diff options
Diffstat (limited to 'lex.c')
| -rw-r--r-- | lex.c | 123 |
1 files changed, 84 insertions, 39 deletions
@@ -50,11 +50,10 @@ MaybeChunk lex_expansion_chunk(RawLine* line, RegionVec* reg_arena, QuotationMod TokenVec inner_toks = lex_level(line, ')', reg_arena); expect_char(line, ')'); expect_char(line, ')'); - push_to_RegionVec( + Region* reg_ptr = push_to_RegionVec( reg_arena, (Region){.mode = ARITHM, .tokens = inner_toks} ); - Region* reg_ptr = ®_arena->items[reg_arena->count]; // TODO: maybe have push() return a pointer [in the macro] return (MaybeChunk){ .some = true, .value = (Chunk){ @@ -66,11 +65,10 @@ MaybeChunk lex_expansion_chunk(RawLine* line, RegionVec* reg_arena, QuotationMod } else { TokenVec inner_toks = lex_level(line, ')', reg_arena); expect_char(line, ')'); - push_to_RegionVec( + Region* reg_ptr = push_to_RegionVec( reg_arena, - (Region){.mode = CMD_SUB, .tokens = inner_toks} + (Region){.mode = CMDS, .tokens = inner_toks} ); - Region* reg_ptr = ®_arena->items[reg_arena->count]; return (MaybeChunk) { .some = true, .value = (Chunk){ @@ -164,65 +162,112 @@ Chunk collect_single_quoted_section(RawLine* line) { } +typedef struct { + TokenVec tokens; + Word word_germ; + String unq_liter_germ; +} LexState; + + +void close_literal_germ(LexState* state) { + if (state->unq_liter_germ.count > 0) { + push_to_Word( + &state->word_germ, + (Chunk){ + .qmode = UN_Q, + .kind = LITERAL, + .as.literal_str = inner(state->unq_liter_germ) + } + ); + state->unq_liter_germ = make_String(256); + } +} + +void close_word_and_lit_germs(LexState* state) { + close_literal_germ(state); + if (state->word_germ.count > 0) { + push_to_TokenVec(&state->tokens, (Token){.kind = TOK_WORD, .as.word = state->word_germ}); + state->word_germ = make_Word(256); + } +} TokenVec lex_level(RawLine* line, char until, RegionVec* reg_arena) { - printf("\n\n==== START LEVEL ===\n\n"); - TokenVec tokens = make_TokenVec(256); - Word curr_chunks = make_Word(256); - String word_buf = make_String(256); + LexState state = { + .tokens = make_TokenVec(256), + .word_germ = make_Word(256), + .unq_liter_germ = make_String(256) + }; while (line->str[line->pos] != until) { - printf("\n\nLoop %d %d %d (%s)\n", tokens.count, curr_chunks.count, word_buf.count, word_buf.items); if (line->str[line->pos] == ' ') { - if (word_buf.count > 0) { - push_to_Word( - &curr_chunks, - (Chunk){.qmode = UN_Q, .kind = LITERAL, .as.literal_str = inner(word_buf)} - ); - word_buf = make_String(256); - } - if (curr_chunks.count > 0) { - push_to_TokenVec(&tokens, (Token){.kind = TOK_WORD, .as.word = curr_chunks}); - curr_chunks = make_Word(256); - } - + close_word_and_lit_germs(&state); while (line->str[line->pos] == ' ') line->pos++; + continue; } switch (line->str[line->pos]) { case '\'': - push_to_Word(&curr_chunks, collect_single_quoted_section(line)); + push_to_Word(&state.word_germ, collect_single_quoted_section(line)); break; case '\"': - lex_double_quoted_section(line, &curr_chunks, reg_arena); + lex_double_quoted_section(line, &state.word_germ, reg_arena); break; case '$': line->pos++; MaybeChunk mchunk = lex_expansion_chunk(line, reg_arena, UN_Q); if (mchunk.some) { - push_to_Word(&curr_chunks, mchunk.value); + push_to_Word(&state.word_germ, mchunk.value); } else { - push_to_String(&word_buf, '$'); + push_to_String(&state.unq_liter_germ, '$'); + } + break; + case '>': + close_word_and_lit_germs(&state); + line->pos++; + + switch (line->str[line->pos]) { + case '(': + line->pos++; + TokenVec sub_regtoks = lex_level(line, ')', reg_arena); + Region subreg = {.mode = CMDS, .tokens = sub_regtoks}; + Region* reg_ptr = push_to_RegionVec(reg_arena, subreg); + line->pos++; + + push_to_TokenVec(&state.tokens, (Token){.kind = TOK_PROCSUB_OUT, .as.procsub_region = reg_ptr}); + break; + case '>': + line->pos++; + push_to_TokenVec(&state.tokens, (Token){.kind = TOK_SPECIAL, .as.spec = WRITE_REDIR_APPEND}); + break; + default: + push_to_TokenVec(&state.tokens, (Token){.kind = TOK_SPECIAL, .as.spec = WRITE_REDIR_TRUNC}); + } + break; + case '<': + close_word_and_lit_germs(&state); + line->pos++; + if (line->str[line->pos] == '(') { + line->pos++; + TokenVec sub_regtoks = lex_level(line, ')', reg_arena); + Region subreg = {.mode = CMDS, .tokens = sub_regtoks}; + Region* reg_ptr = push_to_RegionVec(reg_arena, subreg); + line->pos++; + + push_to_TokenVec(&state.tokens, (Token){.kind = TOK_PROCSUB_IN, .as.procsub_region = reg_ptr}); + break; + } else { + push_to_TokenVec(&state.tokens, (Token){.kind = TOK_SPECIAL, .as.spec = READ_REDIR}); } break; default: - push_to_String(&word_buf, line->str[line->pos++]); + push_to_String(&state.unq_liter_germ, line->str[line->pos++]); } } - - if (word_buf.count > 0) { - push_to_Word( - &curr_chunks, - (Chunk){.qmode = UN_Q, .kind = LITERAL, .as.literal_str = inner(word_buf)} - ); - } - if (curr_chunks.count > 0) { - push_to_TokenVec(&tokens, (Token){.kind = TOK_WORD, .as.word = curr_chunks}); - } - printf("\nFIN: %d %d %d\n\n", tokens.count, curr_chunks.count, word_buf.count); - return tokens; + // TODO: handle until char _and_ EOL + close_word_and_lit_germs(&state); + return state.tokens; } |