Opened 4 years ago

Closed 3 years ago

#1424 closed defect (fixed)

Syndie freenet support seemingly

Reported by: hikiko Owned by:
Priority: minor Milestone: 0.9.27
Component: apps/syndie Version: 0.9.17
Keywords: freenet Cc:
Parent Tickets:

Description

This diff seems to add freenet support to syndie, the gui updates on error or completion, it doesnt update per file(nor are the files uploaded per file).

Uses the same fcp method as before and fixes a little of what seemed like URI mangling that was going on. Added extra checks for possible fcp communication error as well.

Subtickets

Attachments (6)

syndie-freenet.diff (11.1 KB) - added by hikiko 4 years ago.
sorry its early... now the proper diff
syndie-freenet.2.diff (10.9 KB) - added by hikiko 4 years ago.
Fixed white space, added client closing and made the UI inform you that they are being pushed as well as finished pushing
syndie-freenet.3.2.diff (12.7 KB) - added by hikiko 4 years ago.
syndie-freenet.3.diff (12.7 KB) - added by hikiko 4 years ago.
Change broke generate genKey in gui. Also some one kindly pointed out readResults could possibly toss results.
syndie-freenet.4.diff (21.1 KB) - added by hikiko 4 years ago.
FreenetHelper.java (11.2 KB) - added by hikiko 4 years ago.
seemed wisest

Download all attachments as: .zip

Change History (16)

comment:1 follow-up: Changed 4 years ago by killyourtv

Thanks for looking at this.

Does this patch supercede the gist you pasted in #i2p-dev a few days ago? Note that you're also including unreleated changes to build.properties in each patch.

comment:2 in reply to: ↑ 1 ; follow-up: Changed 4 years ago by hikiko

Replying to killyourtv:

Thanks for looking at this.

Does this patch supercede the gist you pasted in #i2p-dev a few days ago? Note that you're also including unreleated changes to build.properties in each patch.

Yes this one supersedes that one, this adds UI updating and proper error checking for FCPv2. Fixed the diff sorry about that. I have just been mtn diff > file in my build environment.

Last edited 4 years ago by hikiko (previous) (diff)

Changed 4 years ago by hikiko

sorry its early... now the proper diff

comment:3 in reply to: ↑ 2 Changed 4 years ago by killyourtv

Replying to hikiko:

Fixed the diff sorry about that. I have just been mtn diff > file in my build environment.

It is/was fine, I just wanted to let you know about it. I could have also edited the patch myself, of course. :)

Thanks again for giving Syndie some love. I'll be testing this soon, but no, I still hadn't gotten around to testing the original one (other than making sure it compiled).

comment:4 follow-up: Changed 4 years ago by zzz

There's a lot of unnecessary whitespace changes in the diff that make it much more difficult to review.

comment:5 in reply to: ↑ 4 Changed 4 years ago by hikiko

Replying to zzz:

There's a lot of unnecessary whitespace changes in the diff that make it much more difficult to review.

Sorry about that when I'm back at home base later I'll put together a clean one that was just the source I was working out of when I was figuring out how it all pieced together.

Changed 4 years ago by hikiko

Fixed white space, added client closing and made the UI inform you that they are being pushed as well as finished pushing

comment:6 Changed 4 years ago by hikiko

Fixed white space changes and now the UI informs you that the files are being pushed rather than just informing you they are done being pushed. Also added bits to make sure the client connection closes on freenets end. Hope this is useful to you all.

Changed 4 years ago by hikiko

Changed 4 years ago by hikiko

Change broke generate genKey in gui. Also some one kindly pointed out readResults could possibly toss results.

comment:7 Changed 4 years ago by hikiko

Due to more investigation into how syndie works alongside freenet this fix is kinda like putting a band-aid on a broken arm, sorry for the wasting of time.

More has to be done, there needs to be a entire Freenet syndie shared server tool/make the freenetpusher tool again that seems to just exist in comments now. The regular user probably shouldn't be creating and posting a normal shared index but rather a index of all local post and all local post every insert(maybe using a series of redirects on upload as well).

Other wise it just doesn't really work at all except for a (very)limited number post. So it fixes it, but not really.

Sorry I rushed to submit this, Ill attempt to work at this more and make a reasonable working solution to these problems.

comment:8 Changed 4 years ago by hikiko

alright so processFreenetPost was for syndication originally as in to upload entire archive with a shared index, peer to peer doesn't really share messages as is in syndie or fetch messages with out indexes, for anything it seems. Every freenet user has to publish a index as well due to this because freenet can only pull messages, a syndication server would receive read addresses/keys from various users and simply watch them and every once in a while repost the entire archive. does this sound about right/acceptable to every one, That is what frost does they just use the WOT(which no ones sure is gonna survive purgedb4o as is) WOT is a service that shares read keys and available services in those key spaces.

So as is yes, the code is correct for a basic freenet users archive syncing.

And possibly it should be modified to build a push a shared-index as well because well, freenet users should probably be posting those to share new boards in a shared archive and such no matter what. In a set up like this as well a new user wouldn't even necessarily have to share their USK with the big shared archive as well in order to eventually appear on it with the main shared archive in this set up.

Shared archive/syndicate creators could use what ever service they wanted to include more USKs-freenet users in their archive directly or it could survive on allot of adhoc person to person even.

So tldr,

-So as is yes, the code is correct for a basic freenet users archive syncing.

-freenetPost/processFreenetPost is the next piece needed which is very similar to pushfreenet but it post entire archives(as the command is current says its self with only returning -1 ATM)

  • to make this more automagic but it would greatly increase dependencies would be to design this to use the WOT. This scenerio actually seems less desirable in some ways, as in increasing dependencies and adding another piece that could change or become outdated. But would make it more magic/automagic.This is what frost does and FMS btw.

comment:9 Changed 4 years ago by hikiko

  • Status changed from new to testing

Last adjustments. Freenetpost method seemed well, pointless, the best option to create a shared-index using buildindex and seting a 3 month limit on uploads, all doable and workable from the sync menu and text interface with out it.

Changed 4 years ago by hikiko

Changed 4 years ago by hikiko

seemed wisest

comment:10 Changed 3 years ago by zzz

  • Milestone changed from 0.9.18 to 0.9.27
  • Resolution set to fixed
  • Status changed from testing to closed
  • Type changed from enhancement to defect

syndie-freenet.4.diff and FreenetHelper?.java applied in 80ec445f3481bddacb17f561d409c56592256f5f to be 1.106/107.
Minor cleanups by me. Untested by me.

Note: See TracTickets for help on using tickets.