changegroup: compute seen files as changesets are added (
issue4750)
Before this patch, addchangegroup() would walk the changelog and compute
the set of seen files between applying changesets and applying
manifests. When cloning large repositories such as mozilla-central,
this consumed a non-trivial amount of time. On my MBP, this walk takes
~10s. On a dainty EC2 instance, this was measured to take ~125s! On the
latter machine, this delay was enough for the Mercurial server to
disconnect the client, thinking it had timed out, thus causing a clone
to abort.
This patch enables the changelog to compute the set of changed files as
new revisions are added. By doing so, we:
* avoid a potentially heavy computation between changelog and manifest
processing by spreading the computation across all changelog additions
* avoid extra reads from the changelog by operating on the data as it is
added
The downside of this is that the add revision callback does result in
extra I/O. Before, we would perform a flush (and subsequent read to
construct the full revision) when new delta chains were created. For
changelogs, this is typically every 2-4 revisions. Using the callback
guarantees there will be a flush after every added revision *and* an
open + read of the changelog to obtain the full revision in order to
read the added files. So, this increases the frequency of these
operations by the average chain length. In the future, the revlog
should be smart enough to know how to read revisions that haven't been
flushed yet, thus eliminating this extra I/O.
On my MBP, the total CPU times for an `hg unbundle` with a local
mozilla-central gzip bundle containing 251,934 changesets and 211,065
files did not have a statistically significant change with this patch,
holding steady around 360s. So, the increased revlog flushing did not
have an effect.
With this patch, there is no longer a visible pause between applying
changeset and manifest data. Before, it sure felt like Mercurial was
lethargic making this transition. Now, the transition is nearly
instantaneous, giving the impression that Mercurial is faster. Of course,
eliminating this pause means that the potential for network disconnect due
to channel inactivity during the changelog walk is eliminated as well.
And that is the impetus behind this change.
hgrc settings command line options final result
quiet verbo debug quiet verbo debug quiet verbo debug
0 False False False False False False -> False False False
1 True False False False False False -> True False False
2 False True False False False False -> False True False
3 True True False False False False -> False False False
4 False False True False False False -> False True True
5 True False True False False False -> False True True
6 False True True False False False -> False True True
7 True True True False False False -> False True True
8 False False False True False False -> True False False
9 True False False True False False -> True False False
10 False True False True False False -> True False False
11 True True False True False False -> True False False
12 False False True True False False -> True False False
13 True False True True False False -> True False False
14 False True True True False False -> True False False
15 True True True True False False -> True False False
16 False False False False True False -> False True False
17 True False False False True False -> False True False
18 False True False False True False -> False True False
19 True True False False True False -> False True False
20 False False True False True False -> False True False
21 True False True False True False -> False True False
22 False True True False True False -> False True False
23 True True True False True False -> False True False
24 False False False True True False -> False False False
25 True False False True True False -> False False False
26 False True False True True False -> False False False
27 True True False True True False -> False False False
28 False False True True True False -> False False False
29 True False True True True False -> False False False
30 False True True True True False -> False False False
31 True True True True True False -> False False False
32 False False False False False True -> False True True
33 True False False False False True -> False True True
34 False True False False False True -> False True True
35 True True False False False True -> False True True
36 False False True False False True -> False True True
37 True False True False False True -> False True True
38 False True True False False True -> False True True
39 True True True False False True -> False True True
40 False False False True False True -> False True True
41 True False False True False True -> False True True
42 False True False True False True -> False True True
43 True True False True False True -> False True True
44 False False True True False True -> False True True
45 True False True True False True -> False True True
46 False True True True False True -> False True True
47 True True True True False True -> False True True
48 False False False False True True -> False True True
49 True False False False True True -> False True True
50 False True False False True True -> False True True
51 True True False False True True -> False True True
52 False False True False True True -> False True True
53 True False True False True True -> False True True
54 False True True False True True -> False True True
55 True True True False True True -> False True True
56 False False False True True True -> False True True
57 True False False True True True -> False True True
58 False True False True True True -> False True True
59 True True False True True True -> False True True
60 False False True True True True -> False True True
61 True False True True True True -> False True True
62 False True True True True True -> False True True
63 True True True True True True -> False True True