[marius-bughiu]

Standalone Visual Basic source repeats the cache-array lookup in every expression-ID branch. Retrieve the array once before dispatch when IDs are dense, then keep each expression's existing cache slot and lazy data-context construction. Guard invalid IDs before touching the activity context. Empty/sparse IDs, C#, partial classes and the location-list overload retain their previous paths. This cohort shows lower allocations but no whole-process elapsed-time improvement; the measured time increases are shown below.

Stack 4/4; base branch codex/corewf-03-share-vb-cache-helpers. Cumulative results start at public CoreWF commit 65caba7c7fc7d141e847775b00f2c9558e8daa59. None of these optimizations relies on duplicate expression text.

Measured on six synthetic authored cases with scenario-local variables and mostly unique expressions, with execution templates assigned versus removed. All validation and analyzer rules are enabled. Each of the five source-built CoreWF states ran twice per variant in forward/reverse order, serially on the same machine and compiler host. Every run deleted the isolated project .local first. The actual loaded CoreWF image hashes, input inventories, diagnostics and output coverage were verified.

Adjacent change (preceding stack state → this state; medians of two runs):

Metric	Templates	Before	After	Delta
Whole process	With	24 s	25 s	+0.684 s
Whole-process managed allocation	With	9.8 GB	9.6 GB	-169.73 MB
Compile	With	14 s	13 s	-0.273 s
Compile managed allocation	With	5.5 GB	5.3 GB	-135.41 MB
ExpressionsCompiler.CompileExpressions	With	8.20 s	7.43 s	-0.769 s
Source-generation thread allocation	With	640.62 MB	608.38 MB	-32.24 MB
Source-parsing thread allocation	With	141.48 MB	138.55 MB	-2.94 MB
Generated source	With	43,327,093 characters	40,538,754 characters	-2,788,339 characters
Peak sampled private memory	With	2.5 GB	2.4 GB	-134.06 MB
Whole process	Without	22 s	24 s	+1.65 s
Whole-process managed allocation	Without	9.4 GB	9.4 GB	-65.63 MB
Compile	Without	12 s	13 s	+1.10 s
Compile managed allocation	Without	5.3 GB	5.2 GB	-125.59 MB
ExpressionsCompiler.CompileExpressions	Without	5.78 s	7.52 s	+1.74 s
Source-generation thread allocation	Without	638.09 MB	605.67 MB	-32.42 MB
Source-parsing thread allocation	Without	139.05 MB	134.25 MB	-4.80 MB
Generated source	Without	43,158,131 characters	40,368,994 characters	-2,789,137 characters
Peak sampled private memory	Without	2.4 GB	2.2 GB	-206.06 MB

Whole-process observations behind the adjacent medians: With templates: before [23 s, 26 s], after [26 s, 24 s]. Without templates: before [23 s, 22 s], after [23 s, 25 s].

Cumulative change (original CoreWF state → this state; same cohort):

Metric	Templates	Before	After	Delta
Whole process	With	28 s	25 s	-2.36 s
Whole-process managed allocation	With	10.0 GB	9.6 GB	-393.77 MB
ValidateProject	With	2.76 s	2.64 s	-0.117 s
ValidateProject managed allocation	With	1.5 GB	1.5 GB	-57.43 MB
AnalyzeProject	With	4.18 s	3.53 s	-0.651 s
AnalyzeProject managed allocation	With	2.1 GB	2.1 GB	-2.62 MB
Compile	With	15 s	13 s	-1.52 s
Compile managed allocation	With	5.7 GB	5.3 GB	-331.49 MB
Source-generation thread allocation	With	689.36 MB	608.38 MB	-80.98 MB
Source-parsing thread allocation	With	147.00 MB	138.55 MB	-8.45 MB
Generated source	With	46,395,569 characters	40,538,754 characters	-5,856,815 characters
Peak sampled private memory	With	2.5 GB	2.4 GB	-94.87 MB
Whole process	Without	30 s	24 s	-5.97 s
Whole-process managed allocation	Without	9.6 GB	9.4 GB	-251.36 MB
ValidateProject	Without	3.39 s	2.60 s	-0.790 s
ValidateProject managed allocation	Without	1.5 GB	1.5 GB	+34.88 MB
AnalyzeProject	Without	5.90 s	3.70 s	-2.20 s
AnalyzeProject managed allocation	Without	2.1 GB	2.1 GB	+1.22 MB
Compile	Without	15 s	13 s	-1.71 s
Compile managed allocation	Without	5.5 GB	5.2 GB	-287.83 MB
Source-generation thread allocation	Without	686.70 MB	605.67 MB	-81.03 MB
Source-parsing thread allocation	Without	147.97 MB	134.25 MB	-13.72 MB
Generated source	Without	46,211,967 characters	40,368,994 characters	-5,842,973 characters
Peak sampled private memory	Without	2.4 GB	2.2 GB	-211.01 MB

Validation: The exact publication commit passed 38 cumulative focused cases and 57 existing cases. This evolves the preceding source-generation test class, preserves its execution checks and root-only helper assertion, and adds invalid-ID/empty-input coverage. Sixteen behavior cases passed on the preceding implementation. The statement test preserves an existing generator rejection; it does not claim statement execution support. Release owner builds used SDK 8.0.426 and net6.0-windows; the compiler host used .NET 10.0.11. The complete CoreWF test suite was not run. Build copies disabled the repository's package-generation/global-cache-cleanup postbuild actions; no such project changes are included in this PR.

These are descriptive six-case diagnostics, not a 1,000-case throughput result or a statistical confidence claim. Host load, GC and run order can affect elapsed time; increases are retained in the tables. Managed allocation is not retained heap size. Nested phases overlap and must not be summed. Deltas use unrounded values; displayed values are rounded independently. Matching compiler output coverage does not establish arbitrary workflow execution or persistence equivalence.

Note

This change and description were prepared with AI assistance; the measurements and test results above were collected locally.