On 04/01/2013 11:25 AM, C. Michael Pilato wrote: > On 03/31/2013 12:31 AM, Blair Zajac wrote: >> Question on svn_repos_load_fs4() and svn_repos_get_fs_build_parser4() (maybe >> other functions have a similar text): >> >> * @note If @a start_rev and @a end_rev are valid revisions, this >> * function presumes the revisions as numbered in @a dumpstream only >> * increase from the beginning of the stream to the end. Gaps in the >> * number sequence are ignored, but upon finding a revision number >> * younger than the specified range, this function may stop loading >> * new revisions regardless of their number. >> >> What does 'may stop' mean? Does it flips a coin ;) Seriously, will it or >> will it not stop, or under which conditions. > > Heh. It's been a long while -- I'll look into that again. On first (again) > glance, I don't see any code which would stop anything.
Okay, so the behavior is straightforward -- if a revision in the dump stream is outside the range, it is ignored (skipped) and processing continues. I remember now what I was thinking when I composed that docstring. I was thinking about how much time would be wasted trudging the whole way through a dump of, say, revisions 0-100000 when the load operation was limited to the revision range 0-10. Today, the API will keep examining the addition 99,990 revisions just in case one of them has a revision number within the desired range. I was building in some future-proofing to allow us to decide later that that's silly -- that the minute we see a revision that's younger than the desired range, we could bail out on processing the dumpstream immediately. Any strong opinions on which behavior we want for this API? -- C. Michael Pilato <cmpil...@collab.net> CollabNet <> www.collab.net <> Enterprise Cloud Development
signature.asc
Description: OpenPGP digital signature
