corrupt map file

[GoogleCodeExporter]

Here is a map file that I used in 0.2 (edgy eft deb files) and I opened it
with the SVN version of today. It is completely messed up. Links go in all
directions.

I corrected them. Cleaned it up. Closed labyrinth, opened it again and,
boom, corrupt again. It did save some changes, but it links stuff at random
it seems.

I guess some compatibility must have been broken? Have no such problems
with maps I created in the svn version. Advice appreciated :| well don't
worry this map was not really important but it should not happen too often.
I hope it can be useful to you to determine the cause.

Original issue reported on code.google.com by nekoh...@gmail.com on 10 Nov 2006 at 3:48

Attachments:

• b6fc1b1af810af9f6385b21fdaa24833ba73e2a9.map

[GoogleCodeExporter]

For example, "blogs" is completely messed up. Try unlinking it from all those 
things
on the other side of the map, then linking it to "méthodologie" and linking
"méthodologie" to "logiciels libres".

Chances are it will have changed next time you open it.

Original comment by nekoh...@gmail.com on 10 Nov 2006 at 3:52

[GoogleCodeExporter]

Oh cr.., it does not only happen with old maps, it happens with new ones too. 
As long
as they are complex enough. Could it be multiple links on a single thought 
would be
the problem?

Not very usable :|

Original comment by nekoh...@gmail.com on 10 Nov 2006 at 2:19

Attachments:

• psychologie.map

[GoogleCodeExporter]

The problem is with the "thought numbers". These are used to identify thoughts 
and
links between them when loading / saving.  Each thought has a number 
associated. 
When loading a map in older versions (0.2), the thought number was internally
incremented each time a thought was found.  This means that when a thought was
deleted, the map closed and reopened, there would be a duplicate thought number 
and
links would screw up on future loading / saving (as they'd go to the first 
created
thought with the correct number).  I fixed this a little while ago in trunk. 
Unfortunately, it appears in the big rewrite, this fix has been lost and the old
behavior has regained its place.  Grr.

Inside both files, there are duplicate thought numbers (first map: 37, 20, 25, 
second
map: 29, 30, 32, 33, 34, 36, 37, 39, 40) and so links are going to the older 
versions
of these.

I'm away right now, but when I return, I'll reapply the fix, which should 
hopefully
fix this properly once and for all.

Original comment by DonScor...@gmail.com on 10 Nov 2006 at 4:16

[GoogleCodeExporter]

Recommitted the fix in SVN.  Attached are the two map files, fixed (i.e. the
offending thought numbers changed to be unique).  I don't know about the links
between them, so I've left as-is.  If you delete the incorrect links and 
recreate
them correctly, they should now stick :)

Thanks.  Marking as Fixed.  If you still have problems (note that older maps may
still have issues.  Without some serious workarounds in the code this can't 
really be
helped :( ) with these maps or newly created maps, please add a comment and 
I'll look
again.

2006-11-12 Don Scorgie <DonScorgie@Blueyonder.co.uk>

        * src/MMapArea.py:
        * src/labyrinth.py:
        * src/Browser.py:
        Fix thought numbers on loading.  Again.  Sigh (issue #32
        and previously issue #6)
        Fix loading of maps from the command line

Original comment by DonScor...@gmail.com on 12 Nov 2006 at 9:59

• Changed state: Fixed

Attachments:

• b6fc1b1af810af9f6385b21fdaa24833ba73e2a9.map
• psychologie.map

[GoogleCodeExporter]

Thank you for fixing the sample maps ;) they work fine now. Hopefully corruption
problems like these should not happen too often anymore ^^

Original comment by nekoh...@gmail.com on 15 Nov 2006 at 5:59