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

Commitf3a9d75

Browse files
committed
Finish cleaning up backend's handling of /* ... */ and -- comments,
per pghackers discussion around 20-Feb. Also add specific error messagesfor unterminated comments and unterminated quoted strings. These thingsare nonissues for input coming from psql, but they do matter for inputcoming from other front ends.
1 parent370186e commitf3a9d75

File tree

1 file changed

+55
-19
lines changed

1 file changed

+55
-19
lines changed

‎src/backend/parser/scan.l

Lines changed: 55 additions & 19 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 1.65 2000/02/21 18:47:02 tgl Exp $
12+
* $Header: /cvsroot/pgsql/src/backend/parser/scan.l,v 1.66 2000/03/11 05:14:06 tgl Exp $
1313
*
1414
*-------------------------------------------------------------------------
1515
*/
@@ -133,22 +133,24 @@ xdstop{dquote}
133133
xdinside[^"]+
134134

135135
/* C-style comments
136-
* Ignored by the scanner and parser.
137136
*
138137
* The "extended comment" syntax closely resembles allowable operator syntax.
139138
* The tricky part here is to get lex to recognize a string starting with
140139
* slash-star as a comment, when interpreting it as an operator would produce
141-
* a longer match --- remember lex will prefer a longer match! So, we have
142-
* to provide a special rule for xcline (a complete comment that could
143-
* otherwise look like an operator), as well as append {op_and_self}* to
144-
* xcstart so that it matches at least as much as {operator} would.
145-
* Then the tie-breaker (first matching rule of same length) wins.
146-
* There is still a problem if someone writes, eg, slash-star-star-slash-plus.
147-
* It'll be taken as an xcstart, rather than xcline and an operator as one
148-
* could wish. I don't see any way around that given lex's behavior;
149-
* that someone will just have to write a space after the comment.
140+
* a longer match --- remember lex will prefer a longer match! Also, if we
141+
* have something like plus-slash-star, lex will think this is a 3-character
142+
* operator whereas we want to see it as a + operator and a comment start.
143+
* The solution is two-fold:
144+
* 1. append {op_and_self}* to xcstart so that it matches as much text as
145+
* {operator} would. Then the tie-breaker (first matching rule of same
146+
* length) ensures xcstart wins. We put back the extra stuff with yyless()
147+
* in case it contains a star-slash that should terminate the comment.
148+
* 2. In the operator rule, check for slash-star within the operator, and
149+
* if found throw it back with yyless(). This handles the plus-slash-star
150+
* problem.
151+
* SQL92-style comments, which start with dash-dash, have similar interactions
152+
* with the operator rule.
150153
*/
151-
xcline\/\*{op_and_self}*\*\/
152154
xcstart\/\*{op_and_self}*
153155
xcstop\*+\/
154156
xcinside([^*]+)|(\*+[^/])
@@ -161,6 +163,7 @@ identifier{letter}{letter_or_digit}*
161163

162164
typecast"::"
163165

166+
/* NB: if you change "self", fix the copy in the operator rule too! */
164167
self[,()\[\].;$\:\+\-\*\/\%\^\<\>\=\|]
165168
op_and_self[\~\!\@\#\^\&\|\`\?\$\:\+\-\*\/\%\<\>\=]
166169
operator{op_and_self}+
@@ -218,27 +221,30 @@ other.
218221
*
219222
* Quoted strings must allow some special characters such as single-quote
220223
* and newline.
221-
* Embedded single-quotes are implemented both in theSQL/92-standard
224+
* Embedded single-quotes are implemented both in theSQL92-standard
222225
* style of two adjacent single quotes "''" and in the Postgres/Java style
223226
* of escaped-quote "\'".
224227
* Other embedded escaped characters are matched explicitly and the leading
225228
* backslash is dropped from the string. - thomas 1997-09-24
226-
* Note that xcline must appear before xcstart, which must appear before
227-
* operator, as explained above! Also whitespace (comment) must appear
228-
* before operator.
229+
* Note that xcstart must appear before operator, as explained above!
230+
* Also whitespace (comment) must appear before operator.
229231
*/
230232

231233
%%
232234
{whitespace}{/* ignore */ }
233235

234-
{xcline}{/* ignore */ }
235-
236-
{xcstart}{BEGIN(xc); }
236+
{xcstart}{
237+
BEGIN(xc);
238+
/* Put back any characters past slash-star; see above */
239+
yyless(2);
240+
}
237241

238242
<xc>{xcstop}{BEGIN(INITIAL); }
239243

240244
<xc>{xcinside}{/* ignore */ }
241245

246+
<xc><<EOF>>{elog(ERROR,"Unterminated /* comment"); }
247+
242248
{xbstart}{
243249
BEGIN(xb);
244250
startlit();
@@ -262,6 +268,7 @@ other.
262268
<xb>{xbcat}{
263269
/* ignore */
264270
}
271+
<xb><<EOF>>{elog(ERROR,"Unterminated binary integer"); }
265272

266273
{xhstart}{
267274
BEGIN(xh);
@@ -278,6 +285,7 @@ other.
278285
literalbuf);
279286
return ICONST;
280287
}
288+
<xh><<EOF>>{elog(ERROR,"Unterminated hexadecimal integer"); }
281289

282290
{xqstart}{
283291
BEGIN(xq);
@@ -296,6 +304,7 @@ other.
296304
<xq>{xqcat}{
297305
/* ignore */
298306
}
307+
<xq><<EOF>>{elog(ERROR,"Unterminated quoted string"); }
299308

300309

301310
{xdstart}{
@@ -310,12 +319,39 @@ other.
310319
<xd>{xdinside}{
311320
addlit(yytext, yyleng);
312321
}
322+
<xd><<EOF>>{elog(ERROR,"Unterminated quoted identifier"); }
313323

314324
{typecast}{return TYPECAST; }
315325

316326
{self}{return yytext[0]; }
317327

318328
{operator}{
329+
/* Check for embedded slash-star or dash-dash */
330+
char *slashstar =strstr((char*)yytext,"/*");
331+
char *dashdash =strstr((char*)yytext,"--");
332+
333+
if (slashstar && dashdash)
334+
{
335+
if (slashstar > dashdash)
336+
slashstar = dashdash;
337+
}
338+
elseif (!slashstar)
339+
slashstar = dashdash;
340+
341+
if (slashstar)
342+
{
343+
int nchars = slashstar - ((char*)yytext);
344+
yyless(nchars);
345+
/* If what we have left is only one char, and it's
346+
* one of the characters matching "self", then
347+
* return it as a character token the same way
348+
* that the "self" rule would have.
349+
*/
350+
if (nchars ==1 &&
351+
strchr(",()[].;$:+-*/%^<>=|", yytext[0]))
352+
return yytext[0];
353+
}
354+
319355
if (strcmp((char*)yytext,"!=") ==0)
320356
yylval.str =pstrdup("<>");/* compatibility */
321357
else

0 commit comments

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