Add virtual-twin type-identity regression tests (#229) by NullVoxPopuli-ai-agent · Pull Request #240 · ember-tooling/ember-eslint-parser

NullVoxPopuli-ai-agent · 2026-07-06T16:54:52Z

Tests extracted from #235 so the fix and its regression tests can be reviewed separately. Merge #235 first — these tests intentionally fail against the current parser; the failures are the #229 bug.

What's covered

Three assertions, each run under both project and projectService (the service mode is feature-detected: its option spelling varies across typescript-eslint majors, and v7's experimental service can't hold .gts files in-project, so it's registered only when a probe parse yields a program that actually contains virtual twins — 3 tests on tsee 7, 6 on tsee 8):

Every virtual twin is grafted — the .ts resolution virtual and (where present) the .mts readDirectory twin share statements with the .gts root.
One type identity across import paths — a class with a private member resolves to the identical type object via the .gts root and via its virtual twins, and is assignable to itself in both directions.
No re-graft of an already-synced program — a canary planted on a twin after binding survives a later parse; without the once-per-program guard the graft rewrites binder state under a live checker on every parse.

Expected failures without #235

Which tests fail depends on the toolchain, because the .mts twin only appears in the program on newer TS versions:

current default toolchain (TS 5.7): the two re-graft canary tests fail (2 failed | 4 passed);
TS 6 matrix cells: the twin-graft and type-identity tests fail as well, since the old code grafts only the .mts twin and leaves the .ts virtual — the copy every extensionless import resolves to — ungrafted.

After #235, all 6 pass on every matrix cell (verified on #235's CI before the split).

🤖 Generated with Claude Code

Covers the graft behavior fixed in ember-tooling#235, under both project and projectService (feature-detected; v7's experimental service can't hold .gts files in-project, so the mode is registered only when a probe parse yields a program that contains virtual twins): - every virtual twin present shares statements with the .gts root, - a private-member class has one type identity across import paths, - an already-synced program is not re-grafted on later parses. These tests FAIL against the parser without ember-tooling#235 — that failure is the ember-tooling#229 bug (distinct class symbols per import path, binder state rewritten under a live checker). Merge ember-tooling#235 first. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

github-actions · 2026-07-06T17:02:50Z

🏎️ Benchmark Comparison

Parse

	Benchmark	Control (p50)	Experiment (p50)	Δ
🟢	gts small	932.55 ms	895.54 ms	-4.0%
⚪	gts medium	1611.63 ms	1599.94 ms	-0.7%
⚪	gts large	813.39 ms	820.17 ms	+0.8%
⚪	gjs small	798.10 ms	794.37 ms	-0.5%
🟢	gjs medium	1387.77 ms	1356.98 ms	-2.2%
🟢	gjs large	638.02 ms	621.16 ms	-2.6%
🟠	hbs small	129.71 ms	134.83 ms	+3.9%
🟠	hbs medium	725.35 ms	750.22 ms	+3.4%
🟠	hbs large	815.97 ms	844.56 ms	+3.5%

🟢 faster · 🔴 slower · 🟠 slightly slower · ⚪ within 2%

Full mitata output

clk: ~3.10 GHz
cpu: AMD EPYC 7763 64-Core Processor
runtime: node 24.18.0 (x64-linux)

benchmark                   avg (min … max) p75 / p99    (min … top 1%)
------------------------------------------- -------------------------------
gts small (control)          986.20 ms/iter    1.01 s       █              
                       (838.34 ms … 1.30 s)    1.13 s ▅  ▅▅▅█ ▅ ▅ ▅  ▅    ▅
                    ( 13.07 mb …  88.44 mb)  42.57 mb █▁▁████▁█▁█▁█▁▁█▁▁▁▁█

gts small (experiment)       946.45 ms/iter 980.60 ms   █                  
                       (819.28 ms … 1.30 s)    1.09 s   █                  
                    ( 12.56 mb …  83.89 mb)  41.88 mb ███▁▁▁██▁▁▁██▁█▁▁▁▁▁█

                             ┌                                            ┐
                                ╷       ┌────────────┬──┐                 ╷
         gts small (control)    ├───────┤            │  ├─────────────────┤
                                ╵       └────────────┴──┘                 ╵
                             ╷   ┌─────────────┬────┐               ╷
      gts small (experiment) ├───┤             │    ├───────────────┤
                             ╵   └─────────────┴────┘               ╵
                             └                                            ┘
                             819.28 ms           975.81 ms           1.13 s

summary
  gts small (experiment)
   1.04x faster than gts small (control)

------------------------------------------- -------------------------------
gts medium (control)            1.69 s/iter    1.72 s  █     █             
                          (1.54 s … 2.09 s)    1.87 s ▅█▅▅ ▅ █   ▅      ▅ ▅
                    ( 80.02 mb …  99.53 mb)  84.32 mb ████▁█▁█▁▁▁█▁▁▁▁▁▁█▁█

gts medium (experiment)         1.66 s/iter    1.69 s     █                
                          (1.50 s … 2.19 s)    1.71 s ▅   █▅▅  ▅ ▅ ▅    ▅▅▅
                    ( 79.72 mb …  99.72 mb)  84.29 mb █▁▁▁███▁▁█▁█▁█▁▁▁▁███

                             ┌                                            ┐
                                 ╷  ┌──────────────┬────┐                 ╷
        gts medium (control)     ├──┤              │    ├─────────────────┤
                                 ╵  └──────────────┴────┘                 ╵
                             ╷    ┌─────────────┬────┐ ╷
     gts medium (experiment) ├────┤             │    ├─┤
                             ╵    └─────────────┴────┘ ╵
                             └                                            ┘
                             1.50 s              1.69 s              1.87 s

summary
  gts medium (experiment)
   1.02x faster than gts medium (control)

------------------------------------------- -------------------------------
gts large (control)          852.69 ms/iter 845.01 ms       ███     █      
                       (779.39 ms … 1.21 s) 875.81 ms ▅  ▅  ███     █     ▅
                    ( 14.87 mb …  86.21 mb)  33.23 mb █▁▁█▁▁███▁▁▁▁▁█▁▁▁▁▁█

gts large (experiment)       877.88 ms/iter 881.67 ms    █                 
                       (789.08 ms … 1.22 s) 981.98 ms    █                 
                    ( 15.01 mb …  87.11 mb)  38.29 mb ████▁█▁█▁▁█▁▁█▁▁▁▁▁▁█

                             ┌                                            ┐
                             ╷     ┌─────────┬    ╷
         gts large (control) ├─────┤         │────┤
                             ╵     └─────────┴    ╵
                               ╷    ┌──────────────┬┐                     ╷
      gts large (experiment)   ├────┤              │├─────────────────────┤
                               ╵    └──────────────┴┘                     ╵
                             └                                            ┘
                             779.39 ms         880.68 ms          981.98 ms

summary
  gts large (control)
   1.03x faster than gts large (experiment)

------------------------------------------- -------------------------------
gjs small (control)          827.98 ms/iter 831.59 ms       █              
                       (752.58 ms … 1.11 s) 903.69 ms  █    █              
                    ( 29.21 mb …  85.87 mb)  37.07 mb ████▁▁█▁▁▁█▁▁▁▁▁█▁▁▁█

gjs small (experiment)       823.65 ms/iter 838.91 ms        █             
                       (736.67 ms … 1.11 s) 912.31 ms    █   █    █        
                    ( 29.42 mb …  85.80 mb)  37.09 mb █▁███▁▁█▁▁▁▁█▁▁▁▁▁▁▁█

                             ┌                                            ┐
                                 ╷ ┌────────────────┬┐                  ╷
         gjs small (control)     ├─┤                │├──────────────────┤
                                 ╵ └────────────────┴┘                  ╵
                             ╷     ┌───────────────┬───┐                  ╷
      gjs small (experiment) ├─────┤               │   ├──────────────────┤
                             ╵     └───────────────┴───┘                  ╵
                             └                                            ┘
                             736.67 ms         824.49 ms          912.31 ms

summary
  gjs small (experiment)
   1.01x faster than gjs small (control)

------------------------------------------- -------------------------------
gjs medium (control)            1.44 s/iter    1.44 s █     █              
                          (1.33 s … 1.83 s)    1.50 s █  ▅ ▅█ ▅  ▅ ▅  ▅   ▅
                    ( 21.00 mb …  88.88 mb)  71.13 mb █▁▁█▁██▁█▁▁█▁█▁▁█▁▁▁█

gjs medium (experiment)         1.42 s/iter    1.44 s       █              
                          (1.30 s … 1.85 s)    1.46 s       █              
                    ( 18.94 mb …  88.99 mb)  70.53 mb █▁▁▁█▁██▁▁█▁▁▁█▁▁█▁██

                             ┌                                            ┐
                                    ╷     ┌─────────────────┬┐            ╷
        gjs medium (control)        ├─────┤                 │├────────────┤
                                    ╵     └─────────────────┴┘            ╵
                             ╷          ┌───────────────┬──┐     ╷
     gjs medium (experiment) ├──────────┤               │  ├─────┤
                             ╵          └───────────────┴──┘     ╵
                             └                                            ┘
                             1.30 s              1.40 s              1.50 s

summary
  gjs medium (experiment)
   1.01x faster than gjs medium (control)

------------------------------------------- -------------------------------
gjs large (control)          680.03 ms/iter 647.81 ms   █                  
                    (617.35 ms … 975.45 ms) 784.17 ms   █▅                 
                    (219.87 kb …  69.98 mb)  43.76 mb ▇▇██▇▁▁▁▇▁▁▁▁▁▁▁▁▁▁▁▇

gjs large (experiment)       669.27 ms/iter 640.83 ms   ███                
                    (595.59 ms … 983.07 ms) 788.31 ms ▅▅███▅  ▅           ▅
                    (886.13 kb …  70.13 mb)  43.78 mb ██████▁▁█▁▁▁▁▁▁▁▁▁▁▁█

                             ┌                                            ┐
                                  ╷   ┌──────────┬                       ╷
         gjs large (control)      ├───┤          │───────────────────────┤
                                  ╵   └──────────┴                       ╵
                             ╷    ┌───────────┬                           ╷
      gjs large (experiment) ├────┤           │───────────────────────────┤
                             ╵    └───────────┴                           ╵
                             └                                            ┘
                             595.59 ms         691.95 ms          788.31 ms

summary
  gjs large (experiment)
   1.02x faster than gjs large (control)

------------------------------------------- -------------------------------
hbs small (control)          148.24 ms/iter 154.02 ms       █           █  
                    (113.03 ms … 288.91 ms) 159.20 ms ▅ ▅  ▅█▅   ▅  ▅   █ ▅
                    ( 51.37 mb …  55.90 mb)  53.49 mb █▁█▁▁███▁▁▁█▁▁█▁▁▁█▁█

hbs small (experiment)       149.59 ms/iter 151.11 ms           █          
                    (111.17 ms … 291.62 ms) 159.99 ms           █     █   █
                    ( 51.22 mb …  55.69 mb)  53.51 mb █▁█▁▁██▁▁▁█▁▁▁▁▁█▁▁▁█

                             ┌                                            ┐
                               ╷         ┌─────────────────────┬────┐    ╷
         hbs small (control)   ├─────────┤                     │    ├────┤
                               ╵         └─────────────────────┴────┘    ╵
                             ╷          ┌───────────────────────┬─┐       ╷
      hbs small (experiment) ├──────────┤                       │ ├───────┤
                             ╵          └───────────────────────┴─┘       ╵
                             └                                            ┘
                             111.17 ms         135.58 ms          159.99 ms

summary
  hbs small (control)
   1.01x faster than hbs small (experiment)

------------------------------------------- -------------------------------
hbs medium (control)         757.70 ms/iter 745.29 ms    █ █  █           █
                    (699.51 ms … 992.38 ms) 795.09 ms ▅  █▅█  █ ▅         █
                    (  8.88 mb …  81.68 mb)  38.01 mb █▁▁███▁▁█▁█▁▁▁▁▁▁▁▁▁█

hbs medium (experiment)      779.92 ms/iter 782.95 ms ██  █                
                    (727.45 ms … 997.89 ms) 840.46 ms ██▅ █   ▅ ▅  ▅      ▅
                    (  8.57 mb …  81.81 mb)  32.61 mb ███▁█▁▁▁█▁█▁▁█▁▁▁▁▁▁█

                             ┌                                            ┐
                             ╷   ┌──────────────┬           ╷
        hbs medium (control) ├───┤              │───────────┤
                             ╵   └──────────────┴           ╵
                                      ╷ ┌──────────────┬┐                 ╷
     hbs medium (experiment)          ├─┤              │├─────────────────┤
                                      ╵ └──────────────┴┘                 ╵
                             └                                            ┘
                             699.51 ms         769.99 ms          840.46 ms

summary
  hbs medium (control)
   1.03x faster than hbs medium (experiment)

------------------------------------------- -------------------------------
hbs large (control)          837.80 ms/iter 819.69 ms      █               
                       (797.73 ms … 1.06 s) 869.97 ms   █  █               
                    ( 53.35 mb …  64.87 mb)  55.49 mb █▁█████▁▁█▁▁▁▁▁▁▁▁▁▁█

hbs large (experiment)       868.41 ms/iter 861.36 ms       ██             
                       (817.20 ms … 1.09 s) 899.91 ms       ██             
                    ( 53.47 mb …  64.84 mb)  55.48 mb ██▁▁▁▁██▁▁▁██▁▁▁▁▁▁▁█

                             ┌                                            ┐
                             ╷   ┌─────────────┬             ╷
         hbs large (control) ├───┤             │─────────────┤
                             ╵   └─────────────┴             ╵
                                      ╷         ┌───────────┬             ╷
      hbs large (experiment)          ├─────────┤           │─────────────┤
                                      ╵         └───────────┴             ╵
                             └                                            ┘
                             797.73 ms         848.82 ms          899.91 ms

summary
  hbs large (control)
   1.04x faster than hbs large (experiment)

### Project mode (type-aware)

	Benchmark	Control (p50)	Experiment (p50)	Δ
🟢	project warm parse x150	618.74 ms	580.41 ms	-6.2%
⚪	project parse+typed x150	624.33 ms	625.00 ms	+0.1%
⚪	projectService warm parse x150	607.25 ms	616.43 ms	+1.5%
🔴	projectService parse+typed x150	705.66 ms	809.54 ms	+14.7%

🟢 faster · 🔴 slower · 🟠 slightly slower · ⚪ within 2%

Full mitata output

clk: ~2.91 GHz
cpu: AMD EPYC 7763 64-Core Processor
runtime: node 24.18.0 (x64-linux)

benchmark                                   avg (min … max) p75 / p99    (min … top 1%)
----------------------------------------------------------- -------------------------------
project warm parse x150 (control)            614.53 ms/iter 622.13 ms                █     
                                    (584.58 ms … 663.41 ms) 630.12 ms ▆▆▁▁▆▆▁▁▁▁▁▁▁▁▆█▆▆▁▁▆
                                  gc(290.31 ms … 310.97 ms) 113.10 mb (109.83 mb…114.81 mb)

project warm parse x150 (experiment)         586.47 ms/iter 598.11 ms █   █             █  
                                    (572.90 ms … 606.59 ms) 600.95 ms █▁█▁██▁█▁▁▁▁█▁▁▁▁▁█▁█
                                  gc(282.61 ms … 318.06 ms) 110.61 mb (107.13 mb…112.49 mb)

                                             ┌                                            ┐
                                                      ╷      ┌────────────────┬─────┐     ╷
           project warm parse x150 (control)          ├──────┤                │     ├─────┤
                                                      ╵      └────────────────┴─────┘     ╵
                                             ╷ ┌────────┬────────┐ ╷
        project warm parse x150 (experiment) ├─┤        │        ├─┤
                                             ╵ └────────┴────────┘ ╵
                                             └                                            ┘
                                             572.90 ms         601.51 ms          630.12 ms

summary
  project warm parse x150 (experiment)
   1.05x faster than project warm parse x150 (control)

----------------------------------------------------------- -------------------------------
project parse+typed x150 (control)           626.15 ms/iter 634.56 ms █  █   █ ██ ██ ███  █
                                    (598.70 ms … 660.14 ms) 642.64 ms █▁▁█▁▁▁█▁██▁██▁███▁▁█
                                  gc(291.42 ms … 305.84 ms) 118.85 mb (117.83 mb…122.53 mb)

project parse+typed x150 (experiment)        626.62 ms/iter 638.11 ms █                  █ 
                                    (601.25 ms … 671.06 ms) 640.23 ms █▁▁█▁▁▁██▁▁▁█▁▁█▁█▁██
                                  gc(297.53 ms … 315.91 ms) 117.24 mb (116.11 mb…121.95 mb)

                                             ┌                                            ┐
                                             ╷               ┌───────────┬────────┐       ╷
          project parse+typed x150 (control) ├───────────────┤           │        ├───────┤
                                             ╵               └───────────┴────────┘       ╵
                                                ╷    ┌────────────────────┬──────────┐  ╷
       project parse+typed x150 (experiment)    ├────┤                    │          ├──┤
                                                ╵    └────────────────────┴──────────┘  ╵
                                             └                                            ┘
                                             598.70 ms         620.67 ms          642.64 ms

summary
  project parse+typed x150 (control)
   1x faster than project parse+typed x150 (experiment)

----------------------------------------------------------- -------------------------------
projectService warm parse x150 (control)     616.90 ms/iter 621.83 ms ██          █        
                                    (596.99 ms … 671.85 ms) 640.01 ms ██▁▁██▁▁▁▁███▁▁▁▁▁▁▁█
                                  gc(301.96 ms … 373.07 ms) 113.59 mb (108.46 mb…116.12 mb)

projectService warm parse x150 (experiment)  630.43 ms/iter 629.82 ms   █                  
                                    (612.34 ms … 670.12 ms) 659.80 ms ▅▅█▁▁▁▅█▁▁▁▁▁▁▁▁▁▁▁▅▅
                                  gc(304.66 ms … 327.75 ms) 111.61 mb (110.98 mb…111.95 mb)

                                             ┌                                            ┐
                                             ╷ ┌───────────┬───┐            ╷
    projectService warm parse x150 (control) ├─┤           │   ├────────────┤
                                             ╵ └───────────┴───┘            ╵
                                                        ╷  ┌─────────┬                    ╷
 projectService warm parse x150 (experiment)            ├──┤         │────────────────────┤
                                                        ╵  └─────────┴                    ╵
                                             └                                            ┘
                                             596.99 ms         628.40 ms          659.80 ms

summary
  projectService warm parse x150 (control)
   1.02x faster than projectService warm parse x150 (experiment)

----------------------------------------------------------- -------------------------------
projectService parse+typed x150 (control)    708.68 ms/iter 718.40 ms             █ █     █
                                    (652.99 ms … 753.50 ms) 737.68 ms █▁▁▁▁▁▁█▁██▁█▁██▁▁▁▁█
                                  gc(308.80 ms … 347.28 ms) 119.58 mb (118.80 mb…124.31 mb)

projectService parse+typed x150 (experiment) 807.58 ms/iter 815.93 ms         █     ▃      
                                    (766.57 ms … 849.41 ms) 833.31 ms ▆▁▁▁▁▆▁▁█▁▁▁▁▆█▆▁▁▁▆▆
                                  gc(342.64 ms … 369.56 ms) 118.38 mb (117.24 mb…122.26 mb)

                                             ┌                                            ┐
                                             ╷         ┌───┬─┐    ╷
   projectService parse+typed x150 (control) ├─────────┤   │ ├────┤
                                             ╵         └───┴─┘    ╵
                                                                         ╷      ┌───┬─┐   ╷
projectService parse+typed x150 (experiment)                             ├──────┤   │ ├───┤
                                                                         ╵      └───┴─┘   ╵
                                             └                                            ┘
                                             652.99 ms         743.15 ms          833.31 ms

summary
  projectService parse+typed x150 (control)
   1.14x faster than projectService parse+typed x150 (experiment)

Three assertions, each under both project and projectService (feature-detected; v7's experimental service can't hold .gts files in-project, so the mode is registered only when a probe parse yields a program that contains virtual twins — 3 tests on tsee 7, 6 on tsee 8): - every virtual twin present shares statements with the .gts root, - a private-member class has one type identity across import paths, - an already-synced program is not re-grafted on later parses. Against the unfixed parser these fail — demonstrated on CI in ember-tooling#240: the re-graft canaries fail on TS 5.7, and the twin-graft/type-identity tests additionally fail on TS 6 cells where the .mts twin shadows the old single-candidate fallback. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

NullVoxPopuli-ai-agent · 2026-07-06T17:08:17Z

Served its purpose: the CI run on this PR demonstrates these tests failing against the unfixed parser — Tests / Tests (Node 20) red with the re-graft canaries failing on TS 5.7 (the TS 6 matrix cells, where the twin-graft and type-identity tests also fail, were skipped by fail-fast). The tests are folded back into #235, where they pass against the fix.

NullVoxPopuli-ai-agent mentioned this pull request Jul 6, 2026

Graft .gts SourceFiles onto every virtual twin, once per program #235

Closed

NullVoxPopuli-ai-agent closed this Jul 6, 2026

NullVoxPopuli-ai-agent deleted the test/virtual-twin-regression branch July 6, 2026 17:08

Navigation Menu

Uh oh!

Add virtual-twin type-identity regression tests (#229) - #240

Add virtual-twin type-identity regression tests (#229)#240
NullVoxPopuli-ai-agent wants to merge 1 commit into
ember-tooling:mainfrom
NullVoxPopuli-ai-agent:test/virtual-twin-regression

NullVoxPopuli-ai-agent commented Jul 6, 2026

Uh oh!

github-actions Bot commented Jul 6, 2026

Uh oh!

NullVoxPopuli-ai-agent commented Jul 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

NullVoxPopuli-ai-agent commented Jul 6, 2026

What's covered

Expected failures without #235

Uh oh!

github-actions Bot commented Jul 6, 2026

🏎️ Benchmark Comparison

Parse

Uh oh!

NullVoxPopuli-ai-agent commented Jul 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants