mirror of
https://github.com/samba-team/samba.git
synced 2025-11-16 20:23:50 +03:00
ace9990060
It is still far from being usable in samba4 but I want to commit so that the work does not get lost by mistake. This is also a good way to get comments if somebody is interested. Sorry Derrell I ended up rewriting large parts of the code but I find this style much more readable. Thanks for the hard work done. Your work was a good reference for me. ah the current code also shows some good numbers sqlite3 generic test: uid search took 0.05 seconds real 0m12.492s user 0m0.492s sys 0m0.345s with tdb we still get better numbers: uid search took 0.46 seconds real 0m0.892s user 0m0.360s sys 0m0.468s but most of the time is spent in adding operations and I think there's still a lot of space for improvement. Simo.
trees.ps contains an explanation of the Genealogical Representation of Trees in Databases which is being used in ldb_sqlite3. Note that we use fgID representation with 4 bytes per level, so we can represent 6.5E+08 subclasses of any object class. This should be adequate for our purposes. :-) The following document is the primary basis for the schema currently being used here: http://www.research.ibm.com/journal/sj/392/shi.html