TODO 4.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293
  1. f{,data}sync() usage could be optimized by batching the calls.
  2. make SSL (connect) timeouts produce a bit more than "Unidentified socket error".
  3. automatically resume upon transient errors, e.g. "connection reset by peer"
  4. or timeout after some data was already transmitted.
  5. possibly also try to handle Exchange's "glitches" somehow.
  6. add support for IMAP UTF-7 (for internationalized mailbox names).
  7. uidvalidity lock timeout handling would be a good idea.
  8. should complain when multiple Channels match the same folders.
  9. propagate folder deletions even when the folders are non-empty.
  10. - verify that "most" of the folders in the Channel are still there.
  11. - refuse to delete unpropagated messages when trashing on the remote side.
  12. - refuse to delete master if it has unpropagated messages. symmetry?
  13. add message expiration based on arrival date (message date would be too
  14. unreliable). MaxAge; probably mutually exclusive to MaxMessages.
  15. add alternative treatments of expired messages. ExpiredMessageMode: Prune
  16. (delete messages like now), Keep (just don't sync) and Archive (move to
  17. separate folder - ArchiveSuffix, default .archive).
  18. add support for event notification callbacks.
  19. it would be also possible to report more differentiated exit codes, but
  20. that seems too limiting in the general case.
  21. make it possible to have different mailbox names for Master and Slave in
  22. Patterns.
  23. - use master:slave for the pattern
  24. - for quoting, use more colons: the longest sequence of colons is the
  25. separator
  26. - this makes Groups mostly useless, as they are mostly a workaround for this
  27. function being missing so far
  28. - this is needed for move detection, which would work only within one Channel
  29. kill the concept of an INBOX, it is a relic from single-channel operation.
  30. if somebody needs it, he can have two stores with different Paths. the path
  31. can name a single (in-)box (curr. broken with maildir). an empty box name
  32. actually means empty, so the IMAP mailbox should use INBOX for Path (can't
  33. make that the default, as it would mess up the NAMESPACE).
  34. add regexp-based mailbox path rewriting to the drivers. user would provide
  35. expressions for both directions. every transformation would be immediately
  36. verified with the inverse transform. PathDelimiter and Flatten would become
  37. special cases of this.
  38. add daemon mode. primary goal: keep imap password in memory.
  39. also: idling mode.
  40. parallel fetching of multiple mailboxes.
  41. TLS session resumption becomes interesting then as well.
  42. imap_set_flags(): group commands for efficiency, don't call back until
  43. imap_commit().
  44. add streaming from fetching to storing.
  45. handle custom flags (keywords).
  46. make use of IMAP CONDSTORE extension (rfc4551; CHANGEDSINCE FETCH Modifier);
  47. make use of IMAP QRESYNC extension (rfc5162) to avoid SEARCH to find vanished
  48. messages.
  49. use MULTIAPPEND and FETCH with multiple messages.
  50. create dummies describing MIME structure of messages bigger than MaxSize.
  51. flagging the dummy would fetch the real message. possibly remove --renew.
  52. note that all interaction needs to happen on the slave side probably.
  53. don't SELECT boxes unless really needed; in particular not for appending,
  54. and in write-only mode not before changes are made.
  55. problem: UIDVALIDITY change detection is delayed, significantly complicating
  56. matters.
  57. some error messages are unhelpful in non-verbose mode due to missing context.
  58. possibly use ^[[1m to highlight error messages.
  59. consider alternative approach to trashing: instead of the current trash-before-
  60. expunge done by mbsync, let MUAs do the trashing (as modern ones typically do).
  61. mbsync wouldn't do any trashing by itself, but should track the moves for
  62. optimization. additionally, there should be a mode to move trashed messages to
  63. the remote store. TrashMode Internal|External, AbsorbRemoteTrash.
  64. a yet different approach to trashing is treating the trash like a normal mailbox.
  65. however, this implies a huge working set.
  66. consider optional use of messages-id (and X-GM-MSGID):
  67. - detection of message moves between folders
  68. - recovery from loss of sync state, migration from other tools