0
0
mirror of https://github.com/vim/vim.git synced 2025-07-26 11:04:33 -04:00

patch 8.2.1462: Vim9: string slice not supported yet

Problem:    Vim9: string slice not supported yet.
Solution:   Add support for string slicing.
This commit is contained in:
Bram Moolenaar 2020-08-15 21:10:16 +02:00
parent 3d1cde8a2f
commit 11107bab7e
9 changed files with 183 additions and 24 deletions

View File

@ -69,7 +69,8 @@ EXTERN char e_const_requires_a_value[]
INIT(= N_("E1021: const requires a value")); INIT(= N_("E1021: const requires a value"));
EXTERN char e_type_or_initialization_required[] EXTERN char e_type_or_initialization_required[]
INIT(= N_("E1022: type or initialization required")); INIT(= N_("E1022: type or initialization required"));
// E1023 unused EXTERN char e_cannot_slice_dictionary[]
INIT(= N_("E1023: cannot slice a dictionary"));
// E1024 unused // E1024 unused
EXTERN char e_using_rcurly_outside_if_block_scope[] EXTERN char e_using_rcurly_outside_if_block_scope[]
INIT(= N_("E1025: using } outside of a block scope")); INIT(= N_("E1025: using } outside of a block scope"));

View File

@ -3699,7 +3699,14 @@ eval_index(
case VAR_STRING: case VAR_STRING:
s = tv_get_string(rettv); s = tv_get_string(rettv);
len = (long)STRLEN(s); len = (long)STRLEN(s);
if (in_vim9script())
{
if (range) if (range)
s = string_slice(s, n1, n2);
else
s = char_from_string(s, n1);
}
else if (range)
{ {
// The resulting variable is a substring. If the indexes // The resulting variable is a substring. If the indexes
// are out of range the result is empty. // are out of range the result is empty.
@ -3718,10 +3725,6 @@ eval_index(
else else
s = vim_strnsave(s + n1, n2 - n1 + 1); s = vim_strnsave(s + n1, n2 - n1 + 1);
} }
else if (in_vim9script())
{
s = char_from_string(s, n1);
}
else else
{ {
// The resulting variable is a string of a single // The resulting variable is a string of a single
@ -5312,6 +5315,69 @@ char_from_string(char_u *str, varnumber_T index)
return vim_strnsave(str + nbyte, MB_CPTR2LEN(str + nbyte)); return vim_strnsave(str + nbyte, MB_CPTR2LEN(str + nbyte));
} }
/*
* Get the byte index for character index "idx" in string "str" with length
* "str_len".
* If going over the end return "str_len".
* If "idx" is negative count from the end, -1 is the last character.
* When going over the start return zero.
*/
static size_t
char_idx2byte(char_u *str, size_t str_len, varnumber_T idx)
{
varnumber_T nchar = idx;
size_t nbyte = 0;
if (nchar >= 0)
{
while (nchar > 0 && nbyte < str_len)
{
nbyte += MB_CPTR2LEN(str + nbyte);
--nchar;
}
}
else
{
nbyte = str_len;
while (nchar < 0 && nbyte > 0)
{
--nbyte;
nbyte -= mb_head_off(str, str + nbyte);
++nchar;
}
}
return nbyte;
}
/*
* Return the slice "str[first:last]" using character indexes.
* Return NULL when the result is empty.
*/
char_u *
string_slice(char_u *str, varnumber_T first, varnumber_T last)
{
size_t start_byte, end_byte;
size_t slen;
if (str == NULL)
return NULL;
slen = STRLEN(str);
start_byte = char_idx2byte(str, slen, first);
if (last == -1)
end_byte = slen;
else
{
end_byte = char_idx2byte(str, slen, last);
if (end_byte < slen)
// end index is inclusive
end_byte += MB_CPTR2LEN(str + end_byte);
}
if (start_byte >= slen || end_byte <= start_byte)
return NULL;
return vim_strnsave(str + start_byte, end_byte - start_byte);
}
/* /*
* Handle: * Handle:
* - expr[expr], expr[expr:expr] subscript * - expr[expr], expr[expr:expr] subscript

View File

@ -60,6 +60,7 @@ int eval_isnamec(int c);
int eval_isnamec1(int c); int eval_isnamec1(int c);
int eval_isdictc(int c); int eval_isdictc(int c);
char_u *char_from_string(char_u *str, varnumber_T index); char_u *char_from_string(char_u *str, varnumber_T index);
char_u *string_slice(char_u *str, varnumber_T first, varnumber_T last);
int handle_subscript(char_u **arg, typval_T *rettv, evalarg_T *evalarg, int verbose); int handle_subscript(char_u **arg, typval_T *rettv, evalarg_T *evalarg, int verbose);
int item_copy(typval_T *from, typval_T *to, int deep, int copyID); int item_copy(typval_T *from, typval_T *to, int deep, int copyID);
void echo_one(typval_T *rettv, int with_space, int *atstart, int *needclr); void echo_one(typval_T *rettv, int with_space, int *atstart, int *needclr);

View File

@ -971,7 +971,7 @@ def Test_disassemble_concat()
assert_equal('aabb', ConcatString()) assert_equal('aabb', ConcatString())
enddef enddef
def StringIndex(): number def StringIndex(): string
let s = "abcd" let s = "abcd"
let res = s[1] let res = s[1]
return res return res

View File

@ -2074,7 +2074,7 @@ def Test_expr7_trailing()
assert_equal(123, d.key) assert_equal(123, d.key)
enddef enddef
def Test_expr7_subscript() def Test_expr7_string_subscript()
let lines =<< trim END let lines =<< trim END
let text = 'abcdef' let text = 'abcdef'
assert_equal('', text[-1]) assert_equal('', text[-1])
@ -2094,6 +2094,28 @@ def Test_expr7_subscript()
assert_equal('f', text[5]) assert_equal('f', text[5])
assert_equal('', text[6]) assert_equal('', text[6])
assert_equal('', text[999]) assert_equal('', text[999])
assert_equal('ábçdëf', text[0:-1])
assert_equal('ábçdëf', text[0 :-1])
assert_equal('ábçdëf', text[0: -1])
assert_equal('ábçdëf', text[0 : -1])
assert_equal('ábçdëf', text[0
:-1])
assert_equal('ábçdëf', text[0:
-1])
assert_equal('ábçdëf', text[0 : -1
])
assert_equal('bçdëf', text[1:-1])
assert_equal('çdëf', text[2:-1])
assert_equal('dëf', text[3:-1])
assert_equal('ëf', text[4:-1])
assert_equal('f', text[5:-1])
assert_equal('', text[6:-1])
assert_equal('', text[999:-1])
assert_equal('ábçd', text[:3])
assert_equal('bçdëf', text[1:])
assert_equal('ábçdëf', text[:])
END END
CheckDefSuccess(lines) CheckDefSuccess(lines)
CheckScriptSuccess(['vim9script'] + lines) CheckScriptSuccess(['vim9script'] + lines)

View File

@ -754,6 +754,8 @@ static char *(features[]) =
static int included_patches[] = static int included_patches[] =
{ /* Add new patch number below this line */ { /* Add new patch number below this line */
/**/
1462,
/**/ /**/
1461, 1461,
/**/ /**/

View File

@ -117,6 +117,7 @@ typedef enum {
// expression operations // expression operations
ISN_CONCAT, ISN_CONCAT,
ISN_STRINDEX, // [expr] string index ISN_STRINDEX, // [expr] string index
ISN_STRSLICE, // [expr:expr] string slice
ISN_LISTINDEX, // [expr] list index ISN_LISTINDEX, // [expr] list index
ISN_SLICE, // drop isn_arg.number items from start of list ISN_SLICE, // drop isn_arg.number items from start of list
ISN_GETITEM, // push list item, isn_arg.number is the index ISN_GETITEM, // push list item, isn_arg.number is the index

View File

@ -3068,6 +3068,7 @@ compile_subscript(
garray_T *stack = &cctx->ctx_type_stack; garray_T *stack = &cctx->ctx_type_stack;
type_T **typep; type_T **typep;
vartype_T vtype; vartype_T vtype;
int is_slice = FALSE;
// list index: list[123] // list index: list[123]
// dict member: dict[key] // dict member: dict[key]
@ -3082,11 +3083,36 @@ compile_subscript(
*arg = skipwhite(p); *arg = skipwhite(p);
if (may_get_next_line_error(p, arg, cctx) == FAIL) if (may_get_next_line_error(p, arg, cctx) == FAIL)
return FAIL; return FAIL;
if (**arg == ':')
// missing first index is equal to zero
generate_PUSHNR(cctx, 0);
else
{
if (compile_expr0(arg, cctx) == FAIL) if (compile_expr0(arg, cctx) == FAIL)
return FAIL; return FAIL;
if (may_get_next_line_error(p, arg, cctx) == FAIL) if (may_get_next_line_error(p, arg, cctx) == FAIL)
return FAIL; return FAIL;
*arg = skipwhite(*arg);
}
if (**arg == ':')
{
*arg = skipwhite(*arg + 1);
if (may_get_next_line_error(p, arg, cctx) == FAIL)
return FAIL;
if (**arg == ']')
// missing second index is equal to end of string
generate_PUSHNR(cctx, -1);
else
{
if (compile_expr0(arg, cctx) == FAIL)
return FAIL;
if (may_get_next_line_error(p, arg, cctx) == FAIL)
return FAIL;
*arg = skipwhite(*arg);
}
is_slice = TRUE;
}
if (**arg != ']') if (**arg != ']')
{ {
emsg(_(e_missbrac)); emsg(_(e_missbrac));
@ -3098,7 +3124,8 @@ compile_subscript(
// we can use the index value type. // we can use the index value type.
// TODO: If we don't know use an instruction to figure it out at // TODO: If we don't know use an instruction to figure it out at
// runtime. // runtime.
typep = ((type_T **)stack->ga_data) + stack->ga_len - 2; typep = ((type_T **)stack->ga_data) + stack->ga_len
- (is_slice ? 3 : 2);
vtype = (*typep)->tt_type; vtype = (*typep)->tt_type;
if (*typep == &t_any) if (*typep == &t_any)
{ {
@ -3109,6 +3136,11 @@ compile_subscript(
} }
if (vtype == VAR_DICT) if (vtype == VAR_DICT)
{ {
if (is_slice)
{
emsg(_(e_cannot_slice_dictionary));
return FAIL;
}
if ((*typep)->tt_type == VAR_DICT) if ((*typep)->tt_type == VAR_DICT)
*typep = (*typep)->tt_member; *typep = (*typep)->tt_member;
else else
@ -3124,12 +3156,24 @@ compile_subscript(
} }
else if (vtype == VAR_STRING) else if (vtype == VAR_STRING)
{ {
*typep = &t_number; *typep = &t_string;
if (generate_instr_drop(cctx, ISN_STRINDEX, 1) == FAIL) if ((is_slice
? generate_instr_drop(cctx, ISN_STRSLICE, 2)
: generate_instr_drop(cctx, ISN_STRINDEX, 1)) == FAIL)
return FAIL;
}
else if (vtype == VAR_BLOB)
{
emsg("Sorry, blob index and slice not implemented yet");
return FAIL; return FAIL;
} }
else if (vtype == VAR_LIST || *typep == &t_any) else if (vtype == VAR_LIST || *typep == &t_any)
{ {
if (is_slice)
{
emsg("Sorry, list slice not implemented yet");
return FAIL;
}
if ((*typep)->tt_type == VAR_LIST) if ((*typep)->tt_type == VAR_LIST)
*typep = (*typep)->tt_member; *typep = (*typep)->tt_member;
if (generate_instr_drop(cctx, ISN_LISTINDEX, 1) == FAIL) if (generate_instr_drop(cctx, ISN_LISTINDEX, 1) == FAIL)
@ -7052,6 +7096,7 @@ delete_instr(isn_T *isn)
case ISN_FOR: case ISN_FOR:
case ISN_LISTINDEX: case ISN_LISTINDEX:
case ISN_STRINDEX: case ISN_STRINDEX:
case ISN_STRSLICE:
case ISN_GETITEM: case ISN_GETITEM:
case ISN_SLICE: case ISN_SLICE:
case ISN_MEMBER: case ISN_MEMBER:

View File

@ -70,7 +70,7 @@ typedef struct {
} ectx_T; } ectx_T;
// Get pointer to item relative to the bottom of the stack, -1 is the last one. // Get pointer to item relative to the bottom of the stack, -1 is the last one.
#define STACK_TV_BOT(idx) (((typval_T *)ectx->ec_stack.ga_data) + ectx->ec_stack.ga_len + idx) #define STACK_TV_BOT(idx) (((typval_T *)ectx->ec_stack.ga_data) + ectx->ec_stack.ga_len + (idx))
void void
to_string_error(vartype_T vartype) to_string_error(vartype_T vartype)
@ -2232,12 +2232,16 @@ call_def_function(
break; break;
case ISN_STRINDEX: case ISN_STRINDEX:
case ISN_STRSLICE:
{ {
varnumber_T n; int is_slice = iptr->isn_type == ISN_STRSLICE;
varnumber_T n1 = 0, n2;
char_u *res; char_u *res;
// string index: string is at stack-2, index at stack-1 // string index: string is at stack-2, index at stack-1
tv = STACK_TV_BOT(-2); // string slice: string is at stack-3, first index at
// stack-2, second index at stack-1
tv = is_slice ? STACK_TV_BOT(-3) : STACK_TV_BOT(-2);
if (tv->v_type != VAR_STRING) if (tv->v_type != VAR_STRING)
{ {
SOURCING_LNUM = iptr->isn_lnum; SOURCING_LNUM = iptr->isn_lnum;
@ -2245,6 +2249,18 @@ call_def_function(
goto on_error; goto on_error;
} }
if (is_slice)
{
tv = STACK_TV_BOT(-2);
if (tv->v_type != VAR_NUMBER)
{
SOURCING_LNUM = iptr->isn_lnum;
emsg(_(e_number_exp));
goto on_error;
}
n1 = tv->vval.v_number;
}
tv = STACK_TV_BOT(-1); tv = STACK_TV_BOT(-1);
if (tv->v_type != VAR_NUMBER) if (tv->v_type != VAR_NUMBER)
{ {
@ -2252,14 +2268,18 @@ call_def_function(
emsg(_(e_number_exp)); emsg(_(e_number_exp));
goto on_error; goto on_error;
} }
n = tv->vval.v_number; n2 = tv->vval.v_number;
// The resulting variable is a string of a single ectx.ec_stack.ga_len -= is_slice ? 2 : 1;
// character. If the index is too big or negative the
// result is empty.
--ectx.ec_stack.ga_len;
tv = STACK_TV_BOT(-1); tv = STACK_TV_BOT(-1);
res = char_from_string(tv->vval.v_string, n); if (is_slice)
// Slice: Select the characters from the string
res = string_slice(tv->vval.v_string, n1, n2);
else
// Index: The resulting variable is a string of a
// single character. If the index is too big or
// negative the result is empty.
res = char_from_string(tv->vval.v_string, n2);
vim_free(tv->vval.v_string); vim_free(tv->vval.v_string);
tv->vval.v_string = res; tv->vval.v_string = res;
} }
@ -3140,6 +3160,7 @@ ex_disassemble(exarg_T *eap)
// expression operations // expression operations
case ISN_CONCAT: smsg("%4d CONCAT", current); break; case ISN_CONCAT: smsg("%4d CONCAT", current); break;
case ISN_STRINDEX: smsg("%4d STRINDEX", current); break; case ISN_STRINDEX: smsg("%4d STRINDEX", current); break;
case ISN_STRSLICE: smsg("%4d STRSLICE", current); break;
case ISN_LISTINDEX: smsg("%4d LISTINDEX", current); break; case ISN_LISTINDEX: smsg("%4d LISTINDEX", current); break;
case ISN_SLICE: smsg("%4d SLICE %lld", case ISN_SLICE: smsg("%4d SLICE %lld",
current, iptr->isn_arg.number); break; current, iptr->isn_arg.number); break;