On 07.01.2008, at 22:31, Daniel Stenberg wrote:

>> Indeed, "negative" ranges are broken with both versions. I'll
>> investigate that. Oh, and I was wrong in another regard: -9 is
>> *not* supposed to mean the first 10 characters but the last 9.
>> Consider the testcase withdrawn as completely bogus.

> Isn't test case 1020 testing a case that should work as it works
> like that with FTP and HTTP?

Yes, but the (curl) range syntax is so obscure to me that I haven't
thought about the special cases carefully.

>> data->state.resume_from += (curl_off_t)file_stat.st_size;

> Sorry for being thick, but are you suggesting the current code is
> fine or not?

I'm suggesting that it's not fine ATM because we're seeking to the end
of the file here (st_size) when range -Y was specified instead of
st_size-Y
but I don't have any more time to work on this right now.

> If not, can you please provide an updated patch?

If no one gets around I'll do a new patch with new test cases based on
your version by the end of the week or so.

Servus,
       Daniel
Received on 2008-01-08