Getting offlineimap to maintain a specific maildir structure

[Sebastian Spaeth]

"Rogério Brito" <rbrito at ime.usp.br> wrote:

>Hi, Sebastian.
>
>First of all, thank you very much for the reply. You don't really
>believe
>how it is apreciated by me.
>
>I was already holding (in Debian-speak) the offlineimap package at
>version
>6.3.4 on all of my machines.
>
>On Aug 25 2012, Sebastian Spaeth wrote:
>> Rogério Brito <rbrito at ime.usp.br> writes:
>> > What I do is: I map gmail's INBOX to the toplevel maildir and,
>then, all 
>> > the other folders get put inside this maildir, with the names
>prefixed by 
>> > a dot and with the separator being a dot.
>> 
>> > | nametrans		= lambda name: re.sub('^', '.', re.sub('^(\[Gmail
>> > \]/|INBOX)', '', name))
>> 
>> > From what I understand, I disabled the bidirectional syncing with 
>> > createfolders=false (in an attempt to mimic as much as possible the
>
>> > behavior of 6.3.4), but I get the following error message:
>> 
>> > | Creating folder .[local]
>> > | ERROR: Aborting sync, folder '.' [acc: 'email']
>> > |   getfolder() asked for a nonexisting folder '.'.
>> 
>> Mmmh, this has nothing to do with nametrans or folder creation, I
>> think. It has something to do with how OfflineImap treats the very
>> top-level folder and there is a bug somewhere.
>
>Yes, I guessed that, but not really sure of what is/was going on.
>
>Just so you know my intention, my goal is to create in my local maildir
>a
>folder structure in a way that I have traditionally used with imap
>servers
>like courier and dovecote.
>
>Basically, the only point remaining is the conversion of gmail's INBOX
>to
>the top level folder, which should be the main maildir.
>
>> You have configured your nametrans rule, so that the GMAIL "INBOX"
>> effective becomes folder ".".
>
>Yes, precisely.
>
>> When we walk the local maildir folders checking if folder "." exists,
>we
>> don't find a folder "." and try to create folder "." which fails.
>
>I'm not really sure. Tried to read the code, but gave up with some
>other
>obligations.
>
>> I will have to check how we deal with this, I believe there were some
>> changes that might have caused this.
>
>If you want, I git bisected to see which commit introduced the change
>where
>things broke:
>
>,----[ git bisect log ]
>| git bisect start
>| # good: [61e50b3b1a81a564697a74384f528b22dfd32d8b] v6.3.4
>| git bisect good 61e50b3b1a81a564697a74384f528b22dfd32d8b
>| # bad: [010941e12e9caaacf312cdb1a8149a1ec2c3e5a6] Merge branch 'next'
>| git bisect bad 010941e12e9caaacf312cdb1a8149a1ec2c3e5a6
>| # bad: [6f361c4d9a25f092fc2dc81b68bdba03c8b4ca67] Don't output
>"Finished in x seconds" in quiet ui
>| git bisect bad 6f361c4d9a25f092fc2dc81b68bdba03c8b4ca67
>| # good: [ee75e0921ff105d19f619973dba188fc6900ff45] Remove 'config' as
>parameter from BaseFolder & derivatives
>| git bisect good ee75e0921ff105d19f619973dba188fc6900ff45
>| # bad: [93b05215fba364564018c52ab257a323c25277d4] Condense
>ui.connecting() function a bit
>| git bisect bad 93b05215fba364564018c52ab257a323c25277d4
>| # skip: [eeef8f4bab057e1e73ab7d1d17898643dee003aa] Don't ask for
>hostname if using a tunnel
>| git bisect skip eeef8f4bab057e1e73ab7d1d17898643dee003aa
>| # skip: [48fdb1441892e1fdade1e13cd5679755bee53a81] Don't output
>thread ID in log
>| git bisect skip 48fdb1441892e1fdade1e13cd5679755bee53a81
>| # bad: [953c58a9c917f467d7256945e207e72f59227a13] Robustify (&fix)
>error throwing on APPEND
>| git bisect bad 953c58a9c917f467d7256945e207e72f59227a13
>| # bad: [3157a8d793876d050bff29a3830c26a43ed8a78a]
>repository.Gmail.getfolder()
>| git bisect bad 3157a8d793876d050bff29a3830c26a43ed8a78a
>| # good: [19ff636390a06fd6c82f6087cb1647e00d34856c] Properly output
>errors when the main thread receives some
>| git bisect good 19ff636390a06fd6c82f6087cb1647e00d34856c
>| # good: [32ca20d0da97b2ed2b377c06f12f4365af37538c] Folder.Maildir: No
>need to store 'uid' in messagelist dict.
>| git bisect good 32ca20d0da97b2ed2b377c06f12f4365af37538c
>| # good: [792243c78f6138904863f60059db051a6c4e4934] Improve
>repository/Maildir.getfolder() to use cached objects
>| git bisect good 792243c78f6138904863f60059db051a6c4e4934
>| # bad: [a279aa7307ca036f93b660ade84099550ff1a4f2] Maildir: Call
>top-level directory '', not '.'
>| git bisect bad a279aa7307ca036f93b660ade84099550ff1a4f2
>`----
>
>Which leads me to:
>
>,----[ git bisect bad ]
>| a279aa7307ca036f93b660ade84099550ff1a4f2 is the first bad commit
>| commit a279aa7307ca036f93b660ade84099550ff1a4f2
>| Author: Sebastian Spaeth <Sebastian at SSpaeth.de>
>| Date:   Mon Sep 19 14:14:44 2011 +0200
>| 
>|     Maildir: Call top-level directory '', not '.'
>|     
>|     If nametrans translates to an empty directory we want to find the
>|     top-level directory by name '' and not by name '.'. This unbreaks
>|     nametrans rules that result in empty folder names.
>|     
>|     Signed-off-by: Sebastian Spaeth <Sebastian at SSpaeth.de>
>| 
>| :040000 040000 9f22f1d8516efbd5776b008ae2f560af746a88f4
>e0f61c67e11efbf034ebe29e89c7c17b6615099f M	offlineimap
>`----
>
>Given what is above, I tried both to change the nametrans to map INBOX
>to
>the simple dot or to the empty string. Neither worked.
>
>> Thanks for the reminder to look into this.
>
>Thanks for your work maintaining offlineimap.

Thanks for the bisect. I'll be able to look into this early next week.

Seb.
-- 
Sent from my tablet. Please excuse my brevity.