theora dev - May 2003 - gcc 3.2 warnings

This is mostly for derf, but others may be interested too. Attached are 
the warnings (stderr log) from compling theora under gcc 3.2.3 
(Debian). There quite a few. Many of them are spurious, but some 
indicate real bugs. In particular it seems likely we have more 
unitialized variables since Dan reports differing encoder output 
depending on the compiler used.

In general, we need to go through the warnings and look for real 
problems on whatever platforms we have access to. gcc, msvc and 
metrowerks codewarrior are a good start. It's good to remove them when 
we can even if they're not bugs to clean up the output, but note that 
by policy we disregard some, such as those warning about using standard 
C implicit type conversion.

  -r

<p>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: warn.log
Type: application/octet-stream
Size: 10559 bytes
Desc: warn.log
Url :
http://lists.xiph.org/pipermail/theora-dev/attachments/20030521/30b2dd3f/warn-0001.obj

What is strange in your log is that it does not catch the only case where a
variable is definately used without being initialized. It is

GrpSad

in scan.c, RowSadScan() method. The others look like bogus warnings.

What worries me is that VC at least catches lots of conversion warnings,
some with possible loss of data. Most are probably ok, but these:

'double' to '_int64'
'double' to '_int32'
'_int64' to 'long'
'_int32' to 'unsigned char'

look suspicious to me. Can anyone comment if these are ok, and the compiler
will do the right thing?

Regards,
Mauricio Piacentini
Tabuleiro

<p><p><p>----- Original Message -----
From: "Ralph Giles" <giles@xiph.org>
To: <theora-dev@xiph.org>
Sent: Wednesday, May 21, 2003 11:42 AM
Subject: [theora-dev] gcc 3.2 warnings

<p>> This is mostly for derf, but others may be interested too.
Attached are
> the warnings (stderr log) from compling theora under gcc 3.2.3
> (Debian). There quite a few. Many of them are spurious, but some
> indicate real bugs. In particular it seems likely we have more
> unitialized variables since Dan reports differing encoder output
> depending on the compiler used.
>
> In general, we need to go through the warnings and look for real
> problems on whatever platforms we have access to. gcc, msvc and
> metrowerks codewarrior are a good start. It's good to remove them when
> we can even if they're not bugs to clean up the output, but note that
> by policy we disregard some, such as those warning about using standard
> C implicit type conversion.
>
>   -r
>
>
--- >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
'theora-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.