Re: crdt.el collaborative real time editing for Emacs

[Qiantan Hong]

Thanks! Actually, it will be incredibly useful if you can get a dump
of CRDT ID state. To do this, M-: (set-buffer-multibyte nil) in 
*Message* (to make it unibyte, CRDT ID are arbitrary binary strings
in general and multibyte buffer does some unicode conversion stuff),
and then M-: (crdt--dump-ids) in the CRDT server/client buffers. This
basically dump out the complete state of CRDT IDs and
should be easy to spot inconsistency between dumps on different
emacsen. Note that it might be very long and get truncated by *Message*.
M-: (setq some-var (crdt--dump-ids)) in the CRDT buffer and then
M-: (insert some-var) in a unibyte buffer is a workaround. Or
any other method that you prefer.

> If undo is done on connected computer, the undo is not shown on the
> server.
Yep haven’t work on undo-redo at all. Any behavior can happen.

> On Oct 5, 2020, at 4:07 AM, Jean Louis <bugs@gnu.support> wrote:
> 
> Hello Qiantan,
> 
> I have the Ethernet switch and few computers in house, I can test
> features.
> 
> * Qiantan Hong <qhong@mit.edu> [2020-10-05 09:45]:
>> Hi Jean,
>> 
>> Thanks again for the report! Currently the code is at very
>> alpha stage so sorry for any bugs in it.
> 
> That is natural.
> 
>> Could you provide a bit more detail on the yanking issue?
> 
> Yes, I have been testing that, first it appeared to work, then when
> yanked from server, it did not appear on the connected computer.
> 
> Then I have saved file on server. Then I have yanked again into file,
> it did not appear on connected computer.
> 
>> Did it just failed for one particular yank operation or all yanking
>> doesn’t work?
> 
> Now I cannot say why it happens, I say it happens.
> 
> If undo is done on connected computer, the undo is not shown on the
> server.
> 
>> I’m more worried about the former case — which means some
>> bug lies in the algorithm implementation and it’s really hard to
>> reproduce — haven’t reproduce it yet.
> 
>> It would be also helpful in this case that you can provide some
>> description of operations happened before this bug is triggered.
> 
>> But if it’s the latter then it might just be some other package
>> interference with the change hook or the change hook is not
>> installed correctly, either case, much easier to fix.
> 
> Sure, I will do that little later, when I have computers one to each
> other, right now they are separate, I have to walk from room to room
> and do action by action.
> 
> What means crdt.el ?
> 
> Many times of editing are collaborative editing, to say that it is
> only collaborative would not be proper, it is better to say "real time
> collaboration" and even better "simultaneous editing":
> https://en.wikipedia.org/wiki/Simultaneous_editing
> 
> The mode should be named in easier fashion, something that is easy for
> users to enable or turn off.
> 
> Later I will give you more detailed transcript what I did, the crdt.el
> leaves some messages too, I can collect those as well. 
> 
> Jean

smime.p7s
Description: S/MIME cryptographic signature