From: tomasfrydrych_at_yahoo.co.uk
Date: Mon Dec 15 2003 - 03:04:51 EST
Hi Leonard,
> > > Every time you save, the version number should be increased...
> >
> >I am not convinced about this.
>
> I am ;).
>
> Why shouldn't it increase the version number? You just
> edited the document, therefore it's a new version of the document.
>
> Think of "Save" as the equivalent of a "cvs checkin".
Except when you save the document, you also obliterate the previous
copy. Now that I think of it, though, you could have a backup program
running in the background while you are working on the doc, and then
the backed up doc would have the same version number as a later
save from the same session with different content.
The history section in the ABW doc looks something like this:
<history version="5" time="xxx" edit-time="xxx" uid="xxxxxxx">
<version id="1" time="xxx" edit-time="xxx" uid="xxxx" />
<version id="2" time="xxx" edit-time="xxx" uid="xxxx" />
<version id="3" time="xxx" edit-time="xxx" uid="xxxx" />
<version id="4" time="xxx" edit-time="xxx" uid="xxxx" />
<version id="5" time="xxx" edit-time="xxx" uid="xxxx" />
</history>
What I will do is to increase the version number (the id attribute of the
last version record) for each save, but will only keep one version
record for the entire ope/save session (so there will be gaps between
the version id's in the table). This will prevent the size of the history
record getting out of hand.
> What random number algorithm are you using ? Have you
> considered a one of the ones for cryptographic purposes?
We have UT_rand() function that Patrick Lam put in place a while back
based on the GNU C lib. You will find more about it in the cpp source
(http://www.abisource.com/lxr/source/abi/src/af/util/xp/ut_rand.cpp).
I am looking into adding some UT_UUID class for generation of proper
uuids; I do not like using rands in their place.
Tomas
This archive was generated by hypermail 2.1.4 : Mon Dec 15 2003 - 03:04:52 EST