On Wed, Jan 14, 2009 at 3:45 AM, Steve Borho <[email protected]> wrote:
> On Tue, Jan 13, 2009 at 7:56 AM, TK Soh <[email protected]> wrote:
>>
>> On Mon, Jan 12, 2009 at 4:58 PM, Steve Borho <[email protected]> wrote:
>> > On Mon, Jan 12, 2009 at 8:16 AM, TK Soh <[email protected]> wrote:
>> >>
>> >> On Mon, Jan 12, 2009 at 2:32 AM, Steve Borho <[email protected]> wrote:
>> >> > Hi TK, I've heard you mention a couple of times you've done some
>> >> > preliminary work on this feature.  Is this available for pulling?
>> >>
>> >> The work need some updating due the changes tortoisehg has seen it I
>> >> last worked on it. I will try update it in a couple of days and share
>> >> it with the team.
>> >
>> > I look forward to seeing this.  Let me know if you need a hand.
>>
>> Here's the patch queue:
>>
>>   http://bitbucket.org/tksoh/tortoisehg-hgshelve/
>>
>> Please give it a try and we can discuss it later. Do keep in mind
>> right now very much a prototype with the basic support for
>> shelve/unshelve.
>
> Finally figured out how to coax the patch queue out of bitbucket (using
> download link)
> Is that supposed to be done easier somehow?
>
> Likes:
>   hunk selection is very nice, well done
>
> Dislikes:
>   no way to unshelve a single file

One of the TODO is to add support to view shelved patches. We can
probably the let user unshelve selected hunks there.

I am also interested in avoid to display shelved hunks alongside the
unshelved ones. But this can be a challenge. Plus some potential
tricky corner case that need to be handled.

>   should it unshelve a file after it has been committed?  this would be more
> record-like

I don't think so. Record works by shelving the changes before
committing [the remaining ones], so it must unshelve after the commit.
In our case, shelve is initiated by user, so unshelve should be so
too.

> Non-sequitors:
>   Does the MS column need to be visible when the working copy has one
> parent?

Turning is on and off will likely create more confusion. And we will
need more screenshot of the same window for the documentation.

>   Having 'unknowns' visible at startup should be a sticky option.  You risk

We want to avoid making sticky option selectively. If we are going to
make 'unknown' sticky, then all the other options should be too.

> forgetting to add/commit new files without it.  You may also miss 
> copies/renames.

I am not sure if the risk is a valid one. An add/copy/rename/....
action is a result of an intention. If the user wants to add, then
they will try look for the file in the list, and learn why it's not
there.

>   What do you think of Qct's policy of pre-selecting files that Mercurial
> would normally
>   commit by default (AMR)?

It should be this way. In fact all FILE specified at argv should be
selected. It seemed strange to have to select the files again, if I
already choose to add them via, either the explorer context menu, or
hgtk add.

I thought Peer has volunteered to work on this quite awhile back.
Maybe I'm mistaken, Peer?

------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Tortoisehg-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tortoisehg-develop

Reply via email to