On Wednesday 26 August 2009 18:58:52 Michael Yip wrote: > Matthew Toseland wrote: > > On Wednesday 26 August 2009 01:38:29 Daniel Cheng wrote: > > > >> On Tue, Aug 25, 2009 at 11:44 PM, Ian Clarke<i...@locut.us> wrote: > >> > >>> On Tue, Aug 25, 2009 at 8:51 AM, Matthew > >>> Toseland<t...@amphibian.dyndns.org> wrote: > >>> > >>>>> It's not a joke but I admit I misunderstood the specification. I was > >>>>> wondering how to detect the end of a AllData reply and I thought there > >>>>> would be something such as "EndMessage". > >>>>> > >>>> No, it ends with the word "Data" on a line by itself and is followed by > >>>> the data. > >>>> > >>> To clarify, the plaintext part of the message ends with "Data" on a > >>> line by itself. > >>> > >>> It is followed by the raw data. It is not necessary to specify when > >>> the raw data ends since its length is provided in the plaintext part > >>> of the message. > >>> > >> The current implementation would be confused if you do this: > >> > >> ClientPut<CR> > >> blbarlbalr=...<CR> > >> blbarlbalr=...<CR> > >> DataLength=1<CR> > >> Data<CR> > >> <LF> > >> > >> This should insert a single byte file with <LF>, but it would get confused. > >> > > > > So the problem is we don't assume the client consistently uses the same > > linefeed marker? File a bug. > > > > Err... this problem started after I've misread the spec. Easy solution > is to parse the DataLength value and just read the number of bytes as > the DataLength value?
Yes, if there are trailing bytes there is always a DataLength value. Some messages sent from the client to the node have multiple DataLength's (with different prefixes), but any message sent to the node will have only one, without a prefix. We always send a newline after "Data", and the data will start after that.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list Devl@freenetproject.org http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
