This is an archive of the discontinued Mercurial Phabricator instance.

Differential D9307

copies-rust: parse the changed-file sidedata directly in rust
ClosedPublic

Authored by marmoute on Nov 12 2020, 10:17 AM.

Details

Reviewers
Alphare
SimonSapin
pulkit
Group Reviewers
hg-reviewers
Commits
rHGe0313b0a6f7e: copies-rust: parse the changed-file sidedata directly in rust
Summary

It does not make much sense to parse the data into python object using slow
python code to later turn them into rust object. We directly pass the binary
blob and use it directly in Rust.

Ideally we could directly read the sidedata in Rust, using a revlog in Rust.
However we do not have this ready to use yet.

This more direct approach provides a nice speedup over the board. Especially five
cases that we previously too slow to return in the previous changeset are not
able to finish.

Notably, we are now significantly faster than the Python version of this code in
all the meaningful cases.

I looked at the various cases that remains significantly slower then the filelog
version and they are currently 3 main source of slowness:

  • The isancestor computation: even if we cache them, if the revs spawn over a large amount of history the ancestry checking is still quite expensive. Using a different approach more centered on the graph we are currently considering might yield significant speed.
  • Merging of the map from the two parents: in some case, this climb up to ⅔ of the time spent in copy tracing. See inline comment for idea to handle this better.
  • Extracting data from the filelog. I would like to think this mostly comes from the fact my test repositories pre-date Valentin Gatien-Baron improvement of the files field (99ebde4fec99) and that more recent revisions will be faster to fetch. Further testing on this aspect is needed.

This revision compared to the previous one:

Repo Case Source-Rev Dest-Rev # of revisions old time new time Difference Factor time per rev

mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 1 revs, 0.000047 s, 0.000049 s, +0.000002 s, × 1.0426, 49 µs/rev
mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 6 revs, 0.000181 s, 0.000114 s, -0.000067 s, × 0.6298, 19 µs/rev
mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 1032 revs, 0.005852 s, 0.004223 s, -0.001629 s, × 0.7216, 4 µs/rev
pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 9 revs, 0.000229 s, 0.000305 s, +0.000076 s, × 1.3319, 33 µs/rev
pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 1 revs, 0.000058 s, 0.000060 s, +0.000002 s, × 1.0345, 60 µs/rev
pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 7 revs, 0.000146 s, 0.000173 s, +0.000027 s, × 1.1849, 24 µs/rev
pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 1 revs, 0.001206 s, 0.000446 s, -0.000760 s, × 0.3698, 446 µs/rev
pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 6 revs, 0.025275 s, 0.010360 s, -0.014915 s, × 0.4099, 1726 µs/rev
pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 4785 revs, 0.080303 s, 0.048002 s, -0.032301 s, × 0.5978, 10 µs/rev
pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 6780 revs, 0.152641 s, 0.075705 s, -0.076936 s, × 0.4960, 11 µs/rev
pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 5441 revs, 0.099107 s, 0.056705 s, -0.042402 s, × 0.5722, 10 µs/rev
pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 43646 revs, 2.137894 s, 0.794685 s, -1.343209 s, × 0.3717, 18 µs/rev
pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 26389 revs, 0.022202 s, 0.020209 s, -0.001993 s, × 0.9102, 0 µs/rev
pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 11316 revs, 0.228946 s, 0.122475 s, -0.106471 s, × 0.5350, 10 µs/rev
netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 2 revs, 0.000186 s, 0.000142 s, -0.000044 s, × 0.7634, 71 µs/rev
netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 2 revs, 0.000133 s, 0.000113 s, -0.000020 s, × 0.8496, 56 µs/rev
netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 3 revs, 0.000320 s, 0.000241 s, -0.000079 s, × 0.7531, 80 µs/rev
netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 9 revs, 0.001339 s, 0.000729 s, -0.000610 s, × 0.5444, 81 µs/rev
netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 1421 revs, 0.015694 s, 0.010198 s, -0.005496 s, × 0.6498, 7 µs/rev
netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 1533 revs, 0.018457 s, 0.015312 s, -0.003145 s, × 0.8296, 9 µs/rev
netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 5750 revs, 0.111691 s, 0.060517 s, -0.051174 s, × 0.5418, 10 µs/rev
netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 67005 revs, 1.166017 s, 0.611102 s, -0.554915 s, × 0.5241, 9 µs/rev
mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 2 revs, 0.000197 s, 0.000164 s, -0.000033 s, × 0.8325, 82 µs/rev
mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 8 revs, 0.000626 s, 0.000334 s, -0.000292 s, × 0.5335, 41 µs/rev
mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 9 revs, 0.000303 s, 0.000463 s, +0.000160 s, × 1.5281, 51 µs/rev
mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 7 revs, 0.001679 s, 0.000730 s, -0.000949 s, × 0.4348, 104 µs/rev
mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 3 revs, 0.006947 s, 0.003522 s, -0.003425 s, × 0.5070, 1174 µs/rev
mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.133070 s, 0.072518 s, -0.060552 s, × 0.5450, 12086 µs/rev
mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.008705 s, 0.005760 s, -0.002945 s, × 0.6617, 3 µs/rev
mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.005913 s, 0.005720 s, -0.000193 s, × 0.9674, 0 µs/rev
mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 7839 revs, 0.101373 s, 0.063310 s, -0.038063 s, × 0.6245, 8 µs/rev
mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 0.046526 s, 0.043608 s, -0.002918 s, × 0.9373, 0 µs/rev
mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 30263 revs, 0.313954 s, 0.204831 s, -0.109123 s, × 0.6524, 6 µs/rev
mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 153721 revs, 3.367395 s, 2.161906 s, -1.205489 s, × 0.6420, 14 µs/rev
mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 210546 revs, 4.691820 s, 3.291831 s, -1.399989 s, × 0.7016, 15 µs/rev
mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 2 revs, 0.001199 s, 0.001213 s, +0.000014 s, × 1.0117, 606 µs/rev
mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 2 revs, 0.001216 s, 0.001225 s, +0.000009 s, × 1.0074, 612 µs/rev
mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 4 revs, 0.000613 s, 0.000564 s, -0.000049 s, × 0.9201, 141 µs/rev
mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 2 revs, 0.001906 s, 0.001549 s, -0.000357 s, × 0.8127, 774 µs/rev
mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 1 revs, 0.092766 s, 0.035918 s, -0.056848 s, × 0.3872, 35918 µs/rev
mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.136074 s, 0.073788 s, -0.062286 s, × 0.5423, 12298 µs/rev
mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.009067 s, 0.006151 s, -0.002916 s, × 0.6784, 3 µs/rev
mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.006243 s, 0.006165 s, -0.000078 s, × 0.9875, 0 µs/rev
mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 6657 revs, 0.114463 s, 0.065421 s, -0.049042 s, × 0.5715, 9 µs/rev
mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 40314 revs, 0.433683 s, 0.313749 s, -0.119934 s, × 0.7235, 7 µs/rev
mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 38690 revs, 0.411278 s, 0.297867 s, -0.113411 s, × 0.7242, 7 µs/rev
mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 54487 revs, 0.155133 s, 0.111300 s, -0.043833 s, × 0.7174, 2 µs/rev
mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 0.048933 s, 0.046202 s, -0.002731 s, × 0.9442, 1 µs/rev
mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 97052 revs, 8.100385 s, 1.999640 s, -6.100745 s, × 0.2469, 20 µs/rev
mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 52031 revs, 1.446720 s, 0.809134 s, -0.637586 s, × 0.5593, 15 µs/rev
mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 363753 revs, killed , 47.406785 s, , , 130 µs/rev
mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 444327 revs, 1.369537 s, 0.996219 s, -0.373318 s, × 0.7274, 2 µs/rev
mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 362229 revs, killed , 47.273399 s, , , 130 µs/rev
mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 359344 revs, killed , 47.419099 s, , , 131 µs/rev
mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 192665 revs, 5.186079 s, 3.512653 s, -1.673426 s, × 0.6773, 18 µs/rev
mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 237259 revs, killed , 44.459049 s, , , 187 µs/rev
mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 391148 revs, killed , 52.837926 s, , , 135 µs/rev

This revision compared to the python code:

Repo Case Source-Rev Dest-Rev # of revisions Python-Time Rust-Time Difference Factor time per rev

mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 1 revs, 0.000044 s, 0.000049 s, +0.000005 s, × 1.1136, 49 µs/rev
mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 6 revs, 0.000138 s, 0.000114 s, -0.000024 s, × 0.8261, 19 µs/rev
mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 1032 revs, 0.005052 s, 0.004223 s, -0.000829 s, × 0.8359, 4 µs/rev
pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 9 revs, 0.000219 s, 0.000305 s, +0.000086 s, × 1.3927, 33 µs/rev
pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 1 revs, 0.000055 s, 0.000060 s, +0.000005 s, × 1.0909, 60 µs/rev
pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 7 revs, 0.000128 s, 0.000173 s, +0.000045 s, × 1.3516, 24 µs/rev
pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 1 revs, 0.001089 s, 0.000446 s, -0.000643 s, × 0.4096, 446 µs/rev
pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 6 revs, 0.017407 s, 0.010360 s, -0.007047 s, × 0.5952, 1726 µs/rev
pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 4785 revs, 0.094175 s, 0.048002 s, -0.046173 s, × 0.5097, 10 µs/rev
pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 6780 revs, 0.238009 s, 0.075705 s, -0.162304 s, × 0.3181, 11 µs/rev
pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 5441 revs, 0.125876 s, 0.056705 s, -0.069171 s, × 0.4505, 10 µs/rev
pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 43646 revs, 3.581556 s, 0.794685 s, -2.786871 s, × 0.2219, 18 µs/rev
pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 26389 revs, 0.016721 s, 0.020209 s, +0.003488 s, × 1.2086, 0 µs/rev
pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 11316 revs, 0.242367 s, 0.122475 s, -0.119892 s, × 0.5053, 10 µs/rev
netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 2 revs, 0.000165 s, 0.000142 s, -0.000023 s, × 0.8606, 71 µs/rev
netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 2 revs, 0.000114 s, 0.000113 s, -0.000001 s, × 0.9912, 56 µs/rev
netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 3 revs, 0.000296 s, 0.000241 s, -0.000055 s, × 0.8142, 80 µs/rev
netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 9 revs, 0.001124 s, 0.000729 s, -0.000395 s, × 0.6486, 81 µs/rev
netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 1421 revs, 0.013060 s, 0.010198 s, -0.002862 s, × 0.7809, 7 µs/rev
netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 1533 revs, 0.017112 s, 0.015312 s, -0.001800 s, × 0.8948, 9 µs/rev
netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 5750 revs, 0.660350 s, 0.060517 s, -0.599833 s, × 0.0916, 10 µs/rev
netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 67005 revs, 10.032499 s, 0.611102 s, -9.421397 s, × 0.0609, 9 µs/rev
mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 2 revs, 0.000189 s, 0.000164 s, -0.000025 s, × 0.8677, 82 µs/rev
mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 8 revs, 0.000462 s, 0.000334 s, -0.000128 s, × 0.7229, 41 µs/rev
mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 9 revs, 0.000270 s, 0.000463 s, +0.000193 s, × 1.7148, 51 µs/rev
mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 7 revs, 0.001474 s, 0.000730 s, -0.000744 s, × 0.4953, 104 µs/rev
mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 3 revs, 0.004806 s, 0.003522 s, -0.001284 s, × 0.7328, 1174 µs/rev
mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.085150 s, 0.072518 s, -0.012632 s, × 0.8517, 12086 µs/rev
mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.007064 s, 0.005760 s, -0.001304 s, × 0.8154, 3 µs/rev
mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.004741 s, 0.005720 s, +0.000979 s, × 1.2065, 0 µs/rev
mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 7839 revs, 0.190133 s, 0.063310 s, -0.126823 s, × 0.3330, 8 µs/rev
mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 0.035651 s, 0.043608 s, +0.007957 s, × 1.2232, 0 µs/rev
mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 30263 revs, 0.440694 s, 0.204831 s, -0.235863 s, × 0.4648, 6 µs/rev
mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 153721 revs, 18.454163 s, 2.161906 s, -16.292257 s, × 0.1172, 14 µs/rev
mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 210546 revs, 31.562719 s, 3.291831 s, -28.270888 s, × 0.1043, 15 µs/rev
mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 2 revs, 0.001189 s, 0.001213 s, +0.000024 s, × 1.0202, 606 µs/rev
mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 2 revs, 0.001204 s, 0.001225 s, +0.000021 s, × 1.0174, 612 µs/rev
mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 4 revs, 0.000586 s, 0.000564 s, -0.000022 s, × 0.9625, 141 µs/rev
mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 2 revs, 0.001845 s, 0.001549 s, -0.000296 s, × 0.8396, 774 µs/rev
mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 1 revs, 0.063822 s, 0.035918 s, -0.027904 s, × 0.5628, 35918 µs/rev
mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.088038 s, 0.073788 s, -0.014250 s, × 0.8381, 12298 µs/rev
mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.007389 s, 0.006151 s, -0.001238 s, × 0.8325, 3 µs/rev
mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.004868 s, 0.006165 s, +0.001297 s, × 1.2664, 0 µs/rev
mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 6657 revs, 0.222450 s, 0.065421 s, -0.157029 s, × 0.2941, 9 µs/rev
mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 40314 revs, 0.370675 s, 0.313749 s, -0.056926 s, × 0.8464, 7 µs/rev
mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 38690 revs, 0.358020 s, 0.297867 s, -0.060153 s, × 0.8320, 7 µs/rev
mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 54487 revs, 0.145235 s, 0.111300 s, -0.033935 s, × 0.7663, 2 µs/rev
mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 0.037606 s, 0.046202 s, +0.008596 s, × 1.2286, 1 µs/rev
mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 97052 revs, 7.382439 s, 1.999640 s, -5.382799 s, × 0.2709, 20 µs/rev
mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 52031 revs, 7.273506 s, 0.809134 s, -6.464372 s, × 0.1112, 15 µs/rev
mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 363753 revs, killed , 47.406785 s, , , 130 µs/rev
mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 444327 revs, 1.074593 s, 0.996219 s, -0.078374 s, × 0.9271, 2 µs/rev
mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 362229 revs, killed , 47.273399 s, , , 130 µs/rev
mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 359344 revs, killed , 47.419099 s, , , 131 µs/rev
mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 192665 revs, 27.746195 s, 3.512653 s, -24.233542 s, × 0.1266, 18 µs/rev
mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 237259 revs, killed , 44.459049 s, , , 187 µs/rev
mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 391148 revs, killed , 52.837926 s, , , 135 µs/rev

This revision compared to the filelog algorithm:

Repo Case Source-Rev Dest-Rev # of revisions filelog sidedata Difference Factor time per rev

mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 1 revs, 0.000906 s, 0.000049 s, -0.000857 s, × 0.0540, 48 µs/rev
mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 6 revs, 0.001844 s, 0.000114 s, -0.001730 s, × 0.0618, 18 µs/rev
mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 1032 revs, 0.018577 s, 0.004223 s, -0.014354 s, × 0.2273, 4 µs/rev
pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 9 revs, 0.005009 s, 0.000305 s, -0.004704 s, × 0.0608, 33 µs/rev
pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 1 revs, 0.209606 s, 0.000060 s, -0.209546 s, × 0.0002, 59 µs/rev
pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 7 revs, 0.017008 s, 0.000173 s, -0.016835 s, × 0.0101, 24 µs/rev
pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 1 revs, 0.019227 s, 0.000446 s, -0.018781 s, × 0.0231, 445 µs/rev
pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 6 revs, 0.765782 s, 0.010360 s, -0.755422 s, × 0.0135, 1726 µs/rev
pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 4785 revs, 1.186068 s, 0.048002 s, -1.138066 s, × 0.0404, 10 µs/rev
pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 6780 revs, 1.266745 s, 0.075705 s, -1.191040 s, × 0.0597, 11 µs/rev
pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 5441 revs, 1.666389 s, 0.056705 s, -1.609684 s, × 0.0340, 10 µs/rev
pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 43646 revs, 0.001070 s, 0.794685 s, +0.793615 s, × 742.69, 18 µs/rev
pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 26389 revs, 1.076269 s, 0.020209 s, -1.056060 s, × 0.0187, 0 µs/rev
pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 11316 revs, 1.355085 s, 0.122475 s, -1.232610 s, × 0.0903, 10 µs/rev
netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 2 revs, 0.028551 s, 0.000142 s, -0.028409 s, × 0.0049, 70 µs/rev
netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 2 revs, 0.157319 s, 0.000113 s, -0.157206 s, × 0.0007, 56 µs/rev
netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 3 revs, 0.025722 s, 0.000241 s, -0.025481 s, × 0.0093, 80 µs/rev
netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 9 revs, 0.053374 s, 0.000729 s, -0.052645 s, × 0.0136, 80 µs/rev
netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 1421 revs, 0.038146 s, 0.010198 s, -0.027948 s, × 0.2673, 7 µs/rev
netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 1533 revs, 0.229215 s, 0.015312 s, -0.213903 s, × 0.0668, 9 µs/rev
netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 5750 revs, 0.974484 s, 0.060517 s, -0.913967 s, × 0.0621, 10 µs/rev
netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 67005 revs, 3.924308 s, 0.611102 s, -3.313206 s, × 0.1557, 9 µs/rev
mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 2 revs, 0.035563 s, 0.000164 s, -0.035399 s, × 0.0046, 81 µs/rev
mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 8 revs, 0.145766 s, 0.000334 s, -0.145432 s, × 0.0022, 41 µs/rev
mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 9 revs, 0.026283 s, 0.000463 s, -0.025820 s, × 0.0176, 51 µs/rev
mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 7 revs, 0.087403 s, 0.000730 s, -0.086673 s, × 0.0083, 104 µs/rev
mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 3 revs, 0.209484 s, 0.003522 s, -0.205962 s, × 0.0168, 1173 µs/rev
mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 2.197867 s, 0.072518 s, -2.125349 s, × 0.0329, 12084 µs/rev
mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.090142 s, 0.005760 s, -0.084382 s, × 0.0638, 3 µs/rev
mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.742658 s, 0.005720 s, -0.736938 s, × 0.0077, 0 µs/rev
mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 7839 revs, 1.166159 s, 0.063310 s, -1.102849 s, × 0.0542, 8 µs/rev
mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 6.721719 s, 0.043608 s, -6.678111 s, × 0.0064, 0 µs/rev
mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 30263 revs, 3.356523 s, 0.204831 s, -3.151692 s, × 0.0610, 6 µs/rev
mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 153721 revs, 15.880822 s, 2.161906 s, -13.718916 s, × 0.1361, 14 µs/rev
mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 210546 revs, 20.781275 s, 3.291831 s, -17.489444 s, × 0.1584, 15 µs/rev
mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 2 revs, 0.084165 s, 0.001213 s, -0.082952 s, × 0.0144, 606 µs/rev
mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 2 revs, 0.503744 s, 0.001225 s, -0.502519 s, × 0.0024, 612 µs/rev
mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 4 revs, 0.021545 s, 0.000564 s, -0.020981 s, × 0.0261, 140 µs/rev
mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 2 revs, 0.240699 s, 0.001549 s, -0.239150 s, × 0.0064, 774 µs/rev
mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 1 revs, 1.100682 s, 0.035918 s, -1.064764 s, × 0.0326, 35882 µs/rev
mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 2.234809 s, 0.073788 s, -2.161021 s, × 0.0330, 12295 µs/rev
mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.091222 s, 0.006151 s, -0.085071 s, × 0.0674, 3 µs/rev
mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 8315 revs, 0.764722 s, 0.006165 s, -0.758557 s, × 0.0080, 0 µs/rev
mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 6657 revs, 1.185655 s, 0.065421 s, -1.120234 s, × 0.0551, 9 µs/rev
mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 40314 revs, 0.089736 s, 0.313749 s, +0.224013 s, × 3.4963, 7 µs/rev
mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 38690 revs, 0.084132 s, 0.297867 s, +0.213735 s, × 3.5404, 7 µs/rev
mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 54487 revs, 7.581932 s, 0.111300 s, -7.470632 s, × 0.0146, 2 µs/rev
mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 45299 revs, 6.671144 s, 0.046202 s, -6.624942 s, × 0.0069, 1 µs/rev
mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 97052 revs, 7.674771 s, 1.999640 s, -5.675131 s, × 0.2605, 20 µs/rev
mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 52031 revs, 9.870343 s, 0.809134 s, -9.061209 s, × 0.0819, 15 µs/rev
mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 363753 revs, 0.094781 s, 47.406785 s, +47.312004 s, × 500.17, 130 µs/rev
mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 444327 revs, 26.690029 s, 0.996219 s, -25.693810 s, × 0.0373, 2 µs/rev
mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 362229 revs, 0.094941 s, 47.273399 s, +47.178458 s, × 497.92, 130 µs/rev
mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 359344 revs, 0.233811 s, 47.419099 s, +47.185288 s, × 202.80, 131 µs/rev
mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 192665 revs, 19.321750 s, 3.512653 s, -15.809097 s, × 0.1817, 18 µs/rev
mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 237259 revs, 21.358350 s, 44.459049 s, +23.100699 s, × 2.0815, 187 µs/rev
mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 391148 revs, 25.328737 s, 52.837926 s, +27.509189 s, × 2.0860, 135 µs/rev

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Event Timeline

marmoute created this revision.Nov 12 2020, 10:17 AM
Alphare requested changes to this revision.Nov 13 2020, 8:17 AM
Alphare added a subscriber: Alphare.

A couple of missing assert!, but otherwise looks good.

rust/hg-core/src/copy_tracing.rs
67

Simply use b""

80

This error message would be better as not enough data for ChangedFiles or something like this

85

I'd add a couple of assert! just to make sure we get a better error message in case of out-of-bounds access.

129

This could be an assert! instead of if ... panic!, but I don't hate this either

160

Could you add consts for those magic numbers?

171

Same as above here. This can also fail because you have an overflow, not just because the data is too short.

220

Drop the return

288

I'd like to avoid adding more pub members with no documentation and a generic name. Maybe add a visibility modifier, and use #[doc(hidden)].

rust/hg-cpython/src/copy_tracing.rs
83

Same as the other commit, use a multi-line string

This revision now requires changes to proceed.Nov 13 2020, 8:17 AM
marmoute marked 6 inline comments as done.Nov 13 2020, 6:35 PM

feedback taking in account, diff upstate to come soon™

rust/hg-core/src/copy_tracing.rs
80

The slice above was already selection to be 4 bytes. So if I understand this correctly (and remember my discussion with Simon correctly) the fact I need to try_into here is simply a language limitation. I don't expect this to fail.

129

I find the if/panic clearer.

160

Sure, I overlooked them while cleaning this up.

171

I am not sure what you mean here.

marmoute marked an inline comment as done.Nov 13 2020, 7:05 PM
marmoute updated this revision to Diff 23501.
Alphare accepted this revision.Nov 16 2020, 6:39 AM
Alphare added inline comments.
rust/hg-core/src/copy_tracing.rs
80

Yep, with the assert! beforehand, there is no way this can fail.

171

Yeah, I was wrong. However checking that you have enough bytes in the index to have a better error message in case of an out-of-bounds panic is still a good idea. I think that you've covered that in the caller function, so it's fine.

marmoute updated this revision to Diff 23812.Nov 29 2020, 6:13 AM
SimonSapin accepted this revision.Dec 1 2020, 4:10 AM
SimonSapin added a subscriber: SimonSapin.

Looks good to me!

rust/hg-core/src/copy_tracing.rs
206

Is this a binary search? Given that off by ones are notoriously easy it’d be nice to use https://doc.rust-lang.org/std/primitive.slice.html#method.binary_search_by_key instead of making our own. But that only searches in slices while self.entry() here works with integer indices, so never mind.

pulkit accepted this revision.Dec 7 2020, 8:37 AM
pulkit added a subscriber: pulkit.
pulkit added inline comments.
rust/hg-core/src/copy_tracing.rs
66–116

This type needs some documentation love. Kindly followup.

This revision is now accepted and ready to land.Dec 7 2020, 8:37 AM

Revision Contents
Changeset List

Diff 24084

mercurial/copies.py

policy, policy,
pycompat, pycompat,
util, util,
) )
from .utils import stringutil from .utils import stringutil
from .revlogutils import flagutil from .revlogutils import (
flagutil,
sidedata as sidedatamod,
)
rustmod = policy.importrust("copy_tracing") rustmod = policy.importrust("copy_tracing")
def _filter(src, dst, t): def _filter(src, dst, t):
"""filters out invalid copies after chaining""" """filters out invalid copies after chaining"""
# When _chain()'ing copies in 'a' (from 'src' via some other commit 'mid') # When _chain()'ing copies in 'a' (from 'src' via some other commit 'mid')
if debug: if debug:
dbg( dbg(
b'debug.copies: time: %f seconds\n' b'debug.copies: time: %f seconds\n'
% (util.timer() - start) % (util.timer() - start)
) )
return cm return cm
def _revinfo_getter(repo): def _revinfo_getter(repo, match):
"""returns a function that returns the following data given a <rev>" """returns a function that returns the following data given a <rev>"
* p1: revision number of first parent * p1: revision number of first parent
* p2: revision number of first parent * p2: revision number of first parent
* changes: a ChangingFiles object * changes: a ChangingFiles object
""" """
cl = repo.changelog cl = repo.changelog
parents = cl.parentrevs parents = cl.parentrevs
# #
# In addition, it would be possible to reach pathological case, were # In addition, it would be possible to reach pathological case, were
# many first parent are met before any second parent is reached. In # many first parent are met before any second parent is reached. In
# that case the cache could grow. If this even become an issue one can # that case the cache could grow. If this even become an issue one can
# safely introduce a maximum cache size. This would trade extra CPU/IO # safely introduce a maximum cache size. This would trade extra CPU/IO
# time to save memory. # time to save memory.
merge_caches = {} merge_caches = {}
alwaysmatch = match.always()
if rustmod is not None and alwaysmatch:
def revinfo(rev):
p1, p2 = parents(rev)
value = None
e = merge_caches.pop(rev, None)
if e is not None:
return e
if flags(rev) & HASCOPIESINFO:
raw = changelogrevision(rev)._sidedata.get(sidedatamod.SD_FILES)
else:
raw = None
value = (p1, p2, raw)
if p1 != node.nullrev and p2 != node.nullrev:
# XXX some case we over cache, IGNORE
merge_caches[rev] = value
return value
else:
def revinfo(rev): def revinfo(rev):
p1, p2 = parents(rev) p1, p2 = parents(rev)
value = None value = None
e = merge_caches.pop(rev, None) e = merge_caches.pop(rev, None)
if e is not None: if e is not None:
return e return e
changes = None changes = None
if flags(rev) & HASCOPIESINFO: if flags(rev) & HASCOPIESINFO:
changes = changelogrevision(rev).changes changes = changelogrevision(rev).changes
value = (p1, p2, changes) value = (p1, p2, changes)
if p1 != node.nullrev and p2 != node.nullrev: if p1 != node.nullrev and p2 != node.nullrev:
# XXX some case we over cache, IGNORE # XXX some case we over cache, IGNORE
merge_caches[rev] = value merge_caches[rev] = value
return value return value
return revinfo return revinfo
def cached_is_ancestor(is_ancestor): def cached_is_ancestor(is_ancestor):
"""return a cached version of is_ancestor""" """return a cached version of is_ancestor"""
cache = {} cache = {}
iterrevs = set(from_head) iterrevs = set(from_head)
iterrevs &= mrset iterrevs &= mrset
iterrevs.update(roots) iterrevs.update(roots)
iterrevs.remove(b.rev()) iterrevs.remove(b.rev())
revs = sorted(iterrevs) revs = sorted(iterrevs)
if repo.filecopiesmode == b'changeset-sidedata': if repo.filecopiesmode == b'changeset-sidedata':
revinfo = _revinfo_getter(repo) revinfo = _revinfo_getter(repo, match)
return _combine_changeset_copies( return _combine_changeset_copies(
revs, children, b.rev(), revinfo, match, isancestor revs, children, b.rev(), revinfo, match, isancestor
) )
else: else:
revinfo = _revinfo_getter_extra(repo) revinfo = _revinfo_getter_extra(repo)
return _combine_changeset_copies_extra( return _combine_changeset_copies_extra(
revs, children, b.rev(), revinfo, match, isancestor revs, children, b.rev(), revinfo, match, isancestor
) )

rust/hg-core/src/copy_tracing.rs

use crate::utils::hg_path::HgPath; use crate::utils::hg_path::HgPath;
use crate::utils::hg_path::HgPathBuf; use crate::utils::hg_path::HgPathBuf;
use crate::Revision; use crate::Revision;
use im_rc::ordmap::DiffItem; use im_rc::ordmap::DiffItem;
use im_rc::ordmap::OrdMap; use im_rc::ordmap::OrdMap;
use std::cmp::Ordering;
use std::collections::HashMap; use std::collections::HashMap;
use std::collections::HashSet; use std::convert::TryInto;
pub type PathCopies = HashMap<HgPathBuf, HgPathBuf>; pub type PathCopies = HashMap<HgPathBuf, HgPathBuf>;
#[derive(Clone, Debug, PartialEq)] #[derive(Clone, Debug, PartialEq)]
struct TimeStampedPathCopy { struct TimeStampedPathCopy {
/// revision at which the copy information was added /// revision at which the copy information was added
rev: Revision, rev: Revision,
/// the copy source, (Set to None in case of deletion of the associated /// the copy source, (Set to None in case of deletion of the associated
/// key) /// key)
path: Option<HgPathBuf>, path: Option<HgPathBuf>,
} }
/// maps CopyDestination to Copy Source (+ a "timestamp" for the operation) /// maps CopyDestination to Copy Source (+ a "timestamp" for the operation)
type TimeStampedPathCopies = OrdMap<HgPathBuf, TimeStampedPathCopy>; type TimeStampedPathCopies = OrdMap<HgPathBuf, TimeStampedPathCopy>;
/// hold parent 1, parent 2 and relevant files actions. /// hold parent 1, parent 2 and relevant files actions.
pub type RevInfo = (Revision, Revision, ChangedFiles); pub type RevInfo<'a> = (Revision, Revision, ChangedFiles<'a>);
/// represent the files affected by a changesets /// represent the files affected by a changesets
/// ///
/// This hold a subset of mercurial.metadata.ChangingFiles as we do not need /// This hold a subset of mercurial.metadata.ChangingFiles as we do not need
/// all the data categories tracked by it. /// all the data categories tracked by it.
pub struct ChangedFiles { /// This hold a subset of mercurial.metadata.ChangingFiles as we do not need
removed: HashSet<HgPathBuf>, /// all the data categories tracked by it.
merged: HashSet<HgPathBuf>, pub struct ChangedFiles<'a> {
salvaged: HashSet<HgPathBuf>, nb_items: u32,
copied_from_p1: PathCopies, index: &'a [u8],
copied_from_p2: PathCopies, data: &'a [u8],
} }
/// Represent active changes that affect the copy tracing. /// Represent active changes that affect the copy tracing.
enum Action<'a> { enum Action<'a> {
/// The parent ? children edge is removing a file /// The parent ? children edge is removing a file
/// ///
/// (actually, this could be the edge from the other parent, but it does /// (actually, this could be the edge from the other parent, but it does
/// not matters) /// not matters)
/// Merged: file had history on both side that needed to be merged /// Merged: file had history on both side that needed to be merged
Merged, Merged,
/// Salvaged: file was candidate for deletion, but survived the merge /// Salvaged: file was candidate for deletion, but survived the merge
Salvaged, Salvaged,
/// Normal: Not one of the two cases above /// Normal: Not one of the two cases above
Normal, Normal,
} }
impl ChangedFiles { type FileChange<'a> = (u8, &'a HgPath, &'a HgPath);
pub fn new(
AlphareUnsubmitted
Done

Simply use b""

Alphare: Simply use `b""`
removed: HashSet<HgPathBuf>, const EMPTY: &[u8] = b"";
merged: HashSet<HgPathBuf>, const COPY_MASK: u8 = 3;
salvaged: HashSet<HgPathBuf>, const P1_COPY: u8 = 2;
copied_from_p1: PathCopies, const P2_COPY: u8 = 3;
copied_from_p2: PathCopies, const ACTION_MASK: u8 = 28;
) -> Self { const REMOVED: u8 = 12;
ChangedFiles { const MERGED: u8 = 8;
removed, const SALVAGED: u8 = 16;
merged,
salvaged, impl<'a> ChangedFiles<'a> {
copied_from_p1, const INDEX_START: usize = 4;
copied_from_p2, const ENTRY_SIZE: u32 = 9;
} const FILENAME_START: u32 = 1;
AlphareUnsubmitted
Not Done

This error message would be better as not enough data for ChangedFiles or something like this

Alphare: This error message would be better as `not enough data for ChangedFiles` or something like this
marmouteAuthorUnsubmitted
Done

The slice above was already selection to be 4 bytes. So if I understand this correctly (and remember my discussion with Simon correctly) the fact I need to try_into here is simply a language limitation. I don't expect this to fail.

marmoute: The slice above was already selection to be 4 bytes. So if I understand this correctly (and…
AlphareUnsubmitted
Not Done

Yep, with the assert! beforehand, there is no way this can fail.

Alphare: Yep, with the `assert!` beforehand, there is no way this can fail.
const COPY_SOURCE_START: u32 = 5;
pub fn new(data: &'a [u8]) -> Self {
assert!(
data.len() >= 4,
AlphareUnsubmitted
Done

I'd add a couple of assert! just to make sure we get a better error message in case of out-of-bounds access.

Alphare: I'd add a couple of `assert!` just to make sure we get a better error message in case of out-of…
"data size ({}) is too small to contain the header (4)",
data.len()
);
let nb_items_raw: [u8; 4] = (&data[0..=3])
.try_into()
.expect("failed to turn 4 bytes into 4 bytes");
let nb_items = u32::from_be_bytes(nb_items_raw);
let index_size = (nb_items * Self::ENTRY_SIZE) as usize;
let index_end = Self::INDEX_START + index_size;
assert!(
data.len() >= index_end,
"data size ({}) is too small to fit the index_data ({})",
data.len(),
index_end
);
let ret = ChangedFiles {
nb_items,
index: &data[Self::INDEX_START..index_end],
data: &data[index_end..],
};
let max_data = ret.filename_end(nb_items - 1) as usize;
assert!(
ret.data.len() >= max_data,
"data size ({}) is too small to fit all data ({})",
data.len(),
index_end + max_data
);
ret
pulkitUnsubmitted
Not Done

This type needs some documentation love. Kindly followup.

pulkit: This type needs some documentation love. Kindly followup.
} }
pub fn new_empty() -> Self { pub fn new_empty() -> Self {
ChangedFiles { ChangedFiles {
removed: HashSet::new(), nb_items: 0,
merged: HashSet::new(), index: EMPTY,
salvaged: HashSet::new(), data: EMPTY,
copied_from_p1: PathCopies::new(), }
copied_from_p2: PathCopies::new(), }
/// internal function to return an individual entry at a given index
fn entry(&'a self, idx: u32) -> FileChange<'a> {
if idx >= self.nb_items {
AlphareUnsubmitted
Not Done

This could be an assert! instead of if ... panic!, but I don't hate this either

Alphare: This could be an `assert!` instead of `if ... panic!`, but I don't hate this either
marmouteAuthorUnsubmitted
Done

I find the if/panic clearer.

marmoute: I find the `if/panic` clearer.
panic!(
"index for entry is higher that the number of file {} >= {}",
idx, self.nb_items
)
}
let flags = self.flags(idx);
let filename = self.filename(idx);
let copy_idx = self.copy_idx(idx);
let copy_source = self.filename(copy_idx);
(flags, filename, copy_source)
}
/// internal function to return the filename of the entry at a given index
fn filename(&self, idx: u32) -> &HgPath {
let filename_start;
if idx == 0 {
filename_start = 0;
} else {
filename_start = self.filename_end(idx - 1)
} }
let filename_end = self.filename_end(idx);
let filename_start = filename_start as usize;
let filename_end = filename_end as usize;
HgPath::new(&self.data[filename_start..filename_end])
}
/// internal function to return the flag field of the entry at a given
/// index
fn flags(&self, idx: u32) -> u8 {
let idx = idx as usize;
self.index[idx * (Self::ENTRY_SIZE as usize)]
AlphareUnsubmitted
Done

Could you add consts for those magic numbers?

Alphare: Could you add `const`s for those magic numbers?
marmouteAuthorUnsubmitted
Done

Sure, I overlooked them while cleaning this up.

marmoute: Sure, I overlooked them while cleaning this up.
}
/// internal function to return the end of a filename part at a given index
fn filename_end(&self, idx: u32) -> u32 {
let start = (idx * Self::ENTRY_SIZE) + Self::FILENAME_START;
let end = (idx * Self::ENTRY_SIZE) + Self::COPY_SOURCE_START;
let start = start as usize;
let end = end as usize;
let raw = (&self.index[start..end])
.try_into()
.expect("failed to turn 4 bytes into 4 bytes");
AlphareUnsubmitted
Not Done

Same as above here. This can also fail because you have an overflow, not just because the data is too short.

Alphare: Same as above here. This can also fail because you have an overflow, not just because the data…
marmouteAuthorUnsubmitted
Done

I am not sure what you mean here.

marmoute: I am not sure what you mean here.
AlphareUnsubmitted
Not Done

Yeah, I was wrong. However checking that you have enough bytes in the index to have a better error message in case of an out-of-bounds panic is still a good idea. I think that you've covered that in the caller function, so it's fine.

Alphare: Yeah, I was wrong. However checking that you have enough bytes in the index to have a better…
u32::from_be_bytes(raw)
}
/// internal function to return index of the copy source of the entry at a
/// given index
fn copy_idx(&self, idx: u32) -> u32 {
let start = (idx * Self::ENTRY_SIZE) + Self::COPY_SOURCE_START;
let end = (idx + 1) * Self::ENTRY_SIZE;
let start = start as usize;
let end = end as usize;
let raw = (&self.index[start..end])
.try_into()
.expect("failed to turn 4 bytes into 4 bytes");
u32::from_be_bytes(raw)
} }
/// Return an iterator over all the `Action` in this instance. /// Return an iterator over all the `Action` in this instance.
fn iter_actions(&self, parent: usize) -> impl Iterator<Item = Action> { fn iter_actions(&self, parent: usize) -> ActionsIterator {
let copies_iter = match parent { ActionsIterator {
1 => self.copied_from_p1.iter(), changes: &self,
2 => self.copied_from_p2.iter(), parent: parent,
_ => unreachable!(), current: 0,
}; }
let remove_iter = self.removed.iter();
let copies_iter = copies_iter.map(|(x, y)| Action::Copied(x, y));
let remove_iter = remove_iter.map(|x| Action::Removed(x));
copies_iter.chain(remove_iter)
} }
/// return the MergeCase value associated with a filename /// return the MergeCase value associated with a filename
fn get_merge_case(&self, path: &HgPath) -> MergeCase { fn get_merge_case(&self, path: &HgPath) -> MergeCase {
if self.salvaged.contains(path) { if self.nb_items == 0 {
return MergeCase::Salvaged;
} else if self.merged.contains(path) {
return MergeCase::Merged;
} else {
return MergeCase::Normal; return MergeCase::Normal;
} }
let mut low_part = 0;
let mut high_part = self.nb_items;
while low_part < high_part {
let cursor = (low_part + high_part - 1) / 2;
SimonSapinUnsubmitted
Done

Is this a binary search? Given that off by ones are notoriously easy it’d be nice to use https://doc.rust-lang.org/std/primitive.slice.html#method.binary_search_by_key instead of making our own. But that only searches in slices while self.entry() here works with integer indices, so never mind.

SimonSapin: Is this a binary search? Given that off by ones are notoriously easy it’d be nice to use https…
let (flags, filename, _source) = self.entry(cursor);
match path.cmp(filename) {
Ordering::Less => low_part = cursor + 1,
Ordering::Greater => high_part = cursor,
Ordering::Equal => {
return match flags & ACTION_MASK {
MERGED => MergeCase::Merged,
SALVAGED => MergeCase::Salvaged,
_ => MergeCase::Normal,
};
}
}
}
MergeCase::Normal
AlphareUnsubmitted
Done

Drop the return

Alphare: Drop the `return`
} }
} }
/// A struct responsible for answering "is X ancestors of Y" quickly /// A struct responsible for answering "is X ancestors of Y" quickly
/// ///
/// The structure will delegate ancestors call to a callback, and cache the /// The structure will delegate ancestors call to a callback, and cache the
/// result. /// result.
#[derive(Debug)] #[derive(Debug)]
let b = (self.inner)(anc, desc); let b = (self.inner)(anc, desc);
self.pairs.insert((anc, desc), b); self.pairs.insert((anc, desc), b);
b b
} }
} }
} }
} }
struct ActionsIterator<'a> {
changes: &'a ChangedFiles<'a>,
parent: usize,
current: u32,
}
impl<'a> Iterator for ActionsIterator<'a> {
type Item = Action<'a>;
fn next(&mut self) -> Option<Action<'a>> {
while self.current < self.changes.nb_items {
let (flags, file, source) = self.changes.entry(self.current);
self.current += 1;
if (flags & ACTION_MASK) == REMOVED {
return Some(Action::Removed(file));
}
let copy = flags & COPY_MASK;
if self.parent == 1 && copy == P1_COPY {
return Some(Action::Copied(file, source));
}
if self.parent == 2 && copy == P2_COPY {
return Some(Action::Copied(file, source));
}
}
return None;
}
}
/// A small struct whose purpose is to ensure lifetime of bytes referenced in
AlphareUnsubmitted
Done

I'd like to avoid adding more pub members with no documentation and a generic name. Maybe add a visibility modifier, and use #[doc(hidden)].

Alphare: I'd like to avoid adding more `pub` members with no documentation and a generic name. Maybe add…
/// ChangedFiles
///
/// It is passed to the RevInfoMaker callback who can assign any necessary
/// content to the `data` attribute. The copy tracing code is responsible for
/// keeping the DataHolder alive at least as long as the ChangedFiles object.
pub struct DataHolder<D> {
/// RevInfoMaker callback should assign data referenced by the
/// ChangedFiles struct it return to this attribute. The DataHolder
/// lifetime will be at least as long as the ChangedFiles one.
pub data: Option<D>,
}
pub type RevInfoMaker<'a, D> =
Box<dyn for<'r> Fn(Revision, &'r mut DataHolder<D>) -> RevInfo<'r> + 'a>;
/// Same as mercurial.copies._combine_changeset_copies, but in Rust. /// Same as mercurial.copies._combine_changeset_copies, but in Rust.
/// ///
/// Arguments are: /// Arguments are:
/// ///
/// revs: all revisions to be considered /// revs: all revisions to be considered
/// children: a {parent ? [childrens]} mapping /// children: a {parent ? [childrens]} mapping
/// target_rev: the final revision we are combining copies to /// target_rev: the final revision we are combining copies to
/// rev_info(rev): callback to get revision information: /// rev_info(rev): callback to get revision information:
/// * first parent /// * first parent
/// * second parent /// * second parent
/// * ChangedFiles /// * ChangedFiles
/// isancestors(low_rev, high_rev): callback to check if a revision is an /// isancestors(low_rev, high_rev): callback to check if a revision is an
/// ancestor of another /// ancestor of another
pub fn combine_changeset_copies<A: Fn(Revision, Revision) -> bool>( pub fn combine_changeset_copies<A: Fn(Revision, Revision) -> bool, D>(
revs: Vec<Revision>, revs: Vec<Revision>,
children: HashMap<Revision, Vec<Revision>>, children: HashMap<Revision, Vec<Revision>>,
target_rev: Revision, target_rev: Revision,
rev_info: &impl Fn(Revision) -> RevInfo, rev_info: RevInfoMaker<D>,
is_ancestor: &A, is_ancestor: &A,
) -> PathCopies { ) -> PathCopies {
let mut all_copies = HashMap::new(); let mut all_copies = HashMap::new();
let mut oracle = AncestorOracle::new(is_ancestor); let mut oracle = AncestorOracle::new(is_ancestor);
for rev in revs { for rev in revs {
// Retrieve data computed in a previous iteration // Retrieve data computed in a previous iteration
let copies = all_copies.remove(&rev); let copies = all_copies.remove(&rev);
let copies = match copies { let copies = match copies {
Some(c) => c, Some(c) => c,
None => TimeStampedPathCopies::default(), // root of the walked set None => TimeStampedPathCopies::default(), // root of the walked set
}; };
let current_children = match children.get(&rev) { let current_children = match children.get(&rev) {
Some(c) => c, Some(c) => c,
None => panic!("inconsistent `revs` and `children`"), None => panic!("inconsistent `revs` and `children`"),
}; };
for child in current_children { for child in current_children {
// We will chain the copies information accumulated for `rev` with // We will chain the copies information accumulated for `rev` with
// the individual copies information for each of its children. // the individual copies information for each of its children.
// Creating a new PathCopies for each `rev` ? `children` vertex. // Creating a new PathCopies for each `rev` → `children` vertex.
let (p1, p2, changes) = rev_info(*child); let mut d: DataHolder<D> = DataHolder { data: None };
let (p1, p2, changes) = rev_info(*child, &mut d);
let parent = if rev == p1 { let parent = if rev == p1 {
1 1
} else { } else {
assert_eq!(rev, p2); assert_eq!(rev, p2);
2 2
}; };
let mut new_copies = copies.clone(); let mut new_copies = copies.clone();

rust/hg-cpython/src/copy_tracing.rs

use cpython::ObjectProtocol; use cpython::ObjectProtocol;
use cpython::PyBool; use cpython::PyBool;
use cpython::PyBytes; use cpython::PyBytes;
use cpython::PyDict; use cpython::PyDict;
use cpython::PyList; use cpython::PyList;
use cpython::PyModule; use cpython::PyModule;
use cpython::PyObject; use cpython::PyObject;
use cpython::PyResult; use cpython::PyResult;
use cpython::PyTuple; use cpython::PyTuple;
use cpython::Python; use cpython::Python;
use hg::copy_tracing::combine_changeset_copies; use hg::copy_tracing::combine_changeset_copies;
use hg::copy_tracing::ChangedFiles; use hg::copy_tracing::ChangedFiles;
use hg::copy_tracing::DataHolder;
use hg::copy_tracing::RevInfo; use hg::copy_tracing::RevInfo;
use hg::utils::hg_path::HgPathBuf; use hg::copy_tracing::RevInfoMaker;
use hg::Revision; use hg::Revision;
/// Combines copies information contained into revision `revs` to build a copy /// Combines copies information contained into revision `revs` to build a copy
/// map. /// map.
/// ///
/// See mercurial/copies.py for details /// See mercurial/copies.py for details
pub fn combine_changeset_copies_wrapper( pub fn combine_changeset_copies_wrapper(
py: Python, py: Python,
}; };
// Wrap the `rev_info_maker` python callback as a Rust closure // Wrap the `rev_info_maker` python callback as a Rust closure
// //
// No errors are expected from the Python side, and they will should only // No errors are expected from the Python side, and they will should only
// happens in case of programing error or severe data corruption. Such // happens in case of programing error or severe data corruption. Such
// errors will raise panic and the rust-cpython harness will turn them into // errors will raise panic and the rust-cpython harness will turn them into
// Python exception. // Python exception.
let rev_info_maker = |rev: Revision| -> RevInfo { let rev_info_maker: RevInfoMaker<PyBytes> =
Box::new(|rev: Revision, d: &mut DataHolder<PyBytes>| -> RevInfo {
let res: PyTuple = rev_info let res: PyTuple = rev_info
.call(py, (rev,), None) .call(py, (rev,), None)
.expect("rust-copy-tracing: python call to `rev_info` failed") .expect("rust-copy-tracing: python call to `rev_info` failed")
.cast_into(py) .cast_into(py)
.expect( .expect(
"rust-copy_tracing: python call to `rev_info` returned \ "rust-copy_tracing: python call to `rev_info` returned \
unexpected non-Tuple value", unexpected non-Tuple value",
); );
let p1 = res.get_item(py, 0).extract(py).expect( let p1 = res.get_item(py, 0).extract(py).expect(
"rust-copy-tracing: \ "rust-copy-tracing: rev_info return is invalid, first item \
rev_info return is invalid, first item is a not a revision", is a not a revision",
); );
let p2 = res.get_item(py, 1).extract(py).expect( let p2 = res.get_item(py, 1).extract(py).expect(
"rust-copy-tracing: \ "rust-copy-tracing: rev_info return is invalid, first item \
rev_info return is invalid, second item is a not a revision", is a not a revision",
); );
let changes = res.get_item(py, 2); let files = match res.get_item(py, 2).extract::<PyBytes>(py) {
Ok(raw) => {
let files; // Give responsability for the raw bytes lifetime to
if !changes // hg-core
AlphareUnsubmitted
Done

Same as the other commit, use a multi-line string

Alphare: Same as the other commit, use a multi-line string
.hasattr(py, "copied_from_p1") d.data = Some(raw);
.expect("rust-copy-tracing: python call to `hasattr` failed") let addrs = d.data.as_ref().expect(
{ "rust-copy-tracing: failed to get a reference to the \
files = ChangedFiles::new_empty(); raw bytes for copy data").data(py);
} else { ChangedFiles::new(addrs)
let p1_copies: PyDict = changes
.getattr(py, "copied_from_p1")
.expect(
"rust-copy-tracing: retrieval of python attribute \
`copied_from_p1` failed",
)
.cast_into(py)
.expect(
"rust-copy-tracing: failed to convert `copied_from_p1` \
to PyDict",
);
let p1_copies: PyResult<_> = p1_copies
.items(py)
.iter()
.map(|(key, value)| {
let key = key.extract::<PyBytes>(py).expect(
"rust-copy-tracing: conversion of copy destination to\
PyBytes failed",
);
let key = key.data(py);
let value = value.extract::<PyBytes>(py).expect(
"rust-copy-tracing: conversion of copy source to \
PyBytes failed",
);
let value = value.data(py);
Ok((
HgPathBuf::from_bytes(key),
HgPathBuf::from_bytes(value),
))
})
.collect();
let p2_copies: PyDict = changes
.getattr(py, "copied_from_p2")
.expect(
"rust-copy-tracing: retrieval of python attribute \
`copied_from_p2` failed",
)
.cast_into(py)
.expect(
"rust-copy-tracing: failed to convert `copied_from_p2` \
to PyDict",
);
let p2_copies: PyResult<_> = p2_copies
.items(py)
.iter()
.map(|(key, value)| {
let key = key.extract::<PyBytes>(py).expect(
"rust-copy-tracing: conversion of copy destination to \
PyBytes failed");
let key = key.data(py);
let value = value.extract::<PyBytes>(py).expect(
"rust-copy-tracing: conversion of copy source to \
PyBytes failed",
);
let value = value.data(py);
Ok((
HgPathBuf::from_bytes(key),
HgPathBuf::from_bytes(value),
))
})
.collect();
let removed: PyObject = changes.getattr(py, "removed").expect(
"rust-copy-tracing: retrieval of python attribute \
`removed` failed",
);
let removed: PyResult<_> = removed
.iter(py)
.expect(
"rust-copy-tracing: getting a python iterator over the \
`removed` set failed",
)
.map(|filename| {
let filename = filename
.expect(
"rust-copy-tracing: python iteration over the \
`removed` set failed",
)
.extract::<PyBytes>(py)
.expect(
"rust-copy-tracing: \
conversion of `removed` item to PyBytes failed",
);
let filename = filename.data(py);
Ok(HgPathBuf::from_bytes(filename))
})
.collect();
let merged: PyObject = changes.getattr(py, "merged").expect(
"rust-copy-tracing: retrieval of python attribute \
`merged` failed",
);
let merged: PyResult<_> = merged
.iter(py)
.expect(
"rust-copy-tracing: getting a python iterator over the \
`merged` set failed",
)
.map(|filename| {
let filename = filename
.expect(
"rust-copy-tracing: python iteration over the \
`merged` set failed",
)
.extract::<PyBytes>(py)
.expect(
"rust-copy-tracing: \
conversion of `merged` item to PyBytes failed",
);
let filename = filename.data(py);
Ok(HgPathBuf::from_bytes(filename))
})
.collect();
let salvaged: PyObject = changes.getattr(py, "salvaged").expect(
"rust-copy-tracing: retrieval of python attribute \
`salvaged` failed",
);
let salvaged: PyResult<_> = salvaged
.iter(py)
.expect(
"rust-copy-tracing: getting a python iterator over the \
`salvaged` set failed",
)
.map(|filename| {
let filename = filename
.expect(
"rust-copy-tracing: python iteration over the \
`salvaged` set failed",
)
.extract::<PyBytes>(py)
.expect(
"rust-copy-tracing: \
conversion of `salvaged` item to PyBytes failed",
);
let filename = filename.data(py);
Ok(HgPathBuf::from_bytes(filename))
})
.collect();
files = ChangedFiles::new(
removed.unwrap(),
merged.unwrap(),
salvaged.unwrap(),
p1_copies.unwrap(),
p2_copies.unwrap(),
);
} }
// value was presumably None, meaning they was no copy data.
Err(_) => ChangedFiles::new_empty(),
};
(p1, p2, files) (p1, p2, files)
}; });
let children: PyResult<_> = children let children: PyResult<_> = children
.items(py) .items(py)
.iter() .iter()
.map(|(k, v)| { .map(|(k, v)| {
let v: &PyList = v.cast_as(py)?; let v: &PyList = v.cast_as(py)?;
let v: PyResult<_> = let v: PyResult<_> =
v.iter(py).map(|child| Ok(child.extract(py)?)).collect(); v.iter(py).map(|child| Ok(child.extract(py)?)).collect();
Ok((k.extract(py)?, v?)) Ok((k.extract(py)?, v?))
}) })
.collect(); .collect();
let res = combine_changeset_copies( let res = combine_changeset_copies(
revs?, revs?,
children?, children?,
target_rev, target_rev,
&rev_info_maker, rev_info_maker,
&is_ancestor_wrap, &is_ancestor_wrap,
); );
let out = PyDict::new(py); let out = PyDict::new(py);
for (dest, source) in res.into_iter() { for (dest, source) in res.into_iter() {
out.set_item( out.set_item(
py, py,
PyBytes::new(py, &dest.into_vec()), PyBytes::new(py, &dest.into_vec()),
PyBytes::new(py, &source.into_vec()), PyBytes::new(py, &source.into_vec()),