Hello John,
Wednesday, April 23, 2008, 6:45:59 AM, you wrote:
> Hi Marcus,
> Thanks for the response. I've gone ahead and added:
> #define YYFILL(n) { if(YYCURSOR >= YYLIMIT) return 0; }
> and:
> [^] {
> return 0;
> }
That should do it :-)
> For the missing ?> case. However, I didn't really understand what you
> meant wrt to the '/0' character. How exactly would I modify the ANYCHAR
> rule to negate this character? Just a side note, I modified my PHP rule
> and it did start working.
If you deal with zero '\0' terminated strings you need to have a special
rule to avoid reading beyond the buffer.
> Also, is there some documentation with regards to states in re2c?
In your local checkout you have an up-to-date manual in htdocs and you also
have an up-to-date man page. However there is no further stuff like lessons.
I hope thae there will be time to add lessons as soon as the state stuff is
finalized so that it is available when the next stable version 0.14.0 will
be released.
marcus
> On Tue, Apr 22, 2008 at 06:14:39PM +0200, Marcus Boerger wrote:
>> Hello John,
>>
>> Friday, April 18, 2008, 8:46:38 AM, you wrote:
>>
>> > I've run into a segfault porting some rules over from flex/bison which
>> > used to work. Admittedly, I'm just scratching the surface and starting
>> > with re2c.
>>
>> > Relevant code:
>>
>> > #include <stdio.h>
>> > #include <stdlib.h>
>> > #include <string.h>
>> > char *scan(char *p)
>> > {
>> > /*!re2c
>> > re2c:define:YYCTYPE = "unsigned char";
>> > re2c:define:YYCURSOR = p;
>> > re2c:yyfill:enable = 0;
>> > re2c:yych:conversion = 1;
>> > re2c:indent:top = 1;
>> > re2c:define:YYMARKER = p;
>> > PLACEHOLDER = "<!!([^>]+)?>[^<]+</!!>";
>> > ANYCHAR = (.|[\n\t]);
>> > PHP = "<\?php"({ANYCHAR})+"?>";
>> > {PHP} {
>> > return "passed";
>> > }
>> > */
>> > }
>>
>> > void main() {
>> > char *test;
>> > test = malloc(4096);
>> > bzero(test,4096);
>> > strcpy(test,"<?php ... ?>");
>> > char *res = scan(test);
>> > printf("%s",res);
>> > free(test);
>> > }
>>
>> > Run with re2c -Fis
>> > re2c 0.13.4
>>
>> > I found that this works as expected if I drop the ? from <?php.
>>
>> The first thing here is the backslash in front of the ?. The second is that
>> you did not provide any code that is to be executed in case something does
>> not match. And last but not least there is a problem with unlimited
>> captures like in your case. The rule ANYCHAR+ has no limitation, hence
>> there will be calls to YYFILL to fill your buffer. If that results in a
>> reallocation you need to update all pointers re2c uses. Now in your case if
>> re2c does find the '<?php' but does not find the '?>' it will simply
>> continue beyond the end of the string. To avoid that you can provide a rule
>> for the '/0' character. Then either add that rule above the current one or
>> change ANYCHAR to not contain '/0'. The other solution is to provide a
>> check for YYFILL that stops matching at the end of the string.
>>
>> > valgrind:
>>
>> > ==14141== Invalid read of size 1
>> > ==14141== at 0x804850B: scan (<stdout>:50)
>> > ==14141== by 0x80485B5: main (lang.r:28)
>> > ==14141== Address 0x4169028 is 0 bytes after a block of size 4,096 alloc'd
>> > ==14141== at 0x4020641: malloc (in
>> > /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
>> > ==14141== by 0x8048579: main (lang.r:25)
>>
>>
>>
>> --
>> Best regards,
>> Marcus mailto:[EMAIL PROTECTED]
>>
--
Best regards,
Marcus mailto:[EMAIL PROTECTED]
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Re2c-general mailing list
Re2c-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/re2c-general
