[sugar] Trial-3 journal problem

Eben Eliason eben.eliason
Wed Sep 19 22:05:46 EDT 2007


Well, as I mentioned, you can do so for now by creating a new "clean"
copy each time you want to modify it.  If you're talking about example
or template files, then I would recommend keeping a clean copy within
the bundle so that you can create a new clean journal entry from that
at will.

Once versioning is in place, this will be a non-issue.  Unfortunately,
until then I don't think there's a better solution. Versioning is
really high on our list, though, and we hope to have it working soon.
It's obviously essential to the system in a number of ways, so we'd
much rather do it right than attempt more workarounds.

- Eben


On 9/19/07, Bert Freudenberg <bert at freudenbergs.de> wrote:
> What we need for etoys is a way to explicitly save a project, and
> then be able to resume from that point any number of times without
> fearing of it to be overwritten automatically.
>
> Any idea how to do that within the journal philosophy?
>
> - Bert -
>
> On Sep 19, 2007, at 19:59 , Eben Eliason wrote:
>
> > 1. Open activity instance A1
> > 2. Get implicit journal entry a1
> > 3. Press keep in A1, saved as journal entry a2
> >
> > (make changes to A1?)
> >
> > 4. Close A1, implicitly saved as a1'
> >
> > - Neither A1 or A2 are open, a1' and a2 are in Journal.
> >
> > 5. Open journal entry a2 as activity A2
> >
> > (make changes to A2?)
> >
> > 6. Close A2
> > 7. Get implicit journal entry a2'
> >
> > - Actually, We should have a1' and a2'.
> >
> > So, the key thing to note is that we're always "working on" some
> > activity, thus modifying it and storing it's latest state to the
> > Journal.  At step 4, a1' and a2 could actually be identical (unless
> > we've made further changes in A1 before closing).  In either case, we
> > have the "good version" you refer to as a2, as created by pressing
> > keep and having that stored state.  Any further adjustments to AI will
> > result in a1'' and so on.
> >
> > Likewise, opening a2 as A2 gives us a new working context, basically
> > like a branch, which starts at the stored state.  Of course, it will
> > still save it's own modifications implicitly just as A1 did (or any
> > activity does), resulting in a2', a2'' and so on.  Pressing the keep
> > button doesn't create a permanently static representation, but a
> > version snapshot that can later be resumed and modified at will.  In
> > order to indefinitely preserve a2 you'd have to press keep each time
> > you resumed it with the intention of making changes. (This problem, of
> > course, disappears when versioning lands, since this is exactly what
> > versioning does for us automatically.)
> >
> > - Eben
> >
> >
> >>
> >> - Bert -
> >>
> >> On Sep 19, 2007, at 18:14 , Eben Eliason wrote:
> >>
> >>> Hi Bert -
> >>>
> >>> I want to clarify this a bit to be sure I understand what's going
> >>> on:
> >>>
> >>> - Open activity instance A1
> >>> - Get implicit journal entry a1
> >>> - Press keep in A1, creating A2 and journal entry a2
> >>> - Close A2
> >>>
> >>> (neither A1 or A2 are open, a1 and a2 are in Journal)
> >>>
> >>> - Open journal entry a2 as activity A2
> >>> - Close A2
> >>> - Get implicit journal entry a2'
> >>>
> >>> (We SHOULD have a1 and a2' in Journal, neither open)
> >>>
> >>> Now, the problem as I read it is that the remaining Journal entries
> >>> are a2 and a2' instead of a1 and a2' as it should be.  Am I
> >>> understanding you correctly?
> >>>
> >>> - Eben
> >>>
> >>> On 9/19/07, Bert Freudenberg <bert at freudenbergs.de> wrote:
> >>>> Hi,
> >>>>
> >>>> for etoys we implemented the storing to the Journal as suggested,
> >>>> namely, on an explicit save (pressing keep button) we create a copy
> >>>> in the datastore, whereas when exiting, we update the existing
> >>>> entry.
> >>>> This is to workaround the missing versioning support.
> >>>>
> >>>> However, in testing we found a flaw in that scheme: When you resume
> >>>> an explicitly saved entry and then just exit the activity, it
> >>>> will do
> >>>> an implicit save and overwrite the saved one, resulting in
> >>>> losing the
> >>>> "good" copy.
> >>>>
> >>>> Is there any idea about how to work around this problem?
> >>>>
> >>>> - Bert -
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> Sugar mailing list
> >>>> Sugar at lists.laptop.org
> >>>> http://lists.laptop.org/listinfo/sugar
>
>
>
>
> _______________________________________________
> Sugar mailing list
> Sugar at lists.laptop.org
> http://lists.laptop.org/listinfo/sugar
>



More information about the Sugar-devel mailing list