Is the Subversion import code supposed to work with incremental delta
input?  I find it crashes right away.

My understanding is it looks in the temporary xfiles table to get the
uuid (therefore rid) of existing files (on a given branch), then gets
the blob against which the delta is to be applied, then runs the delta
to get the new blob.  The crash is due to the source blob being empty,
yet the delta code assumes otherwise.  The source blob is empty because
the rid came up zero, i.e. the default value when performing the
aforementioned query.

gdb lets me make queries into the temporary tables by using the p
command to call db_int() or whatever, which is pretty dang cool.  It
didn't take long to see that xfiles is empty.

To fix, xfiles, xbranches, and xrevisions should be initially populated
with the contents of the repository so incremental imports build off the
result of previous imports.  For xrevisions to work, revision tagging
must have been enabled in the previous imports, so even the first import
must have been treated as an incremental import.

I don't see why this wouldn't be doable.  Before I dive in, I figured
I'd post about it and see if anyone has any comments.

-- 
Andy Goth | <andrew.m.goth/at/gmail/dot/com>

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Reply via email to