diff options
author | Damien George <damien.p.george@gmail.com> | 2017-02-17 12:12:40 +1100 |
---|---|---|
committer | Damien George <damien.p.george@gmail.com> | 2017-02-17 12:12:40 +1100 |
commit | 534b7c368dc2af7720f3aaed0c936ef46d773957 (patch) | |
tree | e1b71ef80c0b883728a8871fe277da33b4889c14 /py/compile.c | |
parent | 773278ec3030ea9ed809c5a248fde2278ce4b557 (diff) |
py: Do adjacent str/bytes literal concatenation in lexer, not compiler.
It's much more efficient in RAM and code size to do implicit literal string
concatenation in the lexer, as opposed to the compiler.
RAM usage is reduced because the concatenation can be done right away in the
tokeniser by just accumulating the string/bytes literals into the lexer's
vstr. Prior to this patch adjacent strings/bytes would create a parse tree
(one node per string/bytes) and then in the compiler a whole new chunk of
memory was allocated to store the concatenated string, which used more than
double the memory compared to just accumulating in the lexer.
This patch also significantly reduces code size:
bare-arm: -204
minimal: -204
unix x64: -328
stmhal: -208
esp8266: -284
cc3200: -224
Diffstat (limited to 'py/compile.c')
-rw-r--r-- | py/compile.c | 59 |
1 files changed, 0 insertions, 59 deletions
diff --git a/py/compile.c b/py/compile.c index 3fb45cd69..15e757d46 100644 --- a/py/compile.c +++ b/py/compile.c @@ -2301,65 +2301,6 @@ STATIC void compile_atom_expr_trailers(compiler_t *comp, mp_parse_node_struct_t } } -STATIC void compile_atom_string(compiler_t *comp, mp_parse_node_struct_t *pns) { - // a list of strings - - // check type of list (string or bytes) and count total number of bytes - int n = MP_PARSE_NODE_STRUCT_NUM_NODES(pns); - size_t n_bytes = 0; - int string_kind = MP_PARSE_NODE_NULL; - for (int i = 0; i < n; i++) { - int pn_kind; - if (MP_PARSE_NODE_IS_LEAF(pns->nodes[i])) { - pn_kind = MP_PARSE_NODE_LEAF_KIND(pns->nodes[i]); - assert(pn_kind == MP_PARSE_NODE_STRING || pn_kind == MP_PARSE_NODE_BYTES); - n_bytes += qstr_len(MP_PARSE_NODE_LEAF_ARG(pns->nodes[i])); - } else { - assert(MP_PARSE_NODE_IS_STRUCT(pns->nodes[i])); - mp_parse_node_struct_t *pns_string = (mp_parse_node_struct_t*)pns->nodes[i]; - if (MP_PARSE_NODE_STRUCT_KIND(pns_string) == PN_string) { - pn_kind = MP_PARSE_NODE_STRING; - } else { - assert(MP_PARSE_NODE_STRUCT_KIND(pns_string) == PN_bytes); - pn_kind = MP_PARSE_NODE_BYTES; - } - n_bytes += pns_string->nodes[1]; - } - if (i == 0) { - string_kind = pn_kind; - } else if (pn_kind != string_kind) { - compile_syntax_error(comp, (mp_parse_node_t)pns, "cannot mix bytes and nonbytes literals"); - return; - } - } - - // if we are not in the last pass, just load a dummy object - if (comp->pass != MP_PASS_EMIT) { - EMIT_ARG(load_const_obj, mp_const_none); - return; - } - - // concatenate string/bytes - vstr_t vstr; - vstr_init_len(&vstr, n_bytes); - byte *s_dest = (byte*)vstr.buf; - for (int i = 0; i < n; i++) { - if (MP_PARSE_NODE_IS_LEAF(pns->nodes[i])) { - size_t s_len; - const byte *s = qstr_data(MP_PARSE_NODE_LEAF_ARG(pns->nodes[i]), &s_len); - memcpy(s_dest, s, s_len); - s_dest += s_len; - } else { - mp_parse_node_struct_t *pns_string = (mp_parse_node_struct_t*)pns->nodes[i]; - memcpy(s_dest, (const char*)pns_string->nodes[0], pns_string->nodes[1]); - s_dest += pns_string->nodes[1]; - } - } - - // load the object - EMIT_ARG(load_const_obj, mp_obj_new_str_from_vstr(string_kind == MP_PARSE_NODE_STRING ? &mp_type_str : &mp_type_bytes, &vstr)); -} - // pns needs to have 2 nodes, first is lhs of comprehension, second is PN_comp_for node STATIC void compile_comprehension(compiler_t *comp, mp_parse_node_struct_t *pns, scope_kind_t kind) { assert(MP_PARSE_NODE_STRUCT_NUM_NODES(pns) == 2); |