eol: don't fallback to use .hgeol from tip (BC)
If no .hgeol were found in the current working directory, eol would fallback to
use the one in tip. That could in some cases give very confusing or wrong
behaviour when it applied wrong filters.
It might be convenient to have plain 'clone' immediately apply 'native'
encoding patterns in the cloned repo. But it is wrong to assume that this
revision is tip, and even more wrong to also apply it when not cloning - for
example when updating between history revisions. The encoding should always
match the content of the current .hgeol . It should never use anything else.
eol: tweak test-eol-clone.t with better descriptions and logging
Expose impact of changes coming next ...
eol: fix update - don't use and apply removed .hgeol patterns
'hg up -C' to revisions with different .hgeol patterns could leave dirty
changes in the working directory. That could make deployment of new .hgeol
filters tricky: they would "occasionally" apply also in branches where they
shouldn't.
Fixed by dropping all old patterns before applying new ones.
eol: cache needs update, also if it has same timestamp as the source
Ignoring same timestamp could (in theory?) cause changes to not be detected.
It might happen quite often that the cache is populated right after .hgeol has
been updated and they thus have the same time stamp second. But we want
correctness, and if it populates the cache so fast, then it can also not be a
big problem to run it again next time when the timestamp has moved on.
localrepo: fix variable binding in handling of old filters
The lambda was referencing oldfn in outer scope without binding the current
value. If oldfn function were reassigned before use, wrong filters could be
used.
Fixed by having oldfn as named parameter default value of the lambda.
eol: test-eol-update.t coverage around update --clean using filters ... badly
This will reveal problems and track their fixes.
copies: drop the findlimit logic
We don't use the limit anymore so we should stop computing that limit.
I did not bother measuring the potential performance gain. I am assuming that
not running any code will be faster that doing some computation and not using
the result.
pathcopies: give up any optimization based on `introrev`
Between
8a0136f69027 and
d98fb3f42f33, we sped up the search for the
introduction revision during path copies. However, further checking show that
finding the introduction revision is still expensive and that we are better off
without it. So we simply drop it and only rely on the linkrev optimisation.
I ran `perfpathcopies` on 6989 pair of revision in the pypy
repository (`hg perfhelper-pathcopies`. The result is massively in favor of
dropping this condition. The result of the copy tracing are unchanged.
Attempt to use a smaller changes preserving linkrev usage were unsuccessful, it
can return wrong result. The following changesets broke test-mv-cp-st-diff.t
- if not f.isintroducedafter(limit):
+ if limit >= 0 and f.linkrev() < limit:
return None
Here are various numbers (before this changeset/after this changesets)
source destination before after saved-time ratio
worth cases
e66f24650daf 695dfb0f493b 1.062843 1.246369 -0.183526 1.172675
c979853a3b6a 8d60fe293e79 1.036985 1.196414 -0.159429 1.153743
22349fa2fc33 fbb1c9fd86c0 0.879926 1.038682 -0.158756 1.180420
682b98f3e672 a4878080a536 0.909952 1.063801 -0.153849 1.169074
5adabc9b9848 920958a93997 0.993622 1.147452 -0.153830 1.154817
worse 1%
dbfbfcf077e9 aea8f2fd3593 1.016595 1.082999 -0.066404 1.065320
worse 5%
c95f1ced15f2 7d29d5e39734 0.453694 0.471156 -0.017462 1.038488
worse 10%
3e144ed1d5b7 2aef0e942480 0.035140 0.037535 -0.002395 1.068156
worse 25%
321fc60db035 801748ba582a 0.009267 0.009325 -0.000058 1.006259
median
2088ce763fc2 e6991321d78b 0.000665 0.000651 0.000014 0.978947
best 25%
915631a97de6 385b31354be6 0.040743 0.040363 0.000380 0.990673
best 10%
ad495c36a765 19c10384d3e7 0.431658 0.411490 0.020168 0.953278
best 5%
d13ae7d283ae 813c99f810ac 1.141404 1.075346 0.066058 0.942126
best 1%
81593cb4a496 99ae11866969 1.833297 0.063823 1.769474 0.034813
best cases
c3b14617fbd7 743a0fcaa4eb 1101.811740 2.735970 1099.075770 0.002483
c3b14617fbd7 9ba6ab77fd29 1116.753953 2.800729 1113.953224 0.002508
058b99d6e81f 57e249b7a3ea 1246.128485 3.042762 1243.085723 0.002442
9a8c361aab49 0354a250d371 1253.111894 3.085796 1250.026098 0.002463
442dbbc53c68 3ec1002a818c 1261.786294 3.138607 1258.647687 0.002487
As one can see, the average case is not really impacted. However, the worth case
we get after this changeset are much better than the one we had before it. We
have 30 pairs where improvements are above 10 minutes.
This reflect in the combined time for all pairs
before: 26256s
after: 1300s (-95%)
If we remove these pathological 30 cases, we still see a significant improvements:
before: 1631s
after: 1245s (-24%)
perf: introduce a `--contains` flag to the `perfdirstate` command
The new flag benchmark a large amount of `filepath in dirstate` call. This will
be useful to compare the Python and Rust implementation of the dirstatemap.
perf: introduce a `--iteration` to `perfdirstate`
This flag benchmark an iteration over all the file in the dirstate. This
will be useful to compare the Python and the Rust implementation of the
dirstate.