Mirrors/miller
1
0
Fork 0
mirror of https://github.com/johnkerl/miller.git synced 2026-08-03 04:52:24 +00:00
Code Issues Projects Releases Wiki Activity

grammar neaten -- do-while doesn't end in curly brace

Browse source
This commit is contained in:
John Kerl 2016-05-26 16:28:32 -07:00
parent 28519cacea
commit e2d56810ea
2 changed files with 31 additions and 32 deletions
Download patch file Download diff file Expand all files Collapse all files

62
c/dsls/mlr_dsl_parse.y
View file

@ -14,7 +14,7 @@
// * parens, commas, semis, line endings, whitespace are all stripped away
// * variable names and literal values remain as leaf nodes of the AST
// * = + - * / ** {function names} remain as non-leaf nodes of the AST
// CST: See the md_cst.c
// CST: See mlr_dsl_cst.c
// ================================================================
}
@ -59,7 +59,7 @@ md_body ::= md_statement_list(B). {
// We handle statements of the form ' ; ; ' by parsing the empty spaces around the semicolons as NOP nodes.
// But, the NOP nodes are immediately stripped here and are not included in the AST we return.
md_statement_list(A) ::= md_unbraced_statement(B). {
md_statement_list(A) ::= md_statement_not_braced_end(B). {
if (B->type == MD_AST_NODE_TYPE_NOP) {
A = mlr_dsl_ast_node_alloc_zary("list", MD_AST_NODE_TYPE_STATEMENT_LIST);
} else {
@ -67,7 +67,7 @@ md_statement_list(A) ::= md_unbraced_statement(B). {
}
}
md_statement_list(A) ::= md_braced_statement(B). {
md_statement_list(A) ::= md_statement_braced_end(B). {
if (B->type == MD_AST_NODE_TYPE_NOP) {
A = mlr_dsl_ast_node_alloc_zary("list", MD_AST_NODE_TYPE_STATEMENT_LIST);
} else {
@ -76,9 +76,9 @@ md_statement_list(A) ::= md_braced_statement(B). {
}
// This could also be done with the list on the left and statement on the right. However,
// curly-braced statements *end* with a semicolon; they don't end with one. So this seems to
// be the right way to differentiate.
md_statement_list(A) ::= md_unbraced_statement(B) MD_TOKEN_SEMICOLON md_statement_list(C). {
// curly-brace-terminated statements *end* with a semicolon; they don't start with one. So this seems
// to be the right way to differentiate.
md_statement_list(A) ::= md_statement_not_braced_end(B) MD_TOKEN_SEMICOLON md_statement_list(C). {
if (B->type == MD_AST_NODE_TYPE_NOP) {
A = C;
} else {
@ -86,7 +86,7 @@ md_statement_list(A) ::= md_unbraced_statement(B) MD_TOKEN_SEMICOLON md_statemen
}
}
md_statement_list(A) ::= md_braced_statement(B) md_statement_list(C). {
md_statement_list(A) ::= md_statement_braced_end(B) md_statement_list(C). {
if (B->type == MD_AST_NODE_TYPE_NOP) {
A = C;
} else {
@ -95,45 +95,45 @@ md_statement_list(A) ::= md_braced_statement(B) md_statement_list(C). {
}
// This allows for trailing semicolon, as well as empty string (or whitespace) between semicolons:
md_unbraced_statement(A) ::= . {
md_statement_not_braced_end(A) ::= . {
A = mlr_dsl_ast_node_alloc_zary("nop", MD_AST_NODE_TYPE_NOP);
}
// Begin/end (non-nestable)
md_braced_statement ::= md_begin_block.
md_braced_statement ::= md_end_block.
md_statement_braced_end ::= md_begin_block.
md_statement_braced_end ::= md_end_block.
// Nested control structures:
md_braced_statement ::= md_cond_block.
md_braced_statement ::= md_while_block.
md_braced_statement ::= md_do_while_block.
md_braced_statement ::= md_for_loop_full_srec.
md_braced_statement ::= md_for_loop_full_oosvar.
md_braced_statement ::= md_for_loop_oosvar.
md_braced_statement ::= md_if_chain.
md_statement_braced_end ::= md_cond_block.
md_statement_braced_end ::= md_while_block.
md_statement_braced_end ::= md_for_loop_full_srec.
md_statement_braced_end ::= md_for_loop_full_oosvar.
md_statement_braced_end ::= md_for_loop_oosvar.
md_statement_braced_end ::= md_if_chain.
// Not valid in begin/end since they refer to srecs:
md_unbraced_statement ::= md_srec_assignment.
md_unbraced_statement ::= md_srec_indirect_assignment.
md_unbraced_statement ::= md_oosvar_from_full_srec_assignment.
md_unbraced_statement ::= md_full_srec_from_oosvar_assignment.
md_statement_not_braced_end ::= md_srec_assignment.
md_statement_not_braced_end ::= md_srec_indirect_assignment.
md_statement_not_braced_end ::= md_oosvar_from_full_srec_assignment.
md_statement_not_braced_end ::= md_full_srec_from_oosvar_assignment.
// Valid in begin/end since they don't refer to srecs (although the RHSs might):
md_unbraced_statement ::= md_bare_boolean.
md_unbraced_statement ::= md_oosvar_assignment.
md_unbraced_statement ::= md_filter.
md_unbraced_statement ::= md_unset.
md_unbraced_statement ::= md_emitf.
md_unbraced_statement ::= md_emitp.
md_unbraced_statement ::= md_emit.
md_unbraced_statement ::= md_dump.
md_statement_not_braced_end ::= md_do_while_block.
md_statement_not_braced_end ::= md_bare_boolean.
md_statement_not_braced_end ::= md_oosvar_assignment.
md_statement_not_braced_end ::= md_filter.
md_statement_not_braced_end ::= md_unset.
md_statement_not_braced_end ::= md_emitf.
md_statement_not_braced_end ::= md_emitp.
md_statement_not_braced_end ::= md_emit.
md_statement_not_braced_end ::= md_dump.
// Valid only within for/while, but we accept them here syntactically and reject them in the AST-to-CST
// conversion, where we can produce much more informative error messages:
md_unbraced_statement(A) ::= MD_TOKEN_BREAK(O). {
md_statement_not_braced_end(A) ::= MD_TOKEN_BREAK(O). {
A = mlr_dsl_ast_node_alloc(O->text, MD_AST_NODE_TYPE_BREAK);
}
md_unbraced_statement(A) ::= MD_TOKEN_CONTINUE(O). {
md_statement_not_braced_end(A) ::= MD_TOKEN_CONTINUE(O). {
A = mlr_dsl_ast_node_alloc(O->text, MD_AST_NODE_TYPE_CONTINUE);
}

1
c/todo.txt
View file

@ -6,7 +6,6 @@ BUGFIXES
================================================================
TOP-OF-LIST SUMMARY
!! do-while not brace-ending!! @ .y
! break/continue: fully support or eliminate from the lexer/parser
- in-loop stack flags for handlers (at pvars?)
- test break & continue both set