On Fri, Jun 15, 2012 at 09:00:26PM +0200, Antonio Diaz Diaz wrote: > Hello Cezary, > > Cezary Sliwa wrote: > >A CD-R disc written in Track-At-Once mode has two unreadable run-out > >blocks at the end of the track. Linux seems to shrink the reported > >length of the track when reading the last sectors of a track fails. If > >this happens, ddrescue cannot be restarted (it reports that the input > >file is shorter than the log data). Actually, one may reinsert the > >disc before a restart, but imagine the disc is hardly recognizable... > > Thanks for reporting this. > > Is this the message ddrescue shows?: > "Bad logfile; last block begins past end of input file."
Yes. > > > >It might be a good idea to allow rescuing data from a file which is > >shorter than the log, even independently of the above issue. > > Ddrescue already allows this, but only when the rest of the logfile is > marked as non-tried. Maybe this requirement can be loosened to "there > are no blocks marked as done in the rest of the logfile". What is the purpose of such checks? One may have a partial (short) copy of the file to rescue he wants to merge in. Done blocks are OK in such a case. Just take care not to truncate the rescued file or the log. Cezary Sliwa > > > Regards, > Antonio. _______________________________________________ Bug-ddrescue mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-ddrescue
