Files
zulip/tools/lib/css_parser.py
Steve Howell c0a6038a95 css: Enforce one selector per line.
While it's sometimes nice to put a few selectors on the same line,
it is generally better to have a consistent way of formatting our
selectors, and most of our code up until now lists them vertically.
This change fixes the linter to enforce one selector per line, and
it cleans up the places in the CSS where we had multiple selectors
on the same line.

The advantages of one-per-line are as followers:
    * cleaner diffs
    * easier to see when multiple areas of the app may have the
      same format
    * less likely to go over 80 cols
    * makes it more clear where we have deep nesting in the
      individual selectors
    * makes it easier for our linting tools to enforce
      whitespace violations

This also fixed an old bug where we had ".landing_page h2, h4", which
sets "h4" styles outside of the landing page.
2017-03-26 16:57:33 -07:00

618 lines
19 KiB
Python

from __future__ import absolute_import
from __future__ import print_function
from six.moves import range
from typing import Callable, List, Tuple, Union
####### Helpers
class Token(object):
def __init__(self, s, line, col):
# type: (str, int, int) -> None
self.s = s
self.line = line
self.col = col
class CssParserException(Exception):
def __init__(self, msg, token):
# type: (str, Token) -> None
self.msg = msg
self.token = token
def __str__(self):
# type: () -> str
return self.msg
def find_end_brace(tokens, i, end):
# type: (List[Token], int, int) -> int
depth = 0
while i < end:
s = tokens[i].s
if s == '{':
depth += 1
elif s == '}':
if depth == 0:
raise CssParserException('unexpected }', tokens[i])
elif depth == 1:
break
depth -= 1
i += 1
else:
raise CssParserException('missing }', tokens[i-1])
return i
def get_whitespace_and_comments(tokens, i, end, line=None):
# type: (List[Token], int, int, int) -> Tuple[int, str]
def is_fluff_token(token):
# type: (Token) -> bool
s = token.s
if ws(s[0]):
return True
elif s.startswith('/*'):
# For CSS comments, the caller may pass in a line
# number to indicate that they only want to get
# comments on the same line. (Subsequent comments
# will be attached to the next actual line of code.)
if line is None:
return True
if tokens[i].line == line:
return True
return False
text = ''
while (i < end) and is_fluff_token(tokens[i]):
s = tokens[i].s
text += s
i += 1
return i, text
############### Begin parsing here
def parse_sections(tokens, start, end):
# type: (List[Token], int, int) -> CssSectionList
i = start
sections = []
while i < end:
start, pre_fluff = get_whitespace_and_comments(tokens, i, end)
if start >= end:
raise CssParserException('unexpected empty section', tokens[end-1])
i = find_end_brace(tokens, start, end)
section_end = i + 1
i, post_fluff = get_whitespace_and_comments(tokens, i+1, end)
section = parse_section(
tokens=tokens,
start=start,
end=section_end,
pre_fluff=pre_fluff,
post_fluff=post_fluff
)
sections.append(section)
section_list = CssSectionList(
tokens=tokens,
sections=sections,
)
return section_list
def parse_section(tokens, start, end, pre_fluff, post_fluff):
# type: (List[Token], int, int, str, str) -> Union[CssNestedSection, CssSection]
assert not ws(tokens[start].s)
assert tokens[end-1].s == '}' # caller should strip trailing fluff
first_token = tokens[start].s
if first_token in ('@media', '@keyframes') or first_token.startswith('@-'):
i, selector_list = parse_selectors_section(tokens, start, end) # not technically selectors
section_list = parse_sections(tokens, i+1, end-1)
nested_section = CssNestedSection(
tokens=tokens,
selector_list=selector_list,
section_list=section_list,
pre_fluff=pre_fluff,
post_fluff=post_fluff,
)
return nested_section
else:
i, selector_list = parse_selectors_section(tokens, start, end)
declaration_block = parse_declaration_block(tokens, i, end)
section = CssSection(
tokens=tokens,
selector_list=selector_list,
declaration_block=declaration_block,
pre_fluff=pre_fluff,
post_fluff=post_fluff,
)
return section
def parse_selectors_section(tokens, start, end):
# type: (List[Token], int, int) -> Tuple[int, CssSelectorList]
start, pre_fluff = get_whitespace_and_comments(tokens, start, end)
assert pre_fluff == ''
i = start
text = ''
while i < end and tokens[i].s != '{':
s = tokens[i].s
text += s
i += 1
selector_list = parse_selectors(tokens, start, i)
return i, selector_list
def parse_selectors(tokens, start, end):
# type: (List[Token], int, int) -> CssSelectorList
i = start
selectors = []
while i < end:
s = tokens[i].s
if s == ',':
selector = parse_selector(tokens, start, i)
selectors.append(selector)
i += 1
start = i
if s.startswith('/*'):
raise CssParserException('Comments in selector section are not allowed', tokens[i])
i += 1
selector = parse_selector(tokens, start, i)
selectors.append(selector)
selector_list = CssSelectorList(
tokens=tokens,
selectors=selectors,
)
return selector_list
def parse_selector(tokens, start, end):
# type: (List[Token], int, int) -> CssSelector
i, pre_fluff = get_whitespace_and_comments(tokens, start, end)
levels = []
last_i = None
while i < end:
token = tokens[i]
i += 1
if not ws(token.s[0]):
last_i = i
levels.append(token)
if last_i is None:
raise CssParserException('Missing selector', tokens[-1])
assert last_i is not None
start, post_fluff = get_whitespace_and_comments(tokens, last_i, end)
selector = CssSelector(
tokens=tokens,
pre_fluff=pre_fluff,
post_fluff=post_fluff,
levels=levels,
)
return selector
def parse_declaration_block(tokens, start, end):
# type: (List[Token], int, int) -> CssDeclarationBlock
assert tokens[start].s == '{' # caller should strip leading fluff
assert tokens[end-1].s == '}' # caller should strip trailing fluff
i = start + 1
declarations = []
while i < end-1:
start = i
i, _ = get_whitespace_and_comments(tokens, i, end)
while (i < end) and (tokens[i].s != ';'):
i += 1
if i < end:
i, _ = get_whitespace_and_comments(tokens, i+1, end, line=tokens[i].line)
declaration = parse_declaration(tokens, start, i)
declarations.append(declaration)
declaration_block = CssDeclarationBlock(
tokens=tokens,
declarations=declarations,
)
return declaration_block
def parse_declaration(tokens, start, end):
# type: (List[Token], int, int) -> CssDeclaration
i, pre_fluff = get_whitespace_and_comments(tokens, start, end)
if (i >= end) or (tokens[i].s == '}'):
raise CssParserException('Empty declaration or missing semicolon', tokens[i-1])
css_property = tokens[i].s
if tokens[i+1].s != ':':
raise CssParserException('We expect a colon here', tokens[i])
i += 2
start = i
while (i < end) and (tokens[i].s != ';') and (tokens[i].s != '}'):
i += 1
css_value = parse_value(tokens, start, i)
semicolon = (i < end) and (tokens[i].s == ';')
if semicolon:
i += 1
_, post_fluff = get_whitespace_and_comments(tokens, i, end)
declaration = CssDeclaration(
tokens=tokens,
pre_fluff=pre_fluff,
post_fluff=post_fluff,
css_property=css_property,
css_value=css_value,
semicolon=semicolon,
)
return declaration
def parse_value(tokens, start, end):
# type: (List[Token], int, int) -> CssValue
i, pre_fluff = get_whitespace_and_comments(tokens, start, end)
if i < end:
value = tokens[i]
else:
raise CssParserException('Missing value', tokens[i-1])
i, post_fluff = get_whitespace_and_comments(tokens, i+1, end)
return CssValue(
tokens=tokens,
value=value,
pre_fluff=pre_fluff,
post_fluff=post_fluff,
)
def handle_prefluff(pre_fluff, indent=False):
# type: (str, bool) -> str
pre_fluff_lines = pre_fluff.split('\n')
formatted_pre_fluff_lines = []
comment_indent = ''
general_indent = ''
if indent:
general_indent = ' '
for i, ln in enumerate(pre_fluff_lines):
line_indent = ''
if ln.strip() != '':
if not i:
line_indent = general_indent
comment_indent = ' '
else:
if comment_indent:
if ('*/' in ln or '*' in ln) and (ln.strip()[:2] in ('*/', '* ', '*')):
line_indent = general_indent
if '*/' in ln:
comment_indent = ''
else:
line_indent = general_indent + comment_indent
else:
line_indent = general_indent
comment_indent = ' '
elif len(pre_fluff_lines) == 1 and indent and ln != '':
line_indent = ' '
formatted_pre_fluff_lines.append(line_indent + ln.strip())
if formatted_pre_fluff_lines[-1] != '':
if formatted_pre_fluff_lines[-1].strip() == '' and indent:
formatted_pre_fluff_lines[-1] = ''
formatted_pre_fluff_lines.append('')
pre_fluff = '\n'.join(formatted_pre_fluff_lines)
res = ''
if indent:
if '\n' in pre_fluff:
res = pre_fluff + ' '
elif pre_fluff == '':
res = ' '
else:
res = pre_fluff.rstrip() + ' '
else:
res = pre_fluff
return res
def handle_postfluff(post_fluff, indent=False, space_after_first_line=False):
# type: (str, bool, bool) -> str
post_fluff_lines = post_fluff.split('\n')
formatted_post_fluff_lines = []
comment_indent = ''
general_indent = ''
if indent:
general_indent = ' '
for i, ln in enumerate(post_fluff_lines):
line_indent = ''
if ln.strip() != '':
if i:
if comment_indent:
if ('*/' in ln or '*' in ln) and (ln.strip()[:2] in ('*/', '* ', '*')):
line_indent = general_indent
if '*/' in ln:
comment_indent = ''
else:
line_indent = general_indent + comment_indent
else:
line_indent = general_indent
comment_indent = ' '
elif indent and not i and len(post_fluff_lines) > 2:
formatted_post_fluff_lines.append('')
line_indent = general_indent
comment_indent = ' '
elif space_after_first_line:
line_indent = ' '
if not i:
comment_indent = ' '
elif not i:
comment_indent = ' '
formatted_post_fluff_lines.append(line_indent + ln.strip())
if len(formatted_post_fluff_lines) == 1 and not space_after_first_line:
if formatted_post_fluff_lines[-1].strip() == '':
if formatted_post_fluff_lines[-1] != '':
formatted_post_fluff_lines[-1] = ' '
else:
formatted_post_fluff_lines.append('')
elif formatted_post_fluff_lines[-1].strip() == '':
formatted_post_fluff_lines[-1] = ''
if len(formatted_post_fluff_lines) == 1 and indent:
formatted_post_fluff_lines.append('')
elif space_after_first_line:
formatted_post_fluff_lines.append('')
post_fluff = '\n'.join(formatted_post_fluff_lines)
return post_fluff
#### Begin CSS classes here
class CssSectionList(object):
def __init__(self, tokens, sections):
# type: (List[Token], List[Union[CssNestedSection, CssSection]]) -> None
self.tokens = tokens
self.sections = sections
def text(self):
# type: () -> str
res = ''.join(section.text() for section in self.sections)
return res
class CssNestedSection(object):
def __init__(self, tokens, selector_list, section_list, pre_fluff, post_fluff):
# type: (List[Token], CssSelectorList, CssSectionList, str, str) -> None
self.tokens = tokens
self.selector_list = selector_list
self.section_list = section_list
self.pre_fluff = pre_fluff
self.post_fluff = post_fluff
def text(self):
# type: () -> str
res = ''
res += self.pre_fluff
res += self.selector_list.text()
res += ' {'
section_list_lines = self.section_list.text().split('\n')
formatted_section_list = []
for ln in section_list_lines:
if ln.strip() == '':
formatted_section_list.append('')
else:
formatted_section_list.append(' ' + ln)
res += '\n'.join(formatted_section_list)
res += '}'
res += self.post_fluff
return res
class CssSection(object):
def __init__(self, tokens, selector_list, declaration_block, pre_fluff, post_fluff):
# type: (List[Token], CssSelectorList, CssDeclarationBlock, str, str) -> None
self.tokens = tokens
self.selector_list = selector_list
self.declaration_block = declaration_block
self.pre_fluff = pre_fluff
self.post_fluff = post_fluff
def text(self):
# type: () -> str
res = ''
res += handle_prefluff(self.pre_fluff)
res += self.selector_list.text()
res += ' '
res += self.declaration_block.text()
res += handle_postfluff(self.post_fluff, space_after_first_line=True)
return res
class CssSelectorList(object):
def __init__(self, tokens, selectors):
# type: (List[Token], List[CssSelector]) -> None
self.tokens = tokens
self.selectors = selectors
def text(self):
# type: () -> str
return ',\n'.join(sel.text() for sel in self.selectors)
class CssSelector(object):
def __init__(self, tokens, pre_fluff, post_fluff, levels):
# type: (List[Token],str, str, List[Token]) -> None
self.tokens = tokens
self.pre_fluff = pre_fluff
self.post_fluff = post_fluff
self.levels = levels
def text(self):
# type: () -> str
res = ' '.join(level.s for level in self.levels)
return res
class CssDeclarationBlock(object):
def __init__(self, tokens, declarations):
# type: (List[Token], List[CssDeclaration]) -> None
self.tokens = tokens
self.declarations = declarations
def text(self):
# type: () -> str
res = '{'
for declaration in self.declarations:
res += declaration.text()
res += '}'
return res
class CssDeclaration(object):
def __init__(self, tokens, pre_fluff, post_fluff, css_property, css_value, semicolon):
# type: (List[Token], str, str, str, CssValue, bool) -> None
self.tokens = tokens
self.pre_fluff = pre_fluff
self.post_fluff = post_fluff
self.css_property = css_property
self.css_value = css_value
self.semicolon = semicolon
def text(self):
# type: () -> str
res = ''
res += handle_prefluff(self.pre_fluff, True)
res += self.css_property
res += ':'
value_text = self.css_value.text()
if '\n' in value_text:
# gradient values can be multi-line
res += value_text.rstrip()
else:
res += ' '
res += value_text.strip()
res += ';'
res += handle_postfluff(self.post_fluff, True, True)
return res
class CssValue(object):
def __init__(self, tokens, value, pre_fluff, post_fluff):
# type: (List[Token], Token, str, str) -> None
self.value = value
self.pre_fluff = pre_fluff
self.post_fluff = post_fluff
assert pre_fluff.strip() == ''
def text(self):
# type: () -> str
return self.pre_fluff + self.value.s + self.post_fluff
def parse(text):
# type: (str) -> CssSectionList
tokens = tokenize(text)
section_list = parse_sections(tokens, 0, len(tokens))
return section_list
#### Begin tokenizer section here
def ws(c):
# type: (str) -> bool
return c in ' \t\n'
def tokenize(text):
# type: (str) -> List[Token]
class State(object):
def __init__(self):
# type: () -> None
self.i = 0
self.line = 1
self.col = 1
tokens = []
state = State()
def add_token(s, state):
# type: (str, State) -> None
# deep copy data
token = Token(s=s, line=state.line, col=state.col)
tokens.append(token)
def legal(offset):
# type: (int) -> bool
return state.i + offset < len(text)
def advance(n):
# type: (int) -> None
for _ in range(n):
state.i += 1
if state.i >= 0 and text[state.i - 1] == '\n':
state.line += 1
state.col = 1
else:
state.col += 1
def looking_at(s):
# type: (str) -> bool
return text[state.i:state.i+len(s)] == s
def get_field(terminator):
# type: (Callable[[str], bool]) -> str
offset = 0
paren_level = 0
while legal(offset) and (paren_level or not terminator(text[state.i + offset])):
c = text[state.i + offset]
if c == '(':
paren_level += 1
elif c == ')':
paren_level -= 1
offset += 1
return text[state.i:state.i+offset]
in_property = False
in_value = False
in_media_line = False
starting_media_section = False
while state.i < len(text):
c = text[state.i]
if c in '{};:,':
if c == ':':
in_property = False
in_value = True
elif c == ';':
in_property = True
in_value = False
elif c in '{':
if starting_media_section:
starting_media_section = False
else:
in_property = True
elif c == '}':
in_property = False
s = c
elif ws(c):
terminator = lambda c: not ws(c)
s = get_field(terminator)
elif looking_at('/*'):
# hacky
old_i = state.i
while (state.i < len(text)) and not looking_at('*/'):
state.i += 1
if not looking_at('*/'):
raise CssParserException('unclosed comment', tokens[-1])
s = text[old_i:state.i+2]
state.i = old_i
elif looking_at('@media'):
s = '@media'
in_media_line = True
starting_media_section = True
elif in_media_line:
in_media_line = False
terminator = lambda c: c == '{'
s = get_field(terminator)
s = s.rstrip()
elif in_property:
terminator = lambda c: ws(c) or c in ':{'
s = get_field(terminator)
elif in_value:
in_value = False
in_property = True
terminator = lambda c: c in ';}'
s = get_field(terminator)
s = s.rstrip()
else:
terminator = lambda c: ws(c) or c == ','
s = get_field(terminator)
add_token(s, state)
advance(len(s))
return tokens