Well I ran the tool sometime back so you will have to see if it has any filtering options. I believe the "official" story is to export everything then use svnadmin dump and filter that using the standard SVN ways. It was not recommended in the documentation, but I have done a partial backup and restore and exported just those projects. The documentation as I said warns against it but it worked for me. If the project is small you could try that. The checkins really need to be completely separate -- if there is any code sharing then you are going to run into problems. It probably worked for me because the directories I exported never referenced each other, separate groups with separate accounts in separate programming languages never touched the other folders.

Jason

David Blaikie wrote:
Yes, the error does appear to be coming from ssphys.exe.

Hm, then should I not be using DESTROY?  It seems that converting the
whole 40GB database would be a bear, likely to fail.  Do you have any
suggestions on how to proceed?  I was hoping if I isolated a small,
young project I could convert that.

When the perl script (responding to ssphys) tells me that a file is
corrupt, would it work to skip that file?

I'm actually at wits' end here - not frustrated, just flummoxed.  Any
hints, ideas or avenues of exploration much appreciated!

thanks

David

-----Original Message-----
From: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> [mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>] On Behalf Of Jason
Winnebeck
Sent: Wednesday, October 10, 2007 2:59 PM
To: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>; Vss2Svn Users
Subject: Re: Memory error

VSS corrupts often. However, the latest version of the source it does
well with. What corrupts are the history records and the versions you
pull out of history and I have seen this first hand on many instances.
The analyze tool really is finding data that is lost, so VSS is a
horrible tool to manage this (hence the benefits of SVN).

I wonder if the error comes from ssphys.exe, as I thought when Perl
programs fail they give an "exception"-like print.

Also I noticed when you DESTROY (delete permanently), it doesn't
actually remove anything from VSS. vss2svn actually recovers that. Just
an interesting side note. You might need to convert and use filters to
get just the data you need.

Jason

David Blinker wrote:
 > Hi, I am attempting to migrate a modest project from VSS to SVN.
 >
 >
 >     * Our VSS database DATA folder is approximately 40GB in size.
 >     * I am using DESTROY to strip away all but the target project
 >       (overnight),
 >     * then ANALYZE (at least 20 hours)
 >     * and vss2svn.pl to convert.
 >
 >
 > However, vss2svn.pl terminates with one of 2 errors:
 >
 >    1. file corrupted, cannot continue or
 >    2. memory error:  "The instruction at "0x28085bc1" referenced
memory
 >       at "0x000000004".  The memory could not be "written"."    (This
 >       last error occurs during "STARTING CACHE FOR PhysicalAction")
 >
 >
 >
 > Can you help me understand what's going on here and what I might do?
 >
 > Also, can you explain why VSS has so many corruptions (reams and reams

 > from ANALYZE) and yet never corrupts the source?  Our developers have
 > never reported any source corruption.
 >
 > David
 >
 >
 >
 > ----------------------------------------------------------------------
 > --
 >
 > _______________________________________________
 > vss2svn-users mailing list
 > Project homepage:
 > http://www.pumacode.org/projects/vss2svn/
 > Subscribe/Unsubscribe/Admin:
 > http://lists.pumacode.org/mailman/listinfo/vss2svn-users-lists.pumacod
> e.org <http://e.org> Mailing list web interface (with searchable archives):
 > http://dir.gmane.org/gmane.comp.version-control.subversion.vss2svn.use
 > r

_______________________________________________
vss2svn-users mailing list
Project homepage:
http://www.pumacode.org/projects/vss2svn/
Subscribe/Unsubscribe/Admin:
http://lists.pumacode.org/mailman/listinfo/vss2svn-users-lists.pumacode <http://lists.pumacode.org/mailman/listinfo/vss2svn-users-lists.pumacode>.
org
Mailing list web interface (with searchable archives):
http://dir.gmane.org/gmane.comp.version-control.subversion.vss2svn.user <http://dir.gmane.org/gmane.comp.version-control.subversion.vss2svn.user>

_______________________________________________
vss2svn-users mailing list
Project homepage:
http://www.pumacode.org/projects/vss2svn/
Subscribe/Unsubscribe/Admin:
http://lists.pumacode.org/mailman/listinfo/vss2svn-users-lists.pumacode.org
Mailing list web interface (with searchable archives):
http://dir.gmane.org/gmane.comp.version-control.subversion.vss2svn.user

Reply via email to