TODO

   1 # vim:sw=2:sts=2:
   2 TODO
   3 ====
   4
   5 Legend:
   6 - [ ] not started
   7 - [-] in-progress
   8 - [x] done
   9 - [~] cancelled
  10
  11 In-progress
  12 -----------
  13
  14 - [-] Convert to Typed Racket
  15   - [x] build executable (otherwise too-slow)
  16   - [-] add signatures
  17     - [x] top-level
  18     - [ ] inner
  19     - [ ] imports
  20 - [-] commands:
  21   - [x] c | crawl
  22     Discover new peers mentioned by known peers.
  23   - [x] r | read
  24     - see timeline ops above
  25   - [ ] w | write
  26     - arg or stdin
  27     - nick expand to URI
  28     - Watch FIFO for lines, then read, timestamp and append [+ upload].
  29       Can be part of a "live" mode, along with background polling and
  30       incremental printing. Sort of an ii-like IRC experience.
  31   - [ ] q | query
  32     - see timeline ops above
  33     - see hashtag and channels above
  34   - [x] d | download
  35     - [ ] options:
  36       - [ ] all - use all known peers
  37       - [ ] fast - all except peers known to be slow or unavailable
  38         REQUIRES: stats
  39   - [x] u | upload
  40     - calls user-configured command to upload user's own timeline file to their server
  41   Looks like a better CLI parser than "racket/cmdline": https://docs.racket-lang.org/natural-cli/
  42   But it is no longer necessary now that I've figured out how to chain (command-line ..) calls.
  43 - [-] Output formats:
  44   - [x] text long
  45   - [x] text short
  46   - [ ] HTML
  47   - [ ] JSON
  48 - [-] Peer discovery
  49   - [-] parse peer refs from peer timelines
  50     - [x] mentions from timeline messages
  51       - [x] @<source.nick source.url>
  52       - [x] @<source.url>
  53     - [ ] "following" from timeline comments: # following = <nick> <uri>
  54   - [ ] Parse User-Agent web access logs.
  55   - [-] Update peer ref file(s)
  56     - [x] peers-all
  57     - [x] peers-mentioned
  58     - [ ] peers-followed (by others, parsed from comments)
  59     - [ ] peers-down (net errors)
  60     - [ ] redirects?
  61     Rough sketch from late 2019:
  62         let read file =
  63             ...
  64         let write file peers =
  65             ...
  66         let fetch peer =
  67             (* Fetch could mean either or both of:
  68              * - fetch peer's we-are-twtxt.txt
  69              * - fetch peer's twtxt.txt and extract mentioned peer URIs
  70              * *)
  71             ...
  72         let test peers =
  73             ...
  74         let rec discover peers_old =
  75             let peers_all =
  76                 Set.fold peers_old ~init:peers_old ~f:(fun peers p ->
  77                     match fetch p with
  78                     | Error _ ->
  79                         (* TODO: Should p be moved to down set here? *)
  80                         log_warning ...;
  81                         peers
  82                     | Ok peers_fetched ->
  83                         Set.union peers peers_fetched
  84                 )
  85             in
  86             if Set.empty (Set.diff peers_old peers_all) then
  87                 peers_all
  88             else
  89                 discover peers_all
  90         let rec loop interval peers_old =
  91             let peers_all = discover peers_old in
  92             let (peers_up, peers_down) = test peers_all in
  93             write "peers-all.txt"  peers_all;
  94             write "peers-up.txt"   peers_up;
  95             write "peers-down.txt" peers_down;
  96             sleep interval;
  97             loop interval peers_all
  98         let () =
  99             loop (Sys.argv.(1)) (read "peers-all.txt")
 100
 101 Backlog
 102 -------
 103 - [ ] user-agent file as CLI option - need to run at least the crawler as another user
 104 - [ ] Support fetching rsync URIs
 105 - [ ] Check for peer duplicates:
 106   - [ ] same nick for N>1 URIs
 107   - [ ] same URI for N>1 nicks
 108 - [ ] Background polling and incremental timeline updates.
 109   We can mark which messages have already been printed and print new ones as
 110   they come in.
 111   REQUIRES: polling
 112 - [ ] Polling mode/command, where tt periodically polls peer timelines
 113 - [ ] nick tiebreaker(s)
 114   - [ ] some sort of a hash of URI?
 115   - [ ] angry-purple-tiger kind if thingie?
 116   - [ ] P2P nick registration?
 117     - [ ] Peers vote by claiming to have seen a nick->uri mapping?
 118           The inherent race condition would be a feature, since all user name
 119           registrations are races.
 120           REQUIRES: blockchain
 121 - [ ] stats
 122   - [ ] download times per peer
 123 - [ ] Support redirects
 124   - should permanent redirects update the peer ref somehow?
 125 - [ ] Support time ranges (i.e. reading the timeline between given time points)
 126 - [ ] optional text wrap
 127 - [ ] write
 128 - [ ] timeline limits
 129 - [ ] peer refs set operations (perhaps better done externally?)
 130 - [ ] timeline as a result of a query (peer ref set op + filter expressions)
 131 - [ ] config files
 132 - [ ] highlight mentions
 133 - [ ] filter on mentions
 134 - [ ] highlight hashtags
 135 - [ ] filter on hashtags
 136 - [ ] hashtags as channels? initial hashtag special?
 137 - [ ] query language
 138 - [ ] console logger colors by level ('error)
 139 - [ ] file logger ('debug)
 140 - [ ] Suport immutable timelines
 141   - store individual messages
 142     - where?
 143       - something like DBM or SQLite - faster
 144       - filesystem - transparent, easily published - probably best
 145   - [ ] block(chain/tree) of twtxts
 146     - distributed twtxt.db
 147     - each twtxt.txt is a ledger
 148     - peers can verify states of ledgers
 149     - peers can publish known nick->url mappings
 150     - peers can vote on nick->url mappings
 151     - we could break time periods into blocks
 152     - how to handle the facts that many(most?) twtxt are unseen by peers
 153     - longest X wins?
 154
 155 Done
 156 ----
 157 - [x] Dedup read-in peers before using them.
 158 - [x] Prevent redundant downloads
 159   - [x] Check ETag
 160   - [x] Check Last-Modified if no ETag was provided
 161     - [x] Parse rfc2822 timestamps
 162 - [x] caching (use cache by default, unless explicitly asked for update)
 163   - [x] value --> cache
 164   - [x] value <-- cache
 165     REQUIRES: d command
 166 - [x] Logger sync before exit.
 167 - [x] Implement rfc3339->epoch
 168 - [x] Remove dependency on rfc3339-old
 169 - [x] remove dependency on http-client
 170 - [x] Build executable
 171   Implies fix of "collection not found" when executing the built executable
 172     outside the source directory:
 173
 174     collection-path: collection not found
 175       collection: "tt"
 176       in collection directories:
 177       context...:
 178        /usr/share/racket/collects/racket/private/collect.rkt:11:53: fail
 179        /usr/share/racket/collects/setup/getinfo.rkt:17:0: get-info
 180        /usr/share/racket/collects/racket/contract/private/arrow-val-first.rkt:555:3
 181        /usr/share/racket/collects/racket/cmdline.rkt:191:51
 182        '|#%mzc:p
 183
 184
 185 Cancelled
 186 ---------
 187 - [~] named timelines/peer-sets
 188   REASON: That is basically files of peers, which we already support.