[vorbis-dev] PATCH: utf8 implementation for oggenc (Was: UTF-8 in comments)

[Andy]

Hello Daniel,

Monday, March 19, 2001, you wrote:

> On Sun, Mar 18, 2001 at 11:57:19AM +1100, Michael Smith wrote:
>> >
>> >Don't bother about that now. Just create that wrapper func/lib/whatever now 
>> >so that it's there, and implent UTF-8 conversion for those systems that 
>> >support it. Others can just write the tags unconverted like now, for now. 
>> >Then, as you find ways to get this working on those other systems, you can 
>> >update the wrapper for those systems. This shouldn't be a problem really 
>> >since there's already oggs with "normal" text in the tags, so that would have 
>> >to be supported anyway..
>> 
>> Well, plain-ascii text is already UTF-8, since it's a strict subset
>> of UTF-8. So for 'normal' text, there's no problem. However, other
>> things won't be supported, ever, and shouldn't be. 
>> 
>> Functionality needs to exist on all platforms, not just some. At the 
>> VERY LEAST, all the widely-used unixes and win32 would be required
>> before I'd add the code.
>> 

> Ok, my two months old baby has been nice to me and i've been able to code
> away on the UTF-8 problem :) I have implemented a generic utf8 conversion
> routine suitable for converting strings before putting them in
> comment-fields. My idea is to use the routine as a fallback if iconv() is not
> available or don't have the needed features (an iconv implementation can
> conform to spec without supporting for example UTF-8) Features:

> * portable (no external library is needed, charsets are statially compiled
> into the binary)
> * simple (~200 lines of code, nothing obfuscated I think)
> * compatible licence (GPL, or whatever the vorbis-tools maintainers choose
> in the future)

> Since this is my first try on actually coding something more than 3 line
> patches in C apart from short school exercises some 5 years ago the code
> might need some going through and restructuring. Please let me hear your
> comments and I will to my best to fix any problems.

> I think it would be better to have *.TXT and make_code_map.pl in a
> subdirectory, could someone with better automake knowledge please enlighten
> me on how to do that?

> I include two charsets for testing purposes (new charsets are trivial to
> add at a cost of about .5 kb each)

> The patch is against vorbis-tools-1.0beta4, since i can't get automake
> dependency handling working it might be neccesary to do a 'cd oggenc; make
> charsetmap.h' to get it to work. I will be looking into this tomorrow.

> cheers/daniel

Call me crazy, but the code sequence:

+       if (!strcasecmp(encoding, "UTF-8")) {
+               /* ideally some checking of the given string should be done */
+               *to = malloc(strlen(from) + 1);
+               strcpy(from, *to);
+               return 0;
+       }

Looks pretty messed up...

I don't think this would have happened at all if "from" had been
declared "const char*" rather than just "char*".

When something is const, declaring it as such can be a boon.

I didn't review the code much further, this one was too easy :-)

Best regards,
 Andy

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.