[RFA] Fix number lexing, again

[RFA] Fix number lexing, again

[Keith Seitz]

[-- Attachment #1: Type: text/plain, Size: 675 bytes --]

Hi,

As someone pointed out, I did not fix all the corner cases with 
linespec_lexer_lex_number.  [See 
http://sourceware.org/ml/gdb/2012-07/msg00078.html - the issue affects 
all platforms]

Since it is still quite common for MI-based UIs to quote the entire 
linespec, we need to terminate numbers successfully on any quotation.

While the list of growing terminations for the string is getting 
tediously long, I don't think inverting the test would be any 
shorter/clearer. WDYT?

Keith

PS. Also recommending this for 7.5.

ChangeLog
2012-07-30  Keith Seitz  <keiths@redhat.com>

	* linespec.c (linespec_lex_number): A number followed
	by quotes is a valid number, too.

[-- Attachment #2: lex_number-quotes.patch --]
[-- Type: text/x-patch, Size: 782 bytes --]

diff --git a/gdb/linespec.c b/gdb/linespec.c
index 3d7f62f..51994c8 100644
--- a/gdb/linespec.c
+++ b/gdb/linespec.c
@@ -391,10 +391,11 @@ linespec_lexer_lex_number (linespec_parser *parser, linespec_token *tokenp)
     }
 
   /* If the next character in the input buffer is not a space, comma,
-     or colon, this input does not represent a number.  */
+     quote, or colon, this input does not represent a number.  */
   if (*PARSER_STREAM (parser) != '\0'
       && !isspace (*PARSER_STREAM (parser)) && *PARSER_STREAM (parser) != ','
-      && *PARSER_STREAM (parser) != ':')
+      && *PARSER_STREAM (parser) != ':'
+      && !strchr (linespec_quote_characters, *PARSER_STREAM (parser)))
     {
       PARSER_STREAM (parser) = LS_TOKEN_STOKEN (*tokenp).ptr;
       return 0;

Re: [RFA] Fix number lexing, again

[Keith Seitz]

[-- Attachment #1: Type: text/plain, Size: 382 bytes --]

On 07/30/2012 09:38 AM, Keith Seitz wrote:

Gah! This time with the tests included!

> ChangeLog
> 2012-07-30  Keith Seitz  <keiths@redhat.com>
>
>      * linespec.c (linespec_lex_number): A number followed
>      by quotes is a valid number, too.

testsuite/ChangeLog
2012-07-30  Keith Seitz  <keiths@redhat.com>

	* gdb.linespec/ls-errs.exp: Check some quote-enclosed
	linespecs.

[-- Attachment #2: lex_number-quotes.patch --]
[-- Type: text/x-patch, Size: 1355 bytes --]

diff --git a/gdb/linespec.c b/gdb/linespec.c
index 3d7f62f..51994c8 100644
--- a/gdb/linespec.c
+++ b/gdb/linespec.c
@@ -391,10 +391,11 @@ linespec_lexer_lex_number (linespec_parser *parser, linespec_token *tokenp)
     }
 
   /* If the next character in the input buffer is not a space, comma,
-     or colon, this input does not represent a number.  */
+     quote, or colon, this input does not represent a number.  */
   if (*PARSER_STREAM (parser) != '\0'
       && !isspace (*PARSER_STREAM (parser)) && *PARSER_STREAM (parser) != ','
-      && *PARSER_STREAM (parser) != ':')
+      && *PARSER_STREAM (parser) != ':'
+      && !strchr (linespec_quote_characters, *PARSER_STREAM (parser)))
     {
       PARSER_STREAM (parser) = LS_TOKEN_STOKEN (*tokenp).ptr;
       return 0;
diff --git a/gdb/testsuite/gdb.linespec/ls-errs.exp b/gdb/testsuite/gdb.linespec/ls-errs.exp
index 7db8ae4..5668397 100644
--- a/gdb/testsuite/gdb.linespec/ls-errs.exp
+++ b/gdb/testsuite/gdb.linespec/ls-errs.exp
@@ -146,6 +146,8 @@ add the_tests "$srcfile:3 foo" unexpected_opt "string" "foo"
 
 foreach x $invalid_offsets {
     add the_tests "$srcfile:$x" invalid_offset_f $x $srcfile
+    add the_tests "\"$srcfile:$x\"" invalid_offset_f $x $srcfile
+    add the_tests "'$srcfile:$x'" invalid_offset_f $x $srcfile
 }
 
 # Test invalid filespecs starting with function.

Re: [RFA] Fix number lexing, again

[Tom Tromey]

>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

Keith> While the list of growing terminations for the string is getting
Keith> tediously long, I don't think inverting the test would be any
Keith> shorter/clearer. WDYT?

I agree.
The patch is ok; so is the test suite patch.

Keith> PS. Also recommending this for 7.5.

Yes please.

Tom

Re: [RFA] Fix number lexing, again

[Keith Seitz]

On 07/30/2012 10:13 AM, Tom Tromey wrote:
>>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:
>
> Keith> While the list of growing terminations for the string is getting
> Keith> tediously long, I don't think inverting the test would be any
> Keith> shorter/clearer. WDYT?
>
> I agree.
> The patch is ok; so is the test suite patch.
>
> Keith> PS. Also recommending this for 7.5.
>
> Yes please.

All committed. Thank you!

Keith

Re: [RFA] Fix number lexing, again

[asmwarrior]

On 2012-7-31 2:05, Keith Seitz wrote:
> All committed. Thank you!
>
> Keith
>
Hi, Keith, I just build the latest CVS HEAD under MinGW+Msys, and I can confirm the fix, thank you.

Yuanhui Zhang