Repository navigation
Look up the cases of a suite in linear time - #482
Merged
Merged
Conversation
samoht
force-pushed
the
fix-case-lookup
branch
5 times, most recently
from
November 12, 2026 12:33
d3aad20 to
244bbda
Compare
Since 1.10.0, compiling a suite looked up each case twice in the list of every case created with the same name, so a suite of many cases with the same name, as in a suite built from test vectors, was quadratic: 40,000 cases named "v" took 5.6s to list, against 0.35s with distinct names. The cases of a suite come in the order they were created in, or in the reverse order for a list literal, so a lookup now tries the neighbours of the previous one before scanning, and each case is looked up once. The same 40,000 cases now take 0.29s, in either order.
samoht
force-pushed
the
fix-case-lookup
branch
from
November 12, 2026 13:43
244bbda to
bce78b2
Compare
jmid
pushed a commit
to ocaml/opam-repository
that referenced
this pull request
Nov 17, 2026
CHANGES: - Fix parallel test runs failing with `Failed 6 times to create symlink`, and a dangling `latest` link (mirage/alcotest#479, @samoht, fixes mirage/alcotest#471 and mirage/alcotest#472 in 1.10.0) - Fix tests failing with `EACCES` on a read-only source tree or with `DUNE_BUILD_DIR` (mirage/alcotest#480, @samoht, fixes mirage/alcotest#471 in 1.10.0) - Honour `ALCOTEST_SOURCE_CODE_POSITION` when set by the program before `run` (mirage/alcotest#481, @samoht, fixes mirage/alcotest#451 in 1.10.0) - Fix slow runs of suites with many cases of the same name (mirage/alcotest#482, @samoht, fixes mirage/alcotest#451 in 1.10.0) - Under Eio, point the failure locations added in 1.10.0 (mirage/alcotest#451) at the test, not at the scheduler (mirage/alcotest#484, fixes mirage/alcotest#478, @samoht, report by @talex5) - Point a failed `check` in a helper at the `check`, as in 1.9.1, not at the test that calls the helper (mirage/alcotest#488, @samoht, fixes mirage/alcotest#451 in 1.10.0)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Since #451, Alcotest records the location of each test case in a table grouped by case name. Each case keeps its own entry: lookups compare the case value itself, so two cases with the same name are told apart. But finding a case scans every entry with its name, so a suite where many cases share a name, such as one generated from test vectors, was quadratic: listing 40,000 cases all named
"v"took 5.6s, against 0.35s with distinct names.Cases are almost always looked up in the order they were created, or in reverse order for a list literal. The lookup now tries the entries next to the previous match before scanning, so the same 40,000 cases take 0.29s.