"""
p[0] = p[1]
def p_then_word(p):
"""then_word : maybe_then_word"""
p[0] = p[1]
def p_done_word(p):
"""done_word : maybe_done_word"""
p[0] = p[1]
def p_do_word(p):
"""do_word : maybe_do_word"""
p[0] = p[1]
def p_until_word(p):
"""until_word : maybe_until_word"""
p[0] = p[1]
def p_assignment_word(p):
"""assignment_word : maybe_assignment_word"""
p[0] = ('assignment_word', p[1][1])
def p_bang_word(p):
"""bang_word : maybe_bang_word"""
p[0] = ('bang_word', p[1][1])
def p_token(p):
"""token : TOKEN
| Fi"""
p[0] = p[1]
def p_empty(p):
'empty :'
p[0] = None
# Error rule for syntax errors
def p_error(p):
msg = []
w = msg.append
if p:
w('%r\n' % p)
w('followed by:\n')
for i in range(5):
n = yacc.token()
if not n:
break
w(' %r\n' % n)
else:
w('Unexpected EOF')
raise sherrors.ShellSyntaxError(''.join(msg))
# Build the parser
try:
import pyshtables
except ImportError:
outputdir = os.path.dirname(__file__)
if not os.access(outputdir, os.W_OK):
outputdir = ''
yacc.yacc(tabmodule = 'pyshtables', outputdir = outputdir, debug = 0)
else:
yacc.yacc(tabmodule = 'pysh.pyshtables', write_tables = 0, debug = 0)
def parse(input, eof=False, debug=False):
"""Parse a whole script at once and return the generated AST and unconsumed
data in a tuple.
NOTE: eof is probably meaningless for now, the parser being unable to work
in pull mode. It should be set to True.
"""
lexer = pyshlex.PLYLexer()
remaining = lexer.add(input, eof)
if lexer.is_empty():
return [], remaining
if debug:
debug = 2
return yacc.parse(lexer=lexer, debug=debug), remaining
#-------------------------------------------------------------------------------
# AST rendering helpers
#-------------------------------------------------------------------------------
def format_commands(v):
"""Return a tree made of strings and lists. Make command trees easier to
display.
"""
if isinstance(v, list):
return [format_commands(c) for c in v]
if isinstance(v, tuple):
if len(v)==2 and isinstance(v[0], str) and not isinstance(v[1], str):
if v[0] == 'async':
return ['AsyncList', map(format_commands, v[1])]
else:
#Avoid decomposing tuples like ('pipeline', Pipeline(...))
return format_commands(v[1])
return format_commands(list(v))
elif isinstance(v, IfCond):
name = ['IfCond']
name += ['if', map(format_commands, v.cond)]
name += ['then', map(format_commands, v.if_cmds)]
name += ['else', map(format_commands, v.else_cmds)]
return name
elif isinstance(v, ForLoop):
name = ['ForLoop']
name += [repr(v.name)+' in ', map(str, v.items)]
name += ['commands', map(format_commands, v.cmds)]
return name
elif isinstance(v, AndOr):
return [v.op, format_commands(v.left), format_commands(v.right)]
elif isinstance(v, Pipeline):
name = 'Pipeline'
if v.reverse_status:
name = '!' + name
return [name, format_commands(v.commands)]
elif isinstance(v, Case):
name = ['Case']
name += [v.name, format_commands(v.items)]
elif isinstance(v, SimpleCommand):
name = ['SimpleCommand']
if v.words:
name += ['words', map(str, v.words)]
if v.assigns:
assigns = [tuple(a[1]) for a in v.assigns]
name += ['assigns', map(str, assigns)]
if v.redirs:
name += ['redirs', map(format_commands, v.redirs)]
return name
elif isinstance(v, RedirectList):
name = ['RedirectList']
if v.redirs:
name += ['redirs', map(format_commands, v.redirs)]
name += ['command', format_commands(v.cmd)]
return name
elif isinstance(v, IORedirect):
return ' '.join(map(str, (v.io_number, v.op, v.filename)))
elif isinstance(v, HereDocument):
return ' '.join(map(str, (v.io_number, v.op, repr(v.name), repr(v.content))))
elif isinstance(v, SubShell):
return ['SubShell', map(format_commands, v.cmds)]
else:
return repr(v)
def print_commands(cmds, output=sys.stdout):
"""Pretty print a command tree."""
def print_tree(cmd, spaces, output):
if isinstance(cmd, list):
for c in cmd:
print_tree(c, spaces + 3, output)
else:
print >>output, ' '*spaces + str(cmd)
formatted = format_commands(cmds)
print_tree(formatted, 0, output)
def stringify_commands(cmds):
"""Serialize a command tree as a string.
Returned string is not pretty and is currently used for unit tests only.
"""
def stringify(value):
output = []
if isinstance(value, list):
formatted = []
for v in value:
formatted.append(stringify(v))
formatted = ' '.join(formatted)
output.append(''.join(['<', formatted, '>']))
else:
output.append(value)
return ' '.join(output)
return stringify(format_commands(cmds))
def visit_commands(cmds, callable):
"""Visit the command tree and execute callable on every Pipeline and
SimpleCommand instances.
"""
if isinstance(cmds, (tuple, list)):
map(lambda c: visit_commands(c,callable), cmds)
elif isinstance(cmds, (Pipeline, SimpleCommand)):
callable(cmds)

View File

@@ -0,0 +1,15 @@
# sherrors.py - shell errors and signals
#
# Copyright 2007 Patrick Mezard
#
# This software may be used and distributed according to the terms
# of the GNU General Public License, incorporated herein by reference.
"""Define shell exceptions and error codes.
"""
class ShellError(Exception):
pass
class ShellSyntaxError(ShellError):
pass