Re: [PATCH] Fix memory leaks in scanner

[Santtu Lakkala <inz inz fi>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 11/02/14 12:34, David King wrote:
> Hi Santtu
>
> On 2014-02-11 11:53, Santtu Lakkala <inz inz fi> wrote:
> > The Scan_Process_Fields_First_Letters_Uppercase() uses
> > g_utf8_strup in roman numeral handling, but does not free the
> > result.
>
> Thanks for the patch, I slightly modified the formatting and pushed
> it to master:
>
> https://git.gnome.org/browse/easytag/commit/?id=ae3dd7b81ba283ba0e9ea3d4aff326deedcc077e
Hmm, now that I think of it, maybe we should use g_ascii_strup() in
roman numeral handling to make it work correctly in Turkish locale
(see setlocale comment in ET_Add_File_To_File_List()).

Also the scanner code is riddled with assumption that strlen(x) ==
strlen(g_utf8_str{up/down}), which is wrong. (again, comment in
ET_Add_File_To_File_List()). This assumption was actually partly
introduced by me, when I did not know better.

I guess most of the scanner code should be rewritten, maybe using
GString and/or regexes where applicable to be totally utf-safe.

- -- 
Santtu Lakkala
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlMBwGQACgkQX9Rc0+po4p3s6QCfeVZCzU30r9OsbygaA40qpngg
5pYAnixgzyMsEgrZv4sda+jGhIRPhN5d
=NLPN
-----END PGP SIGNATURE-----