Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 61.74%. Comparing base (e4f6598) to head (a78c4ee).
⚠️ Report is 401 commits behind head on master.

@@            Coverage Diff             @@
##           master    #3380      +/-   ##
==========================================
- Coverage   61.87%   61.74%   -0.13%     
==========================================
  Files         141      141              
  Lines       12481    12481              
  Branches     1630     1630              
==========================================
- Hits         7722     7706      -16     
- Misses       4038     4054      +16     
  Partials      721      721              

Benchmarks [ tracer ]

Benchmark execution time: 2025-09-18 17:45:37

Comparing candidate commit a78c4ee in PR branch bob/static-integrations with baseline commit e4f6598 in branch master.

Found 11 performance improvements and 2 performance regressions! Performance is the same for 181 metrics, 0 unstable metrics.

scenario:ComposerTelemetryBench/benchTelemetryParsing-opcache

• 🟩 execution_time [-1.880µs; -0.920µs] or [-13.526%; -6.618%]

scenario:ContextPropagationBench/benchExtractHeaders64Bit

• 🟩 execution_time [-93.400ns; -30.600ns] or [-7.575%; -2.482%]

scenario:HookBench/benchHookOverheadInstallHookOnFunction

• 🟩 execution_time [-116.111µs; -102.024µs] or [-19.750%; -17.354%]

scenario:HookBench/benchHookOverheadInstallHookOnFunction-opcache

• 🟩 execution_time [-89.726µs; -73.826µs] or [-16.268%; -13.385%]

scenario:HookBench/benchHookOverheadInstallHookOnMethod

• 🟩 execution_time [-98.362µs; -83.384µs] or [-16.495%; -13.984%]

scenario:HookBench/benchHookOverheadInstallHookOnMethod-opcache

• 🟩 execution_time [-89.046µs; -67.458µs] or [-16.035%; -12.147%]

scenario:LogsInjectionBench/benchLogsInfoInjection

• 🟩 execution_time [-680.155ns; -490.845ns] or [-9.384%; -6.772%]

scenario:LogsInjectionBench/benchLogsInfoInjection-opcache

• 🟩 execution_time [-709.753ns; -340.647ns] or [-7.483%; -3.591%]

scenario:MessagePackSerializationBench/benchMessagePackSerialization

• 🟩 execution_time [-7.943µs; -7.077µs] or [-7.259%; -6.467%]

scenario:MessagePackSerializationBench/benchMessagePackSerialization-opcache

• 🟩 execution_time [-8.603µs; -7.517µs] or [-7.724%; -6.749%]

scenario:PHPRedisBench/benchRedisOverhead

• 🟥 mem_peak [+209.344KB; +209.344KB] or [+4.616%; +4.616%]

scenario:PHPRedisBench/benchRedisOverhead-opcache

• 🟥 mem_peak [+207.536KB; +207.536KB] or [+9.464%; +9.464%]

scenario:TraceSerializationBench/benchSerializeTrace

• 🟩 execution_time [-64.122µs; -53.578µs] or [-13.582%; -11.349%]

Benchmarks [ profiler ]

Benchmark execution time: 2025-09-18 16:54:28

Comparing candidate commit a78c4ee in PR branch bob/static-integrations with baseline commit e4f6598 in branch master.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 28 metrics, 8 unstable metrics.

Hey @bwoebi!

I did not investigate too much, but I think this PR is a cause of a change in behaviour in hooks using static methods.

We are using Instrumentation class that installs all hooks on specific classes/interfaces, something like:

<?php 

final class Instrumentation
{
    public static function commandExecutors(): void
    {
        install_hook(
            CommandExecutor::class.'::execute',
            static function (HookData $hook): void {
                $span = $hook->span();
                // ...
                $span->meta['command_executor.class'] = self::class;
                // ...
            },
        );
    }
}

Previously self would refer to the actual CommandHandler object class, just like $this in non-static closure. Now self returns Instrumentation class.

Is this expected?

@marmichalski Yes, that behaviour is now expected, if you want to use self:: (and have it refer to the other class), just remove static.
There was quite some overhead from rebinding Closures to class scopes, so static Closures are no longer rebound - with the option to have non-static closures bind.