Re: [Libcdio-devel] [RFC] New API iso9660_statv2_t as API/ABI compatible way to read files >= 4 GiB

[Pete Batard]

On 2018.07.09 22:05, Thomas Schmitt wrote:
>      lsn_t extent_lsn;
>      ...
>
>      extent_lsn = from_733 (p_iso9660_dir->extent);
>
>      /* Important: Test here with previous .total_size */
>      if (p_stat->total_size > 0) {
>        /* This is a follow-up extent. Check for a gap. */
>        if (p_stat->lsn + p_stat->total_size / ISO_BLOCKSIZE != extent_lsn
>            || p_stat->total_size % ISO_BLOCKSIZE) {
>
>          /* >>> Gap detected. Throw error. <<< */;
>
>          return NULL;
>        }
>      } else {
>        /* No extent with data before. Record this start as overall start. */
>        p_stat->lsn = extent_lsn;
>      }
>
>      /* Only now update .total_size */
>      p_stat->total_size += from_733 (p_iso9660_dir->size);

Ah, OK. I see now, thanks.

Then, as a pragmatic person, I vote for the "only add total_size" 
approach, as this is the only one we really need, and we can address any 
reports of gaps *when* people can provide a real-life example of a 
gapped ISO.

As long as there's an error message on gap detected, and an error 
thrown, I'm onboard with this.

The only thing I'd like to see is a planned deprecation of size (through 
DO_NOT_WANT_COMPATIBILITY), so that downstream developers get some hint 
that they should update their code.

Regards,

/Pete