It's the last day of the hackathon. The following ideas are written on
the white board (so far).
This is just a scratch pad where we noted some ideas, some
individually and some collectively. Not everything we discussed is
listed; and we didn't discuss all of these. These brief notes are not
meant to be self-explanatory, but just a reminder.
*X* = "important fix"
x = "potentially more-or-less bite-sized task"
move tracking *X*
conflict resolving *X*
log --search -> case-insensitive (-> utf8proc case folding) x
server-side log --search
server-side search
- file names (svn list ...) x
( - file content )
svnadmin load --normalize-revprops *X*
- normalize svn:* props
( - normalize mergeinfo )
svnadmin load --no-flush-to-disk x
svn ls --recursive performance x
( partial commit (one hunk out of many) (-> TortoiseSVN?) )
stash / snapshot / shelving x?
( checkpointing / offline commits )
remote pristines x?
( compressed pristines )
binary patches (what about binary props?) x
svnadmin dump --exclude/--include (-> like svnsync + authz) x
-> unreachable copies as adds
dump[filter]: exclude future copies x?
diff+patch w/ tree changes x?
diff -c<REV> across a copy x
repository upgrade to FSFS f7: *X*
- document ideal procedure for full dump/load with minimal down-time
- implement helper tools
authz:
- reuse/share model amongst connections = cache + lookup
- glob support ("wildcards")
- type-based (file/dir)
- traversal rights
- creation/deletion rights
FS:
- streamy log API
- node API
- enhance support for serf's path@lastChange pattern
- effect:
- eliminates DAG cache
- eliminates fs_id_t from latest API
- log in const memory
ra_session_reuse
kill "no-op changes" (above ~repos layer)
[Not written on the whiteboard, but also discussed:]
migrating the issue tracker (Issuezilla@Tigris to Jira@ASF)
- Julian
