Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit61446e0

Browse files
committed
Improve lexer's error reporting. You get the whole token mentioned now
in parse error messages, not just the part scanned by the last flex rule.For example,select "foo" "bar";used to drawERROR: parser: parse error at or near """which was rather unhelpful. Now it givesERROR: parser: parse error at or near ""bar""Also, error messages concerning bitstring literals and suchlike willquote the source text at you, not the processed internal form of the literal.
1 parent241978b commit61446e0

File tree

3 files changed

+73
-48
lines changed

3 files changed

+73
-48
lines changed

‎src/backend/parser/scan.l

Lines changed: 71 additions & 46 deletions
Original file line numberDiff line numberDiff line change
@@ -9,7 +9,7 @@
99
*
1010
*
1111
* IDENTIFICATION
12-
* $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.92 2002/04/20 21:56:14 petere Exp $
12+
* $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.93 2002/05/01 17:12:07 tgl Exp $
1313
*
1414
*-------------------------------------------------------------------------
1515
*/
@@ -56,6 +56,17 @@ static void addlit(char *ytext, int yleng);
5656
staticvoidaddlitchar(unsignedchar ychar);
5757
staticchar *litbufdup(void);
5858

59+
/*
60+
* When we parse a token that requires multiple lexer rules to process,
61+
* we set token_start to point at the true start of the token, for use
62+
* by yyerror(). yytext will point at just the text consumed by the last
63+
* rule, so it's not very helpful (eg, it might contain just the last
64+
* quote mark of a quoted identifier). But to avoid cluttering every rule
65+
* with setting token_start, we allow token_start = NULL to denote that
66+
* it's okay to use yytext.
67+
*/
68+
staticchar *token_start;
69+
5970
/* Handles to the buffer that the lexer uses internally*/
6071
static YY_BUFFER_STATE scanbufhandle;
6172
staticchar *scanbuf;
@@ -208,7 +219,7 @@ non_newline[^\n\r]
208219

209220
comment("--"{non_newline}*)
210221

211-
whitespace({space}|{comment})
222+
whitespace({space}+|{comment})
212223

213224
/*
214225
* SQL92 requires at least one newline in the whitespace separating
@@ -235,9 +246,16 @@ other.
235246
*/
236247

237248
%%
249+
250+
%{
251+
/* code to execute during start of each call of yylex()*/
252+
token_start =NULL;
253+
%}
254+
238255
{whitespace}{/* ignore */ }
239256

240257
{xcstart}{
258+
token_start = yytext;
241259
xcdepth =0;
242260
BEGIN(xc);
243261
/* Put back any characters past slash-star; see above */
@@ -252,7 +270,11 @@ other.
252270

253271
<xc>{xcstop}{
254272
if (xcdepth <=0)
273+
{
255274
BEGIN(INITIAL);
275+
/* reset token_start for next token */
276+
token_start =NULL;
277+
}
256278
else
257279
xcdepth--;
258280
}
@@ -261,18 +283,18 @@ other.
261283

262284
<xc>{op_chars}{/* ignore */ }
263285

264-
<xc><<EOF>>{elog(ERROR,"Unterminated /* comment"); }
286+
<xc><<EOF>>{yyerror("unterminated /* comment"); }
265287

266288
{xbitstart}{
289+
token_start = yytext;
267290
BEGIN(xbit);
268291
startlit();
269292
addlitchar('b');
270293
}
271294
<xbit>{xbitstop}{
272295
BEGIN(INITIAL);
273296
if (literalbuf[strspn(literalbuf +1,"01") +1] !='\0')
274-
elog(ERROR,"invalid bit string input: '%s'",
275-
literalbuf);
297+
yyerror("invalid bit string input");
276298
yylval.str =litbufdup();
277299
return BITCONST;
278300
}
@@ -284,9 +306,10 @@ other.
284306
<xbit>{xbitcat}{
285307
/* ignore */
286308
}
287-
<xbit><<EOF>>{elog(ERROR,"unterminated bit string literal"); }
309+
<xbit><<EOF>>{yyerror("unterminated bit string literal"); }
288310

289311
{xhstart}{
312+
token_start = yytext;
290313
BEGIN(xh);
291314
startlit();
292315
}
@@ -303,14 +326,14 @@ other.
303326
|| val != (long) ((int32) val)
304327
#endif
305328
)
306-
elog(ERROR,"Bad hexadecimal integer input '%s'",
307-
literalbuf);
329+
yyerror("bad hexadecimal integer input");
308330
yylval.ival = val;
309331
return ICONST;
310332
}
311-
<xh><<EOF>>{elog(ERROR,"Unterminated hexadecimal integer"); }
333+
<xh><<EOF>>{yyerror("unterminated hexadecimal integer"); }
312334

313335
{xqstart}{
336+
token_start = yytext;
314337
BEGIN(xq);
315338
startlit();
316339
}
@@ -335,30 +358,31 @@ other.
335358
<xq>{xqcat}{
336359
/* ignore */
337360
}
338-
<xq><<EOF>>{elog(ERROR,"Unterminated quoted string"); }
361+
<xq><<EOF>>{yyerror("unterminated quoted string"); }
339362

340363

341364
{xdstart}{
365+
token_start = yytext;
342366
BEGIN(xd);
343367
startlit();
344368
}
345369
<xd>{xdstop}{
346370
BEGIN(INITIAL);
347-
if (strlen(literalbuf) ==0)
348-
elog(ERROR,"zero-length delimited identifier");
349-
if (strlen(literalbuf) >= NAMEDATALEN)
371+
if (literallen ==0)
372+
yyerror("zero-length delimited identifier");
373+
if (literallen >= NAMEDATALEN)
350374
{
351-
#ifdef MULTIBYTE
352375
int len;
353-
len =pg_mbcliplen(literalbuf,strlen(literalbuf),NAMEDATALEN-1);
354-
elog(WARNING,"identifier\"%s\" will be truncated to\"%.*s\"",
355-
literalbuf, len, literalbuf);
356-
literalbuf[len] ='\0';
376+
#ifdef MULTIBYTE
377+
len =pg_mbcliplen(literalbuf, literallen,
378+
NAMEDATALEN-1);
357379
#else
358-
elog(WARNING,"identifier\"%s\" will be truncated to\"%.*s\"",
359-
literalbuf, NAMEDATALEN-1, literalbuf);
360-
literalbuf[NAMEDATALEN-1] ='\0';
380+
len = NAMEDATALEN-1;
361381
#endif
382+
elog(NOTICE,"identifier\"%s\" will be truncated to\"%.*s\"",
383+
literalbuf, len, literalbuf);
384+
literalbuf[len] ='\0';
385+
literallen = len;
362386
}
363387
yylval.str =litbufdup();
364388
return IDENT;
@@ -369,7 +393,7 @@ other.
369393
<xd>{xdinside}{
370394
addlit(yytext, yyleng);
371395
}
372-
<xd><<EOF>>{elog(ERROR,"Unterminated quoted identifier"); }
396+
<xd><<EOF>>{yyerror("unterminated quoted identifier"); }
373397

374398
{typecast}{return TYPECAST; }
375399

@@ -383,8 +407,8 @@ other.
383407
* character will match a prior rule, not this one.
384408
*/
385409
intnchars = yyleng;
386-
char *slashstar =strstr((char*)yytext,"/*");
387-
char *dashdash =strstr((char*)yytext,"--");
410+
char *slashstar =strstr(yytext,"/*");
411+
char *dashdash =strstr(yytext,"--");
388412

389413
if (slashstar && dashdash)
390414
{
@@ -395,7 +419,7 @@ other.
395419
elseif (!slashstar)
396420
slashstar = dashdash;
397421
if (slashstar)
398-
nchars = slashstar -((char*)yytext);
422+
nchars = slashstar - yytext;
399423

400424
/*
401425
* For SQL92 compatibility, '+' and '-' cannot be the
@@ -437,15 +461,15 @@ other.
437461
}
438462

439463
/* Convert "!=" operator to "<>" for compatibility */
440-
if (strcmp((char*)yytext,"!=") ==0)
464+
if (strcmp(yytext,"!=") ==0)
441465
yylval.str =pstrdup("<>");
442466
else
443-
yylval.str =pstrdup((char*)yytext);
467+
yylval.str =pstrdup(yytext);
444468
return Op;
445469
}
446470

447471
{param}{
448-
yylval.ival =atol((char*)&yytext[1]);
472+
yylval.ival =atol(yytext +1);
449473
return PARAM;
450474
}
451475

@@ -454,7 +478,7 @@ other.
454478
char* endptr;
455479

456480
errno =0;
457-
val =strtol((char *)yytext, &endptr,10);
481+
val =strtol(yytext, &endptr,10);
458482
if (*endptr !='\0' || errno == ERANGE
459483
#ifdef HAVE_LONG_INT_64
460484
/* if long > 32 bits, check for overflow of int4 */
@@ -463,28 +487,29 @@ other.
463487
)
464488
{
465489
/* integer too large, treat it as a float */
466-
yylval.str =pstrdup((char*)yytext);
490+
yylval.str =pstrdup(yytext);
467491
return FCONST;
468492
}
469493
yylval.ival = val;
470494
return ICONST;
471495
}
472496
{decimal}{
473-
yylval.str =pstrdup((char*)yytext);
497+
yylval.str =pstrdup(yytext);
474498
return FCONST;
475499
}
476500
{real}{
477-
yylval.str =pstrdup((char*)yytext);
501+
yylval.str =pstrdup(yytext);
478502
return FCONST;
479503
}
480504

481505

482506
{identifier}{
483507
ScanKeyword *keyword;
508+
char *ident;
484509
inti;
485510

486511
/* Is it a keyword? */
487-
keyword =ScanKeywordLookup((char*)yytext);
512+
keyword =ScanKeywordLookup(yytext);
488513
if (keyword !=NULL)
489514
return keyword->value;
490515

@@ -496,26 +521,25 @@ other.
496521
* which seems appropriate under SQL99 rules, whereas
497522
* the keyword comparison was NOT locale-dependent.
498523
*/
499-
for (i =0; yytext[i]; i++)
524+
ident =pstrdup(yytext);
525+
for (i =0; ident[i]; i++)
500526
{
501-
if (isupper((unsignedchar)yytext[i]))
502-
yytext[i] =tolower((unsignedchar)yytext[i]);
527+
if (isupper((unsignedchar)ident[i]))
528+
ident[i] =tolower((unsignedchar)ident[i]);
503529
}
504530
if (i >= NAMEDATALEN)
505531
{
506-
#ifdef MULTIBYTE
507532
int len;
508-
len =pg_mbcliplen(yytext,i,NAMEDATALEN-1);
509-
elog(WARNING,"identifier\"%s\" will be truncated to\"%.*s\"",
510-
yytext, len, yytext);
511-
yytext[len] ='\0';
533+
#ifdef MULTIBYTE
534+
len =pg_mbcliplen(ident, i, NAMEDATALEN-1);
512535
#else
513-
elog(WARNING,"identifier\"%s\" will be truncated to\"%.*s\"",
514-
yytext, NAMEDATALEN-1, yytext);
515-
yytext[NAMEDATALEN-1] ='\0';
536+
len = NAMEDATALEN-1;
516537
#endif
538+
elog(NOTICE,"identifier\"%s\" will be truncated to\"%.*s\"",
539+
ident, len, ident);
540+
ident[len] ='\0';
517541
}
518-
yylval.str =pstrdup((char*) yytext);
542+
yylval.str =ident;
519543
return IDENT;
520544
}
521545

@@ -526,7 +550,8 @@ other.
526550
void
527551
yyerror(constchar *message)
528552
{
529-
elog(ERROR,"parser: %s at or near\"%s\"", message, yytext);
553+
elog(ERROR,"parser: %s at or near\"%s\"", message,
554+
token_start ? token_start : yytext);
530555
}
531556

532557

‎src/backend/po/nls.mk

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
11
CATALOG_NAME:= postgres
22
AVAIL_LANGUAGES:= cs de hu ru zh_CN zh_TW
33
GETTEXT_FILES:= + gettext-files
4-
GETTEXT_TRIGGERS:= elog:2 postmaster_error
4+
GETTEXT_TRIGGERS:= elog:2 postmaster_error yyerror

‎src/test/regress/expected/strings.out

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -17,7 +17,7 @@ SELECT 'first line'
1717
' - next line' /* this comment is not allowed here */
1818
' - third line'
1919
AS "Illegal comment within continuation";
20-
ERROR: parser: parse error at or near "'"
20+
ERROR: parser: parse error at or near "' - third line'"
2121
--
2222
-- test conversions between various string types
2323
--

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp