[Bugs] #1408 UNSP: Journal should behave gracefully on entries with unexpected (but syntactically valid) metadata
Sugar Labs Bugs
bugtracker-noreply at sugarlabs.org
Wed Sep 30 23:52:01 EDT 2009
#1408: Journal should behave gracefully on entries with unexpected (but
syntactically valid) metadata
------------------------------------------+---------------------------------
Reporter: garycmartin | Owner: tomeu
Type: defect | Status: assigned
Priority: Unspecified by Maintainer | Milestone: Unspecified by Release Team
Component: journal | Version: Git as of bugdate
Severity: Critical | Keywords: r?
Distribution: Unspecified | Status_field: Unconfirmed
------------------------------------------+---------------------------------
Comment(by alsroot):
Replying to [comment:7 garycmartin]:
> Replying to [comment:6 alsroot]:
> > Gary, could you test this patch in your broken(and not) environment,
> > patch looks invasive and there are doubts about including it to 0.86
>
> Sure. My quick tests so far are not so good :-( Here's the list:
>
> 1) Broken data-store item icons are still showing in the canvas area
with random colours, though the colour in the palette is grey (I see this
may be being resolved in a different ticket).
this random issue was moved to #1435
> 2) When scrolling up in the Journal I still commonly hit screen full's
of grey 0% progress bar, no title, no icon entries as per screen shot
(attaching another example). If scrolling up quickly Sugar can pause for a
while before displaying the grey 0% progress bar entries, though moving
the mouse cursor over the entries appears to redraw them correctly (i.e
these 0% entries are not broken data-store entries, just Journal getting
confused after it scrolls up past a broken data-store entry). Note there
is no issue scrolling down, only up.
>
> 3) The broken DS entries are now titled "Untitled" but the date is
listed as 40 years, 9 months ago (01/01/70). The Journal is listing them
in order correctly so I assume the Journal is getting a valid date from
somewhere for the entry order, perhaps "Unknown" should be the date text
if it's not possible to rescue a real date?
>
> 4) The palette menu is now accessible, so the erased menu item can be
selected, unfortunately selecting erase has no effect (and no tracebacks I
can find) from either the palette menu or from the details view.
>
> 5) The palette allows Start, Start with..., Copy, Send to..., and View
Details. None of which have any effect except View Details which just
shows an empty octet-stream 01/01/70 set of details. Perhaps all these
palette options (except a working erase) should be inactive? FWIW "Start
with..." does correctly say "no activity to start entry" so perhaps
"Start" at least could be disabled via the same mechanism (see screen
shot)?
I guess thats not a good idea after all,
to suppress not valid(that don't have minimal set of properties) ds
entries. Not sure if it could be some kind of filtering on ds side as
well. So, maybe having ds recover feature?
--
Ticket URL: <http://bugs.sugarlabs.org/ticket/1408#comment:8>
Sugar Labs <http://sugarlabs.org/>
Sugar Labs bug tracking system
More information about the Bugs
mailing list