summaryrefslogtreecommitdiff
path: root/src/preproc.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/preproc.c')
-rw-r--r--src/preproc.c318
1 files changed, 318 insertions, 0 deletions
diff --git a/src/preproc.c b/src/preproc.c
new file mode 100644
index 0000000..4cd1c5b
--- /dev/null
+++ b/src/preproc.c
@@ -0,0 +1,318 @@
+#include <ctype.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdbool.h>
+#include <stdio.h>
+
+#include "preproc.h"
+#include "utils.h"
+
+
+// Lexing
+
+typedef struct {
+ char* str;
+ size_t pos;
+} RawLine;
+
+
+String collect_identifier(RawLine* line) {
+ String id_buf = make_String(256);
+ while (line->str[line->pos] == '_' || isalnum(line->str[line->pos])) {
+ push_to_String(&id_buf, line->str[line->pos++]);
+ }
+ return id_buf;
+}
+
+
+
+TokenPtrVec lex_level(RawLine* line, char until, PreprocArena* arena);
+
+
+void expect_char(RawLine* line, char expected) {
+ if (line->str[line->pos] != expected) {
+ exit(45); // TODO: nicer handling
+ }
+ line->pos++;
+}
+
+
+/*
+MaybeChunk lex_expansion_chunk(RawLine* line, RegionVec* reg_arena, QuotationMode qmode) {
+ line->pos++; // Consume $
+ if (line->str[line->pos] == '(') {
+ line->pos++;
+ if (line->str[line->pos] == '(') {
+ exit(1);
+ // TODO: this (and potentially the other parsers must be custom, we need to recognize arithm-internal parens from the terminating `))` )
+
+
+ TokenVec inner_toks = lex_level(line, ')', reg_arena);
+ expect_char(line, ')');
+ expect_char(line, ')');
+ Region* reg_ptr = push_to_RegionVec(
+ reg_arena,
+ (Region){.mode = ARITHM, .tokens = inner_toks}
+ );
+ return (MaybeChunk){
+ .some = true,
+ .value = (Chunk){
+ .qmode = qmode,
+ .kind = REGION_EXP,
+ .as.reg_ptr = reg_ptr
+ },
+ };
+ } else {
+ TokenVec inner_toks = lex_level(line, ')', reg_arena);
+ expect_char(line, ')');
+ Region* reg_ptr = push_to_RegionVec(
+ reg_arena,
+ (Region){.mode = CMDS, .tokens = inner_toks}
+ );
+ return (MaybeChunk) {
+ .some = true,
+ .value = (Chunk){
+ .qmode = qmode,
+ .kind = REGION_EXP,
+ .as.reg_ptr = reg_ptr,
+ }
+ };
+ }
+ } else {
+ String id_name = collect_identifier(line);
+ if (id_name.count > 0) {
+ return (MaybeChunk){
+ .some = true,
+ .value = (Chunk){
+ .qmode = qmode,
+ .kind = VAR_EXP,
+ .as.varname = inner(&id_name),
+ }
+ };
+ }
+ }
+ return (MaybeChunk){.some = false};
+}
+*/
+
+
+void lex_double_quoted_section(RawLine* line, Word* coll, PreprocArena* arena) {
+ line->pos++;
+ String accum = make_String(256);
+ while (line->str[line->pos] != '\"') {
+ if (line->str[line->pos] == '\0') exit(43);
+
+ MaybeChunk exp_chunk = {.some = false};
+
+ if (line->str[line->pos] == '$') {
+ exit(1);
+ /*
+ exp_chunk = lex_expansion_chunk(line, reg_arena, DOUBLE_Q);
+ if (!exp_chunk.some) {
+ push_to_String(&accum, '$');
+ }
+ */
+ } else {
+ push_to_String(&accum, line->str[line->pos++]);
+ }
+ // TODO: other special cases beyond $
+
+ if (exp_chunk.some) {
+ if (accum.count > 0) {
+ Chunk accum_chunk = {.qmode = DOUBLE_Q, .kind = LITERAL, .as.literal_str = inner(&accum)};
+ Chunk* accum_ptr = push_to_ChunkVec(&arena->chunks, accum_chunk);
+ push_to_Word(coll, accum_ptr);
+ accum = make_String(256);
+ }
+ Chunk* exp_ptr = push_to_ChunkVec(&arena->chunks, exp_chunk.value);
+ push_to_Word(coll, exp_ptr);
+ exp_chunk.some = false;
+ }
+ }
+ line->pos++;
+ if (accum.count > 0) {
+ Chunk accum_chunk = {.qmode = DOUBLE_Q, .kind = LITERAL, .as.literal_str = inner(&accum)};
+ Chunk* accum_ptr = push_to_ChunkVec(&arena->chunks, accum_chunk);
+ push_to_Word(coll, accum_ptr);
+ accum = make_String(256);
+ }
+}
+
+
+Chunk next_word_unquoted(RawLine* line, RegionVec* reg_arena) {
+ String buf = make_String(256);
+ for (char c; (c=line->str[line->pos], c != '$' && c != '\0' && c != ' '); line->pos++) {
+ push_to_String(&buf, c);
+ }
+ return (Chunk){.qmode = UN_Q, .kind = LITERAL, .as.literal_str = inner(&buf)};
+}
+
+
+Chunk collect_single_quoted_section(RawLine* line) {
+ line->pos++; // Consume quote mark
+ String chunk_str = make_String(256);
+ while (line->str[line->pos] != '\'') {
+ if (line->str[line->pos] == '\0') exit(44);
+ push_to_String(&chunk_str, line->str[line->pos++]);
+ }
+ line->pos++; // Consume closing quote
+ return (Chunk){
+ .qmode = SINGLE_Q,
+ .kind = LITERAL,
+ .as.literal_str = inner(&chunk_str),
+ };
+}
+
+
+typedef struct {
+ Word word_germ;
+ String unq_liter_germ;
+} LexState;
+
+
+void close_word_and_lit_germs(LexState* state, TokenPtrVec* toks, PreprocArena* arena) {
+ if (state->unq_liter_germ.count > 0) {
+ Chunk chu = {
+ .qmode = UN_Q,
+ .kind = LITERAL,
+ .as.literal_str = inner(&state->unq_liter_germ)
+ };
+ Chunk* chu_ptr = push_to_ChunkVec(&arena->chunks, chu);
+ push_to_Word(&state->word_germ, chu_ptr);
+ state->unq_liter_germ = make_String(256);
+ }
+ if (state->word_germ.count > 0) {
+ Token tok = {.kind = TOK_WORD, .as.word = state->word_germ};
+ Token* tok_ptr = push_to_TokenVec(&arena->tokens, tok);
+ push_to_TokenPtrVec(toks, tok_ptr);
+ state->word_germ = make_Word(256);
+ }
+}
+
+TokenPtrVec lex_level(RawLine* line, char until, PreprocArena* arena) {
+
+ TokenPtrVec line_toks = make_TokenPtrVec(256);
+ LexState state = {
+ .word_germ = make_Word(256),
+ .unq_liter_germ = make_String(256)
+ };
+
+ while (line->str[line->pos] != until) {
+
+ if (line->str[line->pos] == ' ') {
+ close_word_and_lit_germs(&state, &line_toks, arena);
+ while (line->str[line->pos] == ' ') line->pos++;
+ continue;
+ }
+
+ Token* tok_ptr;
+
+ switch (line->str[line->pos]) {
+ case '\'':
+ push_to_Word(&state.word_germ, push_to_ChunkVec(&arena->chunks, collect_single_quoted_section(line)));
+ break;
+ case '\"':
+ lex_double_quoted_section(line, &state.word_germ, arena);
+ break;
+ case '|':
+ line->pos++;
+ close_word_and_lit_germs(&state, &line_toks, arena);
+ tok_ptr = push_to_TokenVec(&arena->tokens, (Token){.kind = TOK_SPECIAL, .as.spec = PIPE});
+ push_to_TokenPtrVec(&line_toks, tok_ptr);
+ break;
+ case '$':
+ exit(1);
+ /*
+ line->pos++;
+ MaybeChunk mchunk = lex_expansion_chunk(line, reg_arena, UN_Q);
+ if (mchunk.some) {
+ push_to_Word(&state.word_germ, mchunk.value);
+ } else {
+ push_to_String(&state.unq_liter_germ, '$');
+ }
+ */
+ break;
+ case '>':
+ close_word_and_lit_germs(&state, &line_toks, arena);
+ line->pos++;
+
+ switch (line->str[line->pos]) {
+ case '(':
+ line->pos++;
+ TokenPtrVec sub_regtoks = lex_level(line, ')', arena);
+ line->pos++;
+ Region subreg = {.mode = CMDS, .tokens = sub_regtoks};
+ Region* subreg_ptr = push_to_RegionVec(&arena->regions, subreg);
+ Token tok = {.kind = TOK_PROCSUB_OUT, .as.procsub_region = subreg_ptr};
+ tok_ptr = push_to_TokenVec(&arena->tokens, tok);
+ push_to_TokenPtrVec(&line_toks, tok_ptr);
+ break;
+ case '>':
+ line->pos++;
+ tok_ptr = push_to_TokenVec(&arena->tokens, (Token){.kind = TOK_SPECIAL, .as.spec = WRITE_REDIR_APPEND});
+ push_to_TokenPtrVec(&line_toks, tok_ptr);
+ break;
+ default:
+ tok_ptr = push_to_TokenVec(&arena->tokens, (Token){.kind = TOK_SPECIAL, .as.spec = WRITE_REDIR_TRUNC});
+ push_to_TokenPtrVec(&line_toks, tok_ptr);
+ }
+ break;
+ case '<':
+ close_word_and_lit_germs(&state, &line_toks, arena);
+ line->pos++;
+ if (line->str[line->pos] == '(') {
+ line->pos++;
+ TokenPtrVec sub_regtoks = lex_level(line, ')', arena);
+ line->pos++;
+ Region subreg = {.mode = CMDS, .tokens = sub_regtoks};
+ Region* reg_ptr = push_to_RegionVec(&arena->regions, subreg);
+ Token reg_tok = {.kind = TOK_PROCSUB_IN, .as.procsub_region = reg_ptr};
+ Token* reg_tok_ptr = push_to_TokenVec(&arena->tokens, reg_tok);
+ push_to_TokenPtrVec(&line_toks, reg_tok_ptr);
+ } else {
+ Token* redir_ptr = push_to_TokenVec(&arena->tokens, (Token){.kind = TOK_SPECIAL, .as.spec = READ_REDIR});
+ push_to_TokenPtrVec(&line_toks, redir_ptr);
+ }
+ break;
+ default:
+ push_to_String(&state.unq_liter_germ, line->str[line->pos++]);
+ }
+ }
+ // TODO: handle until char _and_ EOL
+ close_word_and_lit_germs(&state, &line_toks, arena);
+ return line_toks;
+}
+
+
+TokenPtrVec lex(char* input, PreprocArena* arena) {
+ RawLine line = {.str = input, .pos = 0};
+ TokenPtrVec toks = lex_level(&line, '\0', arena);
+ Token* eol_ptr = push_to_TokenVec(&arena->tokens, (Token){.kind = TOK_SPECIAL, .as.spec = EOL});
+ push_to_TokenPtrVec(&toks, eol_ptr);
+ return toks;
+}
+
+
+
+
+// Parsing
+
+Pipeline parse_tokstream(Token* toks) {
+ size_t tok_pos = 0;
+ Pipeline result = make_Pipeline(256);
+ TokenVec tok_buf = make_TokenVec(256);
+ while (true) {
+ Token curr_tok = toks[tok_pos++];
+ if (curr_tok.kind == TOK_SPECIAL && curr_tok.as.spec == PIPE) {
+ push_to_Pipeline(&result, (PipelineElement){.cmd_toks = tok_buf, .conn = CONN_PIPE});
+ tok_buf = make_TokenVec(256);
+ } else if (curr_tok.kind == TOK_SPECIAL && curr_tok.as.spec == EOL) {
+ push_to_Pipeline(&result, (PipelineElement){.cmd_toks = tok_buf, .conn = CONN_EOL});
+ break;
+ } else {
+ push_to_TokenVec(&tok_buf, curr_tok);
+ }
+ }
+ return result;
+}
generated by cgit v1.2.3 (git 2.46.0) at 2026-07-11 18:23:49 +0000