Fixed 1260, but this now gives rise to another bug: you see person bookmarks
in the eg the event bookmark menu.
I think this is because actions are added, whereas the code should delete
and reconstruct actions instead, not catching the AttributeError.
You are right about 1261, a minor issue though. I also don't know, but there
is a dirty flag somewhere, that must be set on change to trigger a save. I
should have to hunt it down in the code myself.
2007/9/29, James G. Sack (jim) <jgsack@...>:
> I have been working on
> 0001187: Runaway Gramps XML file sizes
> and traced the problem to an obvious and easily fixable bug, which I
> have submitted a patch for.
> Developers may wish to note the bug and my suggested fix.
> __init__(self, val=):
> ##self.v = val # wrong -- v is an alias
> self.v = list(val) # want a copy (not an alias)
> In this case, multiple instances got the same (initially empty) list,
> and (eg) insertions ended up showing on every instance's list!
> But, in the process, I have uncovered 2 new problems which I have
> recorded as new bugs
> 0001260: previously saved person bookmarks not shown on bookmark view
> 0001261: bookmarks data change insufficient to force database write
> On 1260,, I have traced the problem to the statement
> GrampsDb/_GrampsXMLDB in function load()
> (at line 61 in r8631)
> The 'set' ends up clearing data read from the input file.
> ==> I don't know what the purpose of the above code is, so I don't know
> what is wrong or how to fix it. Can anybody comment on this?
> On 1261, I suspect there is an established mechanism for forcing the
> database write on exit, but I haven't even looked for that yet.
> ==> Can somebody give advice on the correct thing to do to signify that
> an update is required on exit?
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> Gramps-bugs mailing list