Jacob Rief wrote:
Resetting the internal flags to a state equivalent to a decoder just after
creation can't work. Suppose the next header has not been yet fed to
LZ_decompress_write. You'll have to switch to a "search" state, then write
data until the header is found.
I assumed that the API client already forwarded its steam to the
beginning of the next member, which means that the first bytes written
using LZ_decompress_write(decoder, ...) are 'LZIP...'. That's the way
I did it for now.
This would unnecessarily skip a whole member if the next header has been
already written with LZ_decompress_write and the error happens before
the decoder reads the header from the internal buffer.
Same if the header has been partially written with LZ_decompress_write.
It seems to me that switching to a "search" state, writing data, and
letting the decoder find the header is the safest approach.
Regards,
Antonio.
_______________________________________________
Lzip-bug mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/lzip-bug
