view libinterp/parse-tree/octave.gperf @ 29724:c19f8cbe0fd5

initial implementation of parsing for arguments validaton block (bug #59405) This change allows parsing of arguments validation blocks. Octave should now accept the arguments block syntax in a mostly Matlab compatible way. Multiple argument blocks are allowed. All arguments blocks must appear before any other exectuable statements in a function. Similar to "methods", "properties", etc., "arguments" is defined as a keyword in the octave.gperf file so that converting "arguments" to the ARGUMENTS token in the lexer and parser is simplified but it is not really treated as a reserved keyword in the language. One known problem with the current approach is that function [...] = f (...) arguments = 13; ... end will result in a parse error. A simple workaround is to place another statement (that is not an arguments block) ahead of the "arguments = ..." line in the function. Fixing this problem generally might require a different parsing method that allows a different type of lookahead than we currently use. NOTE: arguments blocks do not currently perform any actions. Since they may provide default values and/or transform arguments to different types or values, ignoring the arguments block can lead to incorrect results. Octave also currently allows arguments blocks in nested functions though they should be rejected. Special treatment of "arguments" and "endarguments" may be disabled by defining the macro DISABLE_ARGUMENTS_VALIDATION_BLOCK. With this macro defined, Octave's lexer will never return the ARGUMENTS token, so the parser will fail to parse these program elements and the behavior should be the same as prior to this change. * pt-args-block.h, pt-args-block.cc: New files. * libinterp/parse-tree/module.mk: Update. * lex.h, lex.ll (lexical_feedback::m_arguments_is_keyword): New member variable. (lexical_feedback::reset): Reset m_arguments_is_keyword to false. (iskeyword, Fiskeyword): Also handle "arguments" as a special case. (base_lexer::make_keyword_token): Handle arguments and endarguments. * parse.h, oct-parse.yy (base_parser::make_arguments_block, base_parser::make_args_attribute_list, base_parser::make_arg_validation, base_parser::make_args_validation_list, base_parser::append_args_validation_list, base_parser::make_arg_size_spec, base_parser::make_arg_validation_fcns): New functions. (function_body, at_first_executable_stmt, function_body1, arguments_block, arguments_beg, args_attr_list, args_validation_list, arg_validation, size_spec, class_name, validation_fcns, default_value): New non-terminals. (ARGUMENTS): New token. (function): Use new function_body and function_body1 non-terminals to accept arguments blocks, but only at the beginning of a function. (fcn_name): Set lexer.m_arguments_is_keyword to true after parsing function name. (param_list_beg): Set lexer.m_arguments_is_keyword to false while parsing function parameter list. (param_list_beg): Reset lexer.m_arguments_is_keyword to true after parsing function parameter list. (tree_arguments_block_type, tree_args_block_attribute_list_type, tree_args_block_validation_list_type, tree_arg_size_spec_type, tree_arg_validation_type, tree_arg_validation_fcns_type): New non-terminal types. Also declare %destructors for them. * octave.gperf (octave_kw_id): New IDs, arguments_kw and endarguments_kw. Provide entries for arguments and endarguments keywords. * pt-all.h: Include pt-args-block.h. * pt-bp.h, pt-bp.cc (tree_breakpoint::visit_arguments_block, tree_breakpoint::visit_args_block_attribute_list, tree_breakpoint::visit_args_block_validation_list, tree_breakpoint::visit_arg_validation, tree_breakpoint::visit_arg_size_spec, tree_breakpoint::visit_arg_validation_fcns): New virtual functions for arguments block elements. * pt-eval.h, pt-eval.cc (tree_evaluator::visit_arguments_block, tree_evaluator::visit_args_block_attribute_list, tree_evaluator::visit_args_block_validation_list, tree_evaluator::visit_arg_validation, tree_evaluator::visit_arg_size_spec, tree_evaluator::visit_arg_validation_fcns): New virtual functions for arguments block elements. * pt-pr-code.h, pt-pr-code.cc (tree_print_code::visit_arguments_block, tree_print_code::visit_args_block_attribute_list, tree_print_code::visit_args_block_validation_list, tree_print_code::visit_arg_validation, tree_print_code::visit_arg_size_spec, tree_print_code::visit_arg_validation_fcns): New virtual functions for arguments block elements. * pt-walk.h, pt-walk.cc (tree_walker::visit_arguments_block, tree_walker::visit_args_block_attribute_list, tree_walker::visit_args_block_validation_list, tree_walker::visit_arg_validation, tree_walker::visit_arg_size_spec, tree_walker::visit_arg_validation_fcns): New virtual functions for arguments block elements. * token.h (end_tok_type): New type, arguments_end.
author John W. Eaton <jwe@octave.org>
date Tue, 01 Jun 2021 13:34:57 -0400
parents 7854d5752dd2
children 796f54d4ddbf
line wrap: on
line source

%{

////////////////////////////////////////////////////////////////////////
//
// Copyright (C) 1995-2021 The Octave Project Developers
//
// See the file COPYRIGHT.md in the top-level directory of this
// distribution or <https://octave.org/copyright/>.
//
// This file is part of Octave.
//
// Octave is free software: you can redistribute it and/or modify it
// under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// Octave is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with Octave; see the file COPYING.  If not, see
// <https://www.gnu.org/licenses/>.
//
// NOTE: gperf 2.7.2 will silently generate bad code if there are blank
// lines following the "%{" marker above.  This comment block seems to be
// handled correctly.
//
////////////////////////////////////////////////////////////////////////

enum octave_kw_id
{
  arguments_kw,
  break_kw,
  case_kw,
  catch_kw,
  classdef_kw,
  continue_kw,
  do_kw,
  else_kw,
  elseif_kw,
  end_kw,
  end_try_catch_kw,
  end_unwind_protect_kw,
  endarguments_kw,
  endclassdef_kw,
  endenumeration_kw,
  endevents_kw,
  endfor_kw,
  endfunction_kw,
  endif_kw,
  endmethods_kw,
  endparfor_kw,
  endproperties_kw,
  endspmd_kw,
  endswitch_kw,
  endwhile_kw,
  enumeration_kw,
  events_kw,
  for_kw,
  function_kw,
  get_kw,
  global_kw,
  if_kw,
  magic_file_kw,
  magic_line_kw,
  methods_kw,
  otherwise_kw,
  parfor_kw,
  persistent_kw,
  properties_kw,
  return_kw,
  set_kw,
  spmd_kw,
  switch_kw,
  try_kw,
  until_kw,
  unwind_protect_kw,
  unwind_protect_cleanup_kw,
  while_kw
};

%}

struct octave_kw { const char *name; int tok; octave_kw_id kw_id; };

%%
arguments, ARGUMENTS, arguments_kw
break, BREAK, break_kw
case, CASE, case_kw
catch, CATCH, catch_kw
classdef, CLASSDEF, classdef_kw
continue, CONTINUE, continue_kw
do, DO, do_kw
else, ELSE, else_kw
elseif, ELSEIF, elseif_kw
end, END, end_kw
end_try_catch, END, end_try_catch_kw
end_unwind_protect, END, end_unwind_protect_kw
endarguments, END, endarguments_kw
endclassdef, END, endclassdef_kw
endenumeration, END, endenumeration_kw
endevents, END, endevents_kw
endfor, END, endfor_kw
endfunction, END, endfunction_kw
endif, END, endif_kw
endmethods, END, endmethods_kw
endparfor, END, endparfor_kw
endproperties, END, endproperties_kw
endspmd, END, endspmd_kw
endswitch, END, endswitch_kw
endwhile, END, endwhile_kw
enumeration, ENUMERATION, enumeration_kw
events, EVENTS, events_kw
for, FOR, for_kw
function, FCN, function_kw
get, GET, get_kw
global, GLOBAL, global_kw
if, IF, if_kw
methods, METHODS, methods_kw
otherwise, OTHERWISE, otherwise_kw
parfor, PARFOR, parfor_kw
persistent, PERSISTENT, persistent_kw
properties, PROPERTIES, properties_kw
return, FUNC_RET, return_kw
set, SET, set_kw
spmd, SPMD, spmd_kw
switch, SWITCH, switch_kw
try, TRY, try_kw
until, UNTIL, until_kw
unwind_protect, UNWIND, unwind_protect_kw
unwind_protect_cleanup, CLEANUP, unwind_protect_cleanup_kw
while, WHILE, while_kw
__FILE__, DQ_STRING, magic_file_kw
__LINE__, NUMBER, magic_line_kw