chore: support Ubuntu 26.04

[lvaroqui]

Run the CI on Ubuntu 26.04 alongside 24.04: the unit tests and the basic run test on ubuntu-24.04 and ubuntu-26.04 runners, and the executor tests on ubuntu:24.04 and ubuntu:26.04 base images.

The executor tests run in Docker, so the Dockerfile now takes a BASE_IMAGE build argument, set by tests/docker/run.sh from CODSPEED_TEST_BASE_IMAGE (default ubuntu:24.04). Each base image gets its own image tags, setup container and target volume: binaries built against 26.04's glibc do not run in a 24.04 container, and the setup image cleanup would otherwise delete the other base's image.

The workflow sets CODSPEED_VALGRIND_BUILD_FROM_SOURCE=false (forwarded into the test containers), so a system we publish no valgrind package for fails instead of falling back to a source build that takes minutes and hides the missing package. No valgrind-codspeed package is published for 26.04 yet, so the simulation runs on 26.04 are expected to fail until it is: the 14 simulation executor tests and the whole basic-run-test (ubuntu-26.04) job. Note that codspeed setup only warns when valgrind cannot be installed, so the failure surfaces at codspeed run, not at setup.

With the fallback left on, the full executor suite passes locally on ubuntu:26.04 (valgrind-codspeed built from source, every shard green).

Refs COD-3305, which stays open until valgrind-codspeed is published for 26.04.

[codspeed]

Merging this PR will degrade performance by 70.23%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ 8 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• sleep 1
• encode_events_realistic[4]
• encode_events_realistic[2]
• encode_events_realistic[16]
• find_module_events_with_stacks[1000000]
• find_module_events[1000000]
• encode_events_realistic[8]
• encode_events_realistic[1]

ℹ️ 6 benchmarks have no flame graph

The flamegraph could not be generated because we failed to resolve the symbols in the profiling trace. This happens when we fail to harvest information about language frames.

• find_module_events_with_stacks[100000000]
• find_module_events[10000000]
• find_module_events[100000000]
• find_module_events_with_stacks[10000000]
• find_module_events[1000000]
• find_module_events_with_stacks[1000000]

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
❌ 2 (👁 2) regressed benchmarks
✅ 38 untouched benchmarks
⏩ 6 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	WallTime	memtrack track dd	1.2 s	1.1 s	+11.72%
👁	Memory	sleep 1	15.6 KB	103.4 KB	-84.92%
👁 ⚠️	Simulation	sleep 1	245.4 µs	1,566.4 µs	-84.33%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing cod-3305-support-ubuntu-2604 (4f33c00) with main (5ff86d7)}

Footnotes

1. 6 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[greptile-apps]

Confidence Score: 5/5

[High impact] No new actionable issue remains; the README update appears safe to merge.

Summary

Adds Ubuntu 26.04 coverage, package pins, and support documentation.

• CodSpeed recognizes Ubuntu 26.04 when choosing Valgrind packages.
• CI now runs CodSpeed checks on Ubuntu 26.04.
• Docker executor tests can use separate Ubuntu base images.
• Sudo keeps command environment variables on Ubuntu 26.04.

Previous-thread context:

• lvaroqui accepted running unit tests on one Ubuntu version because they do not depend on the OS. The host smoke-test coverage was restored.
• Both Ubuntu 26.04 targets now appear in ALL_VALGRIND_TARGETS.
• lvaroqui accepted omitting the source-build setting from the setup-image hash because CI disables source builds and does not cache Docker images. The previous thread treated the local-cache concern as non-actionable.
• lvaroqui marked missing-package failures on Ubuntu 26.04 as expected until publication.

_{Reviews (5) · Last reviewed commit: "docs: list Ubuntu 26.04 as officially su..." · Reviewed by Greptile}

[GuillaumeLagrange]

olgtm