bug-global
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: htags generates incorrect anchor when -s option is specified.


From: Shuichi KITAGUCHI
Subject: Re: htags generates incorrect anchor when -s option is specified.
Date: Mon, 29 May 2017 22:26:18 +0900 (JST)

Hi,

>  
> http://cvs.savannah.gnu.org/viewcvs/global/libparser/C.c?cvsroot=global&r1=1.14&r2=1.15
>  
> http://cvs.savannah.gnu.org/viewcvs/global/libparser/Cpp.c?cvsroot=global&r1=1.23&r2=1.24

I applied this patch and get correct result.
Thanks!

> Hi.
> 
>> 'ST1' is linked to '../R/1.html', but 'ST2' is linked to
>> '../S/2.html'. I think 'ST2' should be linked to '../R/2.html'.
> 
> This was caused by a bug in the built-in parser of gtags.
> 
> I have fixed the bug in the repository.
>  
> http://cvs.savannah.gnu.org/viewcvs/global/libparser/C.c?cvsroot=global&r1=1.14&r2=1.15
>  
> http://cvs.savannah.gnu.org/viewcvs/global/libparser/Cpp.c?cvsroot=global&r1=1.23&r2=1.24
> 
> Thank you for the report.
> 
> 
> On Mon, 29 May 2017 19:18:11 +0900 (JST)
> Shuichi KITAGUCHI <address@hidden> wrote:
> 
>> Hi,
>> 
>> > Hi,
>> > Could you please send information to do a reproduction test?
>> > o Which version of FreeBSD?
>> 
>> 11.0-RELEASE (from binary package) and 12-CURRENT.
>> 
>> > o What is the htags's command line?
>> 
>> -s -a -n -g
>> 
>> > o Where is the incorrect anchor link?
>> 
>> When prepare following two header files and execute htags
>> with above options,
>> 
>> [a.h]
>> struct st1 {
>>         ST1 *st1_ptr;
>> };
>> 
>> struct st2 {
>>         ST1 st2_st1;
>> };
>> 
>> struct st3 {
>>         ST2 st3_st2;
>> };
>> 
>> [b.h]
>> typedef struct st1      ST1;
>> typedef struct st2      ST2;
>> typedef struct st3      ST3;
>> 
>> HTML/S/2.html contains below lines.
>> 
>> <a id='L1' name='L1'></a>   1 <strong class='reserved'>typedef</strong> 
>> <strong class='reserved'>struct</strong> <a href='../S/1.html#L1' 
>> title='Defined at 1 in a.h.'>st1</a>      <a href='../R/1.html' 
>> title='Multiple referred from 3 places.'>ST1</a>;
>> <a id='L2' name='L2'></a>   2 <strong class='reserved'>typedef</strong> 
>> <strong class='reserved'>struct</strong> <a href='../S/1.html#L5' 
>> title='Defined at 5 in a.h.'>st2</a>      <a href='../S/2.html#L2' 
>> title='Defined at 2 in b.h.'>ST2</a>;
>> 
>> 'ST1' is linked to '../R/1.html', but 'ST2' is linked to
>> '../S/2.html'. I think 'ST2' should be linked to '../R/2.html'.
>> 
>> > Thank you in advance.
>> > 
>> > Regards,
>> > Shigio
>> 
>> Thanks,
>> 
>> > 2017-05-28 22:37 GMT+09:00 Shuichi KITAGUCHI <address@hidden>:
>> > 
>> >> Hi,
>> >>
>> >> I've found that htags command generates incorrect anchor link when -s
>> >> option is specified on FreeBSD and Cygwin. Linux is ok.  Upon
>> >> investigation, this seems to be caused by behaviour of qsort().
>> >>
>> >> When anchor->lineno is same and cmp()@htags/anchor.c returns 0, some
>> >> qsort() impelmentation might swap objects. This causes that type 'R'
>> >> is upper than type 'T', and anchor_get() cannot find correct anchor.
>> >>
>> >> If appling below patch to htags/anchor.c, htags seems generate correct
>> >> link.
>> >>
>> >>  static int
>> >>  cmp(const void *s1, const void *s2)
>> >>  {
>> >> +#if 1
>> >> +       int diff = ((struct anchor *)s1)->lineno - ((struct anchor
>> >> *)s2)->lineno;
>> >> +       /* when lineno is same, preserve current order. */
>> >> +       return diff == 0 ? -1 : diff;
>> >> +#else
>> >>         return ((struct anchor *)s1)->lineno - ((struct anchor
>> >> *)s2)->lineno;
>> >> +#endif
>> >>  }
>> >>  /*
>> >>   * Pointers (as lineno).
>> >>
>> >> Is this right fix?
>> >>
>> >> Thanks,
>> >>
>> >> _______________________________________________
>> >> Bug-global mailing list
>> >> address@hidden
>> >> https://lists.gnu.org/mailman/listinfo/bug-global
>> >>
>> > 
>> > 
>> > 
>> > -- 
>> > Shigio YAMAGUCHI <address@hidden>
>> > PGP fingerprint:
>> > 26F6 31B4 3D62 4A92 7E6F  1C33 969C 3BE3 89DD A6EB
>> 
>> _______________________________________________
>> Bug-global mailing list
>> address@hidden
>> https://lists.gnu.org/mailman/listinfo/bug-global
> 
> -- 
> 岩本 秀樹 <address@hidden>
> 



reply via email to

[Prev in Thread] Current Thread [Next in Thread]