* bbackde at googlemail.com <bbackde at googlemail.com> [2007-09-26 17:49:59]:
> And because you think that you became aggressive immediately? No, if I > would want to blackmail you because I'm too silly to implement your > TestDDA I would have said "don't make it mandatory". I didn't mean to be rude... I apologize if I was. > > But the point is (and I feel you don't want to accept this because its > your "baby") that the alternate proposal is more straight forward. Well, first of all it's not my baby. I didn't have the original idea and the implementation has been discussed : you even took part to the debate! By the way I'm not convinced that bothering with testdda is worth it (and haven't ever been). I implemented it so that toad would focus on other, more important things (his concern was securing freenet on a multi-user system so that running gateways would be possible; testdda is part of the security infrastructure). > Isn't it true that it is easier for clients to handle the TestDDA in > one row with the original request, instead of having to resend the > original request after processing the TestDDA? The current implementation allows that. If you want to implement a dumb, simple client you just have to either use direct transfers or implement FileHash. Both don't require any tracking client-side. > The TestDDA doesn't > really fit into all of the other defined node requests, because it is > not tied to a specific request, but a request cannot be done without > TestDDA. Thats why I thought about an alternate solution. The original > implementation should be kept so clients can act like Thaw: just > sending TestDDA before each request to get the authorization for DDA. > This way Thaw doesn't have to handle TestDDA during the request... > I don't think that thaw is behaving as it should atm > Btw: checking the wiki I didn't found any clue what is send from the > node when the source directory for a ClientPUT is read only and the > node can't create the file that the client must read. > The client replies nothing/whatever it wants ... and the test will fail. > On 9/26/07, Florent Daigni?re <nextgens at freenetproject.org> wrote: > > * Matthew Toseland <toad at amphibian.dyndns.org> [2007-09-26 15:01:18]: > > > > > On Tuesday 25 September 2007 23:24, Florent Daigni?re wrote: > > > > * Matthew Toseland <toad at amphibian.dyndns.org> [2007-09-25 20:22:24]: > > > > > > > > > On Tuesday 25 September 2007 17:07, bbackde at googlemail.com wrote: > > > > > > My point is not to change the principles of the TestDDA > > > > > > implementation. The proposal describes a different way of > > > > > > implementation that fits for stateless clients. The current > > > > > > implementation requires that you send out-of-order testdda requests > > > > > > if > > > > > > the node sends an error, and then the client have to resend the > > > > > > original request. The proposal ties the testdda to the initiating > > > > > > request, and the node remembers the request until the testdda is > > > > > > finished. > > > > > > > > > > The basic principle here seems sound. Making clients' life a bit > > > > > easier is > > > > > generally a good thing. Nextgens? > > > > > > > > Send a patch or fill in a ticket on the BTS. We discussed it 6 months > > > > ago > > > > > > > (http://archives.freenetproject.org/message/20070414.083225.647d5e15.en.html) > > > > suggestions would have been welcome then but aren't anymore . I ended up > > > > implementing what we agreed on and have no plan to spend any time > > > > on that in the near future. By the way if you really want to make a > > > > basic, simple client I suggest you compute the FileHash and send it > > > > everytime. > > > > > > I wasn't asking you to implement it, merely for your opinion on the idea. > > > I > > > will file a bug. > > > > Well that's what I understood ... If frost doesn't support TestDDA we > > are screwed since it's mandatory. > > > > NextGen$ > > > > -----BEGIN PGP SIGNATURE----- > > Version: GnuPG v1.4.6 (GNU/Linux) > > > > iD8DBQFG+ncsU/Z/dHFfxtcRAuyqAKDMywqBjD8vBXGpjPHVndPv7+wdgwCgx9l+ > > 3qYxt5ZkiA15D1xUlnviMlI= > > =UOlf > > -----END PGP SIGNATURE----- > > > > _______________________________________________ > > Devl mailing list > > Devl at freenetproject.org > > http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl > > > > > -- > __________________________________________________ > GnuPG key: (0x48DBFA8A) > Keyserver: pgpkeys.pca.dfn.de > Fingerprint: > 477D F057 1BD4 1AE7 8A54 8679 6690 E2EC 48DB FA8A > __________________________________________________ -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20070926/71e8afe1/attachment.pgp>
