diff options
author | Justin M. Keyes <justinkz@gmail.com> | 2017-04-09 03:59:55 +0200 |
---|---|---|
committer | GitHub <noreply@github.com> | 2017-04-09 03:59:55 +0200 |
commit | cc8f640fb16a4a697e2ad380b9973048d5142a33 (patch) | |
tree | 1952c82321810c475c63ba3a8df5658ccc2d8b0b | |
parent | 3a6b8c28c813f8e65851c8e89c7cf56b4d6cf03f (diff) | |
parent | 967fa96eb2ca263c0fefd9e9be74e9862e59d740 (diff) | |
download | rneovim-cc8f640fb16a4a697e2ad380b9973048d5142a33.tar.gz rneovim-cc8f640fb16a4a697e2ad380b9973048d5142a33.tar.bz2 rneovim-cc8f640fb16a4a697e2ad380b9973048d5142a33.zip |
Merge #6478 from ZyX-I/conv-fixes
Remove conversions which are not needed when &encoding is UTF-8
-rw-r--r-- | runtime/doc/eval.txt | 4 | ||||
-rw-r--r-- | runtime/doc/options.txt | 4 | ||||
-rw-r--r-- | runtime/doc/starting.txt | 24 | ||||
-rw-r--r-- | src/nvim/digraph.h | 4 | ||||
-rw-r--r-- | src/nvim/eval/decode.c | 24 | ||||
-rw-r--r-- | src/nvim/eval/encode.c | 55 | ||||
-rw-r--r-- | src/nvim/ex_cmds.h | 4 | ||||
-rw-r--r-- | src/nvim/ex_getln.h | 6 | ||||
-rw-r--r-- | src/nvim/fold.h | 8 | ||||
-rw-r--r-- | src/nvim/memfile_defs.h | 4 | ||||
-rw-r--r-- | src/nvim/move.h | 4 | ||||
-rw-r--r-- | src/nvim/shada.c | 230 | ||||
-rw-r--r-- | test/unit/eval/decode_spec.lua | 7 |
13 files changed, 67 insertions, 311 deletions
diff --git a/runtime/doc/eval.txt b/runtime/doc/eval.txt index e21f5357a2..16f9a2ea6e 100644 --- a/runtime/doc/eval.txt +++ b/runtime/doc/eval.txt @@ -4944,8 +4944,8 @@ json_decode({expr}) *json_decode()* json_encode({expr}) *json_encode()* Convert {expr} into a JSON string. Accepts - |msgpack-special-dict| as the input. Will not convert |Funcref|s, - mappings with non-string keys (can be created as + |msgpack-special-dict| as the input. Will not convert + |Funcref|s, mappings with non-string keys (can be created as |msgpack-special-dict|), values with self-referencing containers, strings which contain non-UTF-8 characters, pseudo-UTF-8 strings which contain codepoints reserved for diff --git a/runtime/doc/options.txt b/runtime/doc/options.txt index d212e029aa..c30a88f48d 100644 --- a/runtime/doc/options.txt +++ b/runtime/doc/options.txt @@ -5151,8 +5151,8 @@ A jump table for the options with a short description can be found at |Q_op|. saved. When not included, the value of 'history' is used. *shada-c* c Dummy option, kept for compatibility reasons. Has no actual - effect. Current encoding state is described in - |shada-encoding|. + effect: ShaDa always uses UTF-8 and 'encoding' value is fixed + to UTF-8 as well. *shada-f* f Whether file marks need to be stored. If zero, file marks ('0 to '9, 'A to 'Z) are not stored. When not present or when diff --git a/runtime/doc/starting.txt b/runtime/doc/starting.txt index daf6ad9ca2..2d1dd22222 100644 --- a/runtime/doc/starting.txt +++ b/runtime/doc/starting.txt @@ -1097,23 +1097,6 @@ SHADA FILE NAME *shada-file-name* default and the name given with 'shada' or "-i" (unless it's NONE). -CHARACTER ENCODING *shada-encoding* - -The text in the ShaDa file is UTF-8-encoded. Normally you will always work -with the same 'encoding' value, and this works just fine. However, if you -read the ShaDa file with value for 'encoding' different from utf-8 and -'encoding' used when writing ShaDa file, some of the text (non-ASCII -characters) may be invalid as Neovim always attempts to convert the text in -the ShaDa file from the UTF-8 to the current 'encoding' value. Filenames are -never converted, affected elements are: - -- history strings; -- variable values; -- register values; -- last used search and substitute patterns; -- last used substitute replacement string. - - MANUALLY READING AND WRITING *shada-read-write* Two commands can be used to read and write the ShaDa file manually. This @@ -1221,8 +1204,11 @@ exactly four MessagePack objects: 3. Third goes the length of the fourth entry. Unsigned integer as well, used for fast skipping without parsing. 4. Fourth is actual entry data. All currently used ShaDa entries use - containers to hold data: either map or array. Exact format depends on the - entry type: + containers to hold data: either map or array. All string values in those + containers are either binary (applies to filenames) or UTF-8, yet parser + needs to expect that invalid bytes may be present in a UTF-8 string. + + Exact format depends on the entry type: Entry type (name) Entry data ~ 1 (Header) Map containing data that describes the generator diff --git a/src/nvim/digraph.h b/src/nvim/digraph.h index d0f10eaf78..1b73ccaf3f 100644 --- a/src/nvim/digraph.h +++ b/src/nvim/digraph.h @@ -1,8 +1,8 @@ #ifndef NVIM_DIGRAPH_H #define NVIM_DIGRAPH_H -#include "nvim/types.h" // for char_u -#include "nvim/ex_cmds_defs.h" // for exarg_T +#include "nvim/types.h" +#include "nvim/ex_cmds_defs.h" #ifdef INCLUDE_GENERATED_DECLARATIONS # include "digraph.h.generated.h" diff --git a/src/nvim/eval/decode.c b/src/nvim/eval/decode.c index fb31a65971..a7dc6b205d 100644 --- a/src/nvim/eval/decode.c +++ b/src/nvim/eval/decode.c @@ -222,8 +222,6 @@ static inline int json_decoder_pop(ValuesStackItem obj, /// Parse JSON double-quoted string /// -/// @param[in] conv Defines conversion necessary to convert UTF-8 string to -/// &encoding. /// @param[in] buf Buffer being converted. /// @param[in] buf_len Length of the buffer. /// @param[in,out] pp Pointer to the start of the string. Must point to '"'. @@ -240,8 +238,7 @@ static inline int json_decoder_pop(ValuesStackItem obj, /// value when decoder is restarted, otherwise unused. /// /// @return OK in case of success, FAIL in case of error. -static inline int parse_json_string(vimconv_T *const conv, - const char *const buf, const size_t buf_len, +static inline int parse_json_string(const char *const buf, const size_t buf_len, const char **const pp, ValuesStack *const stack, ContainerStack *const container_stack, @@ -416,20 +413,6 @@ static inline int parse_json_string(vimconv_T *const conv, } PUT_FST_IN_PAIR(fst_in_pair, str_end); #undef PUT_FST_IN_PAIR - if (conv->vc_type != CONV_NONE) { - size_t str_len = (size_t) (str_end - str); - char *const new_str = (char *) string_convert(conv, (char_u *) str, - &str_len); - if (new_str == NULL) { - emsgf(_("E474: Failed to convert string \"%.*s\" from UTF-8"), - (int) str_len, str); - xfree(str); - goto parse_json_string_fail; - } - xfree(str); - str = new_str; - str_end = new_str + str_len; - } if (hasnul) { typval_T obj; list_T *const list = tv_list_alloc(); @@ -626,9 +609,6 @@ int json_decode_string(const char *const buf, const size_t buf_len, EMSG(_("E474: Attempt to decode a blank string")); return FAIL; } - vimconv_T conv = { .vc_type = CONV_NONE }; - convert_setup(&conv, (char_u *) "utf-8", p_enc); - conv.vc_fail = true; int ret = OK; ValuesStack stack = KV_INITIAL_VALUE; ContainerStack container_stack = KV_INITIAL_VALUE; @@ -774,7 +754,7 @@ json_decode_string_cycle_start: break; } case '"': { - if (parse_json_string(&conv, buf, buf_len, &p, &stack, &container_stack, + if (parse_json_string(buf, buf_len, &p, &stack, &container_stack, &next_map_special, &didcomma, &didcolon) == FAIL) { // Error message was already given diff --git a/src/nvim/eval/encode.c b/src/nvim/eval/encode.c index 26f9aaa27d..d74913a481 100644 --- a/src/nvim/eval/encode.c +++ b/src/nvim/eval/encode.c @@ -11,7 +11,7 @@ #include <math.h> #include "nvim/eval/encode.h" -#include "nvim/buffer_defs.h" // vimconv_T +#include "nvim/buffer_defs.h" #include "nvim/eval.h" #include "nvim/eval/typval.h" #include "nvim/garray.h" @@ -29,10 +29,6 @@ #define utf_ptr2char(b) utf_ptr2char((char_u *)b) #define utf_ptr2len(b) ((size_t)utf_ptr2len((char_u *)b)) #define utf_char2len(b) ((size_t)utf_char2len(b)) -#define string_convert(a, b, c) \ - ((char *)string_convert((vimconv_T *)a, (char_u *)b, c)) -#define convert_setup(vcp, from, to) \ - (convert_setup(vcp, (char_u *)from, (char_u *)to)) const char *const encode_special_var_names[] = { [kSpecialVarNull] = "null", @@ -537,17 +533,6 @@ int encode_read_from_list(ListReaderState *const state, char *const buf, } \ } while (0) -/// Last used p_enc value -/// -/// Generic pointer: it is not used as a string, only pointer comparisons are -/// performed. Must not be freed. -static const void *last_p_enc = NULL; - -/// Conversion setup for converting from last_p_enc to UTF-8 -static vimconv_T p_enc_conv = { - .vc_type = CONV_NONE, -}; - /// Escape sequences used in JSON static const char escapes[][3] = { [BS] = "\\b", @@ -579,33 +564,15 @@ static inline int convert_to_json_string(garray_T *const gap, } else { size_t utf_len = len; char *tofree = NULL; - if (last_p_enc != (const void *) p_enc) { - p_enc_conv.vc_type = CONV_NONE; - convert_setup(&p_enc_conv, p_enc, "utf-8"); - p_enc_conv.vc_fail = true; - last_p_enc = p_enc; - } - if (p_enc_conv.vc_type != CONV_NONE) { - tofree = string_convert(&p_enc_conv, buf, &utf_len); - if (tofree == NULL) { - emsgf(_("E474: Failed to convert string \"%.*s\" to UTF-8"), - utf_len, utf_buf); - return FAIL; - } - utf_buf = tofree; - } size_t str_len = 0; - // Encode character as \u0000 if - // 1. It is an ASCII control character (0x0 .. 0x1F, 0x7F). - // 2. &encoding is not UTF-8 and code point is above 0x7F. - // 3. &encoding is UTF-8 and code point is not printable according to - // utf_printable(). - // This is done to make it possible to :echo values when &encoding is not - // UTF-8. -#define ENCODE_RAW(p_enc_conv, ch) \ - (ch >= 0x20 && (p_enc_conv.vc_type == CONV_NONE \ - ? utf_printable(ch) \ - : ch < 0x7F)) + // Encode character as \uNNNN if + // 1. It is an ASCII control character (0x0 .. 0x1F; 0x7F not + // utf_printable and thus not checked specially). + // 2. Code point is not printable according to utf_printable(). + // This is done to make resulting values displayable on screen also not from + // Neovim. +#define ENCODE_RAW(ch) \ + (ch >= 0x20 && utf_printable(ch)) for (size_t i = 0; i < utf_len;) { const int ch = utf_ptr2char(utf_buf + i); const size_t shift = (ch == 0? 1: utf_ptr2len(utf_buf + i)); @@ -636,7 +603,7 @@ static inline int convert_to_json_string(garray_T *const gap, utf_len - (i - shift), utf_buf + i - shift); xfree(tofree); return FAIL; - } else if (ENCODE_RAW(p_enc_conv, ch)) { + } else if (ENCODE_RAW(ch)) { str_len += shift; } else { str_len += ((sizeof("\\u1234") - 1) @@ -666,7 +633,7 @@ static inline int convert_to_json_string(garray_T *const gap, break; } default: { - if (ENCODE_RAW(p_enc_conv, ch)) { + if (ENCODE_RAW(ch)) { ga_concat_len(gap, utf_buf + i, shift); } else if (ch < SURROGATE_FIRST_CHAR) { ga_concat_len(gap, ((const char[]) { diff --git a/src/nvim/ex_cmds.h b/src/nvim/ex_cmds.h index 792e2f772f..b564cde56c 100644 --- a/src/nvim/ex_cmds.h +++ b/src/nvim/ex_cmds.h @@ -6,8 +6,8 @@ #include "nvim/os/time.h" #include "nvim/pos.h" #include "nvim/eval/typval.h" -#include "nvim/buffer_defs.h" // for buf_T and win_T -#include "nvim/ex_cmds_defs.h" // for exarg_T +#include "nvim/buffer_defs.h" +#include "nvim/ex_cmds_defs.h" // flags for do_ecmd() #define ECMD_HIDE 0x01 // don't free the current buffer diff --git a/src/nvim/ex_getln.h b/src/nvim/ex_getln.h index a29c8297d5..051564fbe1 100644 --- a/src/nvim/ex_getln.h +++ b/src/nvim/ex_getln.h @@ -3,9 +3,9 @@ #include "nvim/eval/typval.h" #include "nvim/ex_cmds.h" -#include "nvim/ex_cmds_defs.h" // for exarg_T -#include "nvim/os/time.h" // for Timestamp -#include "nvim/regexp_defs.h" // for regmatch_T +#include "nvim/ex_cmds_defs.h" +#include "nvim/os/time.h" +#include "nvim/regexp_defs.h" /* Values for nextwild() and ExpandOne(). See ExpandOne() for meaning. */ #define WILD_FREE 1 diff --git a/src/nvim/fold.h b/src/nvim/fold.h index 2393f4ef47..f35b328fb1 100644 --- a/src/nvim/fold.h +++ b/src/nvim/fold.h @@ -1,12 +1,12 @@ #ifndef NVIM_FOLD_H #define NVIM_FOLD_H -#include <stdio.h> // for FILE +#include <stdio.h> #include "nvim/pos.h" -#include "nvim/garray.h" // for garray_T -#include "nvim/types.h" // for char_u -#include "nvim/buffer_defs.h" // for win_T +#include "nvim/garray.h" +#include "nvim/types.h" +#include "nvim/buffer_defs.h" /* * Info used to pass info about a fold from the fold-detection code to the diff --git a/src/nvim/memfile_defs.h b/src/nvim/memfile_defs.h index 57d8abbe30..b3c2f3564c 100644 --- a/src/nvim/memfile_defs.h +++ b/src/nvim/memfile_defs.h @@ -3,10 +3,10 @@ #include <stdint.h> #include <stdbool.h> -#include <stdlib.h> // for size_t +#include <stdlib.h> #include "nvim/types.h" -#include "nvim/pos.h" // for linenr_T +#include "nvim/pos.h" /// A block number. /// diff --git a/src/nvim/move.h b/src/nvim/move.h index 1cdf1f6f52..00fbcc580f 100644 --- a/src/nvim/move.h +++ b/src/nvim/move.h @@ -2,8 +2,8 @@ #define NVIM_MOVE_H #include <stdbool.h> -#include "nvim/buffer_defs.h" // for win_T -#include "nvim/pos.h" // for linenr_T +#include "nvim/buffer_defs.h" +#include "nvim/pos.h" #ifdef INCLUDE_GENERATED_DECLARATIONS # include "move.h.generated.h" diff --git a/src/nvim/shada.c b/src/nvim/shada.c index c7b95958e0..8c5d6dff65 100644 --- a/src/nvim/shada.c +++ b/src/nvim/shada.c @@ -73,15 +73,10 @@ KHASH_SET_INIT_STR(strset) (vim_rename((char_u *)a, (char_u *)b)) #define mb_strnicmp(a, b, c) \ (mb_strnicmp((char_u *)a, (char_u *)b, c)) -#define has_non_ascii(a) (has_non_ascii((char_u *)a)) -#define string_convert(a, b, c) \ - ((char *)string_convert((vimconv_T *)a, (char_u *)b, c)) #define path_shorten_fname_if_possible(b) \ ((char *)path_shorten_fname_if_possible((char_u *)b)) #define buflist_new(ffname, sfname, ...) \ (buflist_new((char_u *)ffname, (char_u *)sfname, __VA_ARGS__)) -#define convert_setup(vcp, from, to) \ - (convert_setup(vcp, (char_u *)from, (char_u *)to)) #define os_isdir(f) (os_isdir((char_u *) f)) #define regtilde(s, m) ((char *) regtilde((char_u *) s, m)) #define path_tail_with_sep(f) ((char *) path_tail_with_sep((char_u *)f)) @@ -413,8 +408,6 @@ typedef struct sd_read_def { const char *error; ///< Error message in case of error. uintmax_t fpos; ///< Current position (amount of bytes read since ///< reader structure initialization). May overflow. - vimconv_T sd_conv; ///< Structure used for converting encodings of some - ///< items. } ShaDaReadDef; struct sd_write_def; @@ -435,8 +428,6 @@ typedef struct sd_write_def { ShaDaWriteCloser close; ///< Close function. void *cookie; ///< Data describing object written to. const char *error; ///< Error message in case of error. - vimconv_T sd_conv; ///< Structure used for converting encodings of some - ///< items. } ShaDaWriteDef; #ifdef INCLUDE_GENERATED_DECLARATIONS @@ -803,7 +794,7 @@ static int open_shada_file_for_reading(const char *const fname, return error; } - convert_setup(&sd_reader->sd_conv, "utf-8", p_enc); + assert(STRCMP(p_enc, "utf-8") == 0); return 0; } @@ -1899,127 +1890,24 @@ shada_pack_entry_error: } #undef PACK_STRING -/// Write single ShaDa entry, converting it if needed +/// Write single ShaDa entry and free it afterwards /// -/// @warning Frees entry after packing. +/// Will not free if entry could not be freed. /// /// @param[in] packer Packer used to write entry. -/// @param[in] sd_conv Conversion definitions. -/// @param[in] entry Entry written. If entry.can_free_entry is false then -/// it assumes that entry was not converted, otherwise it -/// is assumed that entry was already converted. +/// @param[in] entry Entry written. /// @param[in] max_kbyte Maximum size of an item in KiB. Zero means no /// restrictions. -static ShaDaWriteResult shada_pack_encoded_entry(msgpack_packer *const packer, - const vimconv_T *const sd_conv, - PossiblyFreedShadaEntry entry, - const size_t max_kbyte) - FUNC_ATTR_NONNULL_ALL +static inline ShaDaWriteResult shada_pack_pfreed_entry( + msgpack_packer *const packer, PossiblyFreedShadaEntry entry, + const size_t max_kbyte) + FUNC_ATTR_NONNULL_ALL FUNC_ATTR_ALWAYS_INLINE { ShaDaWriteResult ret = kSDWriteSuccessfull; + ret = shada_pack_entry(packer, entry.data, max_kbyte); if (entry.can_free_entry) { - ret = shada_pack_entry(packer, entry.data, max_kbyte); shada_free_shada_entry(&entry.data); - return ret; - } -#define RUN_WITH_CONVERTED_STRING(cstr, code) \ - do { \ - bool did_convert = false; \ - if (sd_conv->vc_type != CONV_NONE && has_non_ascii((cstr))) { \ - char *const converted_string = string_convert(sd_conv, (cstr), NULL); \ - if (converted_string != NULL) { \ - (cstr) = converted_string; \ - did_convert = true; \ - } \ - } \ - code \ - if (did_convert) { \ - xfree((cstr)); \ - } \ - } while (0) - switch (entry.data.type) { - case kSDItemUnknown: - case kSDItemMissing: { - assert(false); - } - case kSDItemSearchPattern: { - RUN_WITH_CONVERTED_STRING(entry.data.data.search_pattern.pat, { - ret = shada_pack_entry(packer, entry.data, max_kbyte); - }); - break; - } - case kSDItemHistoryEntry: { - RUN_WITH_CONVERTED_STRING(entry.data.data.history_item.string, { - ret = shada_pack_entry(packer, entry.data, max_kbyte); - }); - break; - } - case kSDItemSubString: { - RUN_WITH_CONVERTED_STRING(entry.data.data.sub_string.sub, { - ret = shada_pack_entry(packer, entry.data, max_kbyte); - }); - break; - } - case kSDItemVariable: { - if (sd_conv->vc_type != CONV_NONE) { - typval_T tgttv; - var_item_copy(sd_conv, &entry.data.data.global_var.value, &tgttv, - true, 0); - tv_clear(&entry.data.data.global_var.value); - entry.data.data.global_var.value = tgttv; - } - ret = shada_pack_entry(packer, entry.data, max_kbyte); - break; - } - case kSDItemRegister: { - bool did_convert = false; - if (sd_conv->vc_type != CONV_NONE) { - size_t first_non_ascii = 0; - for (size_t i = 0; i < entry.data.data.reg.contents_size; i++) { - if (has_non_ascii(entry.data.data.reg.contents[i])) { - first_non_ascii = i; - did_convert = true; - break; - } - } - if (did_convert) { - entry.data.data.reg.contents = - xmemdup(entry.data.data.reg.contents, - (entry.data.data.reg.contents_size - * sizeof(entry.data.data.reg.contents[0]))); - for (size_t i = 0; i < entry.data.data.reg.contents_size; i++) { - if (i >= first_non_ascii) { - entry.data.data.reg.contents[i] = get_converted_string( - sd_conv, - entry.data.data.reg.contents[i], - strlen(entry.data.data.reg.contents[i])); - } else { - entry.data.data.reg.contents[i] = - xstrdup(entry.data.data.reg.contents[i]); - } - } - } - } - ret = shada_pack_entry(packer, entry.data, max_kbyte); - if (did_convert) { - for (size_t i = 0; i < entry.data.data.reg.contents_size; i++) { - xfree(entry.data.data.reg.contents[i]); - } - xfree(entry.data.data.reg.contents); - } - break; - } - case kSDItemHeader: - case kSDItemGlobalMark: - case kSDItemJump: - case kSDItemBufferList: - case kSDItemLocalMark: - case kSDItemChange: { - ret = shada_pack_entry(packer, entry.data, max_kbyte); - break; - } } -#undef RUN_WITH_CONVERTED_STRING return ret; } @@ -2556,11 +2444,7 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, break; } typval_T tgttv; - if (sd_writer->sd_conv.vc_type != CONV_NONE) { - var_item_copy(&sd_writer->sd_conv, &vartv, &tgttv, true, 0); - } else { - tv_copy(&vartv, &tgttv); - } + tv_copy(&vartv, &tgttv); ShaDaWriteResult spe_ret; if ((spe_ret = shada_pack_entry(packer, (ShadaEntry) { .type = kSDItemVariable, @@ -2811,9 +2695,8 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, do { \ for (size_t i_ = 0; i_ < ARRAY_SIZE(wms_array); i_++) { \ if (wms_array[i_].data.type != kSDItemMissing) { \ - if (shada_pack_encoded_entry(packer, &sd_writer->sd_conv, \ - wms_array[i_], \ - max_kbyte) == kSDWriteFailed) { \ + if (shada_pack_pfreed_entry(packer, wms_array[i_], max_kbyte) \ + == kSDWriteFailed) { \ ret = kSDWriteFailed; \ goto shada_write_exit; \ } \ @@ -2823,8 +2706,8 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, PACK_WMS_ARRAY(wms->global_marks); PACK_WMS_ARRAY(wms->registers); for (size_t i = 0; i < wms->jumps_size; i++) { - if (shada_pack_encoded_entry(packer, &sd_writer->sd_conv, wms->jumps[i], - max_kbyte) == kSDWriteFailed) { + if (shada_pack_pfreed_entry(packer, wms->jumps[i], max_kbyte) + == kSDWriteFailed) { ret = kSDWriteFailed; goto shada_write_exit; } @@ -2832,8 +2715,8 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, #define PACK_WMS_ENTRY(wms_entry) \ do { \ if (wms_entry.data.type != kSDItemMissing) { \ - if (shada_pack_encoded_entry(packer, &sd_writer->sd_conv, wms_entry, \ - max_kbyte) == kSDWriteFailed) { \ + if (shada_pack_pfreed_entry(packer, wms_entry, max_kbyte) \ + == kSDWriteFailed) { \ ret = kSDWriteFailed; \ goto shada_write_exit; \ } \ @@ -2860,9 +2743,8 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, for (size_t i = 0; i < file_markss_to_dump; i++) { PACK_WMS_ARRAY(all_file_markss[i]->marks); for (size_t j = 0; j < all_file_markss[i]->changes_size; j++) { - if (shada_pack_encoded_entry(packer, &sd_writer->sd_conv, - all_file_markss[i]->changes[j], - max_kbyte) == kSDWriteFailed) { + if (shada_pack_pfreed_entry(packer, all_file_markss[i]->changes[j], + max_kbyte) == kSDWriteFailed) { ret = kSDWriteFailed; goto shada_write_exit; } @@ -2886,8 +2768,8 @@ static ShaDaWriteResult shada_write(ShaDaWriteDef *const sd_writer, if (dump_one_history[i]) { hms_insert_whole_neovim_history(&wms->hms[i]); HMS_ITER(&wms->hms[i], cur_entry, { - if (shada_pack_encoded_entry( - packer, &sd_writer->sd_conv, (PossiblyFreedShadaEntry) { + if (shada_pack_pfreed_entry( + packer, (PossiblyFreedShadaEntry) { .data = cur_entry->data, .can_free_entry = cur_entry->can_free_entry, }, max_kbyte) == kSDWriteFailed) { @@ -3038,8 +2920,6 @@ shada_write_file_nomerge: {} verbose_leave(); } - convert_setup(&sd_writer.sd_conv, p_enc, "utf-8"); - const ShaDaWriteResult sw_ret = shada_write(&sd_writer, (nomerge ? NULL : &sd_reader)); @@ -3327,29 +3207,6 @@ static ShaDaReadResult msgpack_read_uint64(ShaDaReadDef *const sd_reader, return kSDReadStatusSuccess; } -/// Convert or copy and return a string -/// -/// @param[in] sd_conv Conversion definition. -/// @param[in] str String to convert. -/// @param[in] len String length. -/// -/// @return [allocated] converted string or copy of the original string. -static inline char *get_converted_string(const vimconv_T *const sd_conv, - const char *const str, - const size_t len) - FUNC_ATTR_NONNULL_ALL FUNC_ATTR_MALLOC FUNC_ATTR_WARN_UNUSED_RESULT -{ - if (!has_non_ascii_len(str, len)) { - return xmemdupz(str, len); - } - size_t new_len = len; - char *const new_str = string_convert(sd_conv, str, &new_len); - if (new_str == NULL) { - return xmemdupz(str, len); - } - return new_str; -} - #define READERR(entry_name, error_desc) \ RERR "Error while reading ShaDa file: " \ entry_name " entry at position %" PRIu64 " " \ @@ -3427,10 +3284,7 @@ static inline char *get_converted_string(const vimconv_T *const sd_conv, sizeof(*unpacked.data.via.map.ptr)); \ ad_ga.ga_len++; \ } -#define CONVERTED(str, len) ( \ - sd_reader->sd_conv.vc_type != CONV_NONE \ - ? get_converted_string(&sd_reader->sd_conv, (str), (len)) \ - : xmemdupz((str), (len))) +#define CONVERTED(str, len) (xmemdupz((str), (len))) #define BIN_CONVERTED(b) CONVERTED(b.ptr, b.size) #define SET_ADDITIONAL_DATA(tgt, name) \ do { \ @@ -3803,30 +3657,14 @@ shada_read_next_item_start: (char) unpacked.data.via.array.ptr[2].via.u64; } size_t strsize; - if (sd_reader->sd_conv.vc_type == CONV_NONE - || !has_non_ascii_len(unpacked.data.via.array.ptr[1].via.bin.ptr, - unpacked.data.via.array.ptr[1].via.bin.size)) { -shada_read_next_item_hist_no_conv: - strsize = ( - unpacked.data.via.array.ptr[1].via.bin.size - + 1 // Zero byte - + 1); // Separator character - entry->data.history_item.string = xmalloc(strsize); - memcpy(entry->data.history_item.string, - unpacked.data.via.array.ptr[1].via.bin.ptr, - unpacked.data.via.array.ptr[1].via.bin.size); - } else { - size_t len = unpacked.data.via.array.ptr[1].via.bin.size; - char *const converted = string_convert( - &sd_reader->sd_conv, unpacked.data.via.array.ptr[1].via.bin.ptr, - &len); - if (converted != NULL) { - strsize = len + 2; - entry->data.history_item.string = xrealloc(converted, strsize); - } else { - goto shada_read_next_item_hist_no_conv; - } - } + strsize = ( + unpacked.data.via.array.ptr[1].via.bin.size + + 1 // Zero byte + + 1); // Separator character + entry->data.history_item.string = xmalloc(strsize); + memcpy(entry->data.history_item.string, + unpacked.data.via.array.ptr[1].via.bin.ptr, + unpacked.data.via.array.ptr[1].via.bin.size); entry->data.history_item.string[strsize - 2] = 0; entry->data.history_item.string[strsize - 1] = entry->data.history_item.sep; @@ -3859,16 +3697,6 @@ shada_read_next_item_hist_no_conv: "be converted to the VimL value")), initial_fpos); goto shada_read_next_item_error; } - if (sd_reader->sd_conv.vc_type != CONV_NONE) { - typval_T tgttv; - var_item_copy(&sd_reader->sd_conv, - &entry->data.global_var.value, - &tgttv, - true, - 0); - tv_clear(&entry->data.global_var.value); - entry->data.global_var.value = tgttv; - } SET_ADDITIONAL_ELEMENTS(unpacked.data.via.array, 2, entry->data.global_var.additional_elements, "variable"); diff --git a/test/unit/eval/decode_spec.lua b/test/unit/eval/decode_spec.lua index 0b2a423cd6..0c444b33f2 100644 --- a/test/unit/eval/decode_spec.lua +++ b/test/unit/eval/decode_spec.lua @@ -2,7 +2,6 @@ local helpers = require('test.unit.helpers')(after_each) local itp = helpers.gen_itp(it) local cimport = helpers.cimport -local to_cstr = helpers.to_cstr local eq = helpers.eq local neq = helpers.neq local ffi = helpers.ffi @@ -72,7 +71,7 @@ describe('json_decode_string()', function() end itp('does not overflow in error messages', function() - local saved_p_enc = decode.p_enc + collectgarbage('restart') check_failure(']test', 1, 'E474: No container to close: ]') check_failure('[}test', 2, 'E474: Closing list with curly bracket: }') check_failure('{]test', 2, @@ -105,10 +104,6 @@ describe('json_decode_string()', function() check_failure('"\194"test', 3, 'E474: Only UTF-8 strings allowed: \194"') check_failure('"\252\144\128\128\128\128"test', 8, 'E474: Only UTF-8 code points up to U+10FFFF are allowed to appear unescaped: \252\144\128\128\128\128"') check_failure('"test', 1, 'E474: Expected string end: "') - decode.p_enc = to_cstr('latin1') - check_failure('"\\uABCD"test', 8, - 'E474: Failed to convert string "ꯍ" from UTF-8') - decode.p_enc = saved_p_enc check_failure('-test', 1, 'E474: Missing number after minus sign: -') check_failure('-1.test', 3, 'E474: Missing number after decimal dot: -1.') check_failure('-1.0etest', 5, 'E474: Missing exponent: -1.0e') |