NEEDS TESTING!!! [was Re: [PATCH] Obviate the need for explicit 'remember_tree_state/restore_tree_state' calls]

[Dirk Hohndel]

Linus Torvalds <torvalds at linux-foundation.org> writes:
> This implements my previous commentary:
>
> Linus Torvalds <torvalds at linux-foundation.org> writes:
>>
>> We should probably aim to do the whole "restore tree state" in
>> dive_list_update_dives() unconditionally. Right now, almost all of the
>> callers do it, and the two remaining cases (main.c and webservice.c
>> should probably do it too.
>> 
>> Then, if we just made the collapse/expand callbacks update the
>> trip->expanded flag, we could remove all the "save/restore" state
>> entirely, and updating the divelist would just automatically maintain the
>> previous state.
>
> Needs more testing. It looks fine, it continues to remove more lines than 
> it adds, and it passes my trivial tests. But more people should check.

I totally agree. It looks sane, it seems to work, but it REALLY REALLY
REALLY needs testing.

Sadly almost all of our testers are way ahead of Linus and I (timezone
wise), so if you are still awake (or already awake), please get the
latest git and test trip modifications and expanded/collapsed state of
trips when changing sort order, etc.

I'd also love to see "tested and it works ok for me" emails to get a
====================================================================
feeling that this has, indeed, been tested.
===========================================


So if you are one of the ~35 people on this list who have never sent an
email and just lurked and tested... now is your chance!

Thanks

/D