[tivris]

Related Issue(s)	Fixes #6211
Has Unit Tests (y/n)	y
Documentation Included (y/n)	n
Generative AI was used in this contribution (y/n)	y

---

Change Description

Increases the receive timeout that force_recv_timeout in Drv/Ip/test/ut/SocketTestHelper.cpp sets from 50 us to 50 ms. The comment added with it in 8b4e870082 said 50 ms; #5705 reworded the comment to 50 us without changing the value.

Rationale

With #5706's 10 ms polling waits, 50 us receive timeouts overflow the reconnect-test histories on macOS (#6211). Doubling those histories in #5724 was not enough; at 50 ms the measured runs record 20 to 36 entries of the 2000.

Testing/Review Recommendations

macOS (Apple M5):

• Drv_TcpServer_ut_exe, Drv_Udp_ut_exe, Drv_TcpClient_ut_exe, Drv_Ip_Tcp_test, Drv_Ip_Udp_test: 20 of 20 runs pass for each, at normal priority and under taskpolicy -c background. Unmodified devel: TcpServer and Udp abort in 20 of 20 under background.
• fprime-util check --all: 135 of 135 pass.
• clang-format --dry-run -Werror and git diff --check pass.

Linux (arm64 VM): Drv_Udp_ut_exe takes about 0.5 s longer; the other four are unchanged.

Expected, not measured with gcov: Drv/TcpServer line coverage may drop from 64/74 to 63/74, because TcpServerComponentImpl.cpp:62 (RECV_NO_DATA) was only reached through the incidental 50 us timeouts.

Future Work

A deterministic RECV_NO_DATA test for TcpServer would cover that line again.

Contributor Checklist

• I have the rights to contribute this change set under the Apache-2.0 license.
• I agree to be bound by the contributor guidelines.

AI Usage (see policy)

• This contribution follows the F´ AI development skills (entered through AGENTS.md).

The change is AI-generated.

[thomas-bc]

Small and efficient, much appreciated. Thanks.

[thomas-bc]

I'm going to merge #6217 first and see if (1) the error shows up in devel CI and (2) this PR fixes it

[thomas-bc]

That fixed it ! 🎉