cmake: build `stubgss` library for libtests to match autotools #17653

vszakats · 2025-06-17T18:24:01Z

Used by test 2056 and 2057, in a way that's Linux- & autotools-specific.
This patch builds it for all Unix, syncing cmake with autotools.

Adapt the two tests to find the library in CMake builds as well.

Tested OK on Linux. (CI does not test this. The corresponding jobs build
in debug mode, while the LD_PRELOAD feature is locked to non-debug.)

On macOS it didn't load without building everything for aarch64e arch:
"../bld/tests/libtest/libstubgss.dylib' (mach-o file, but is
an incompatible architecture (have 'arm64', need 'arm64e'))"
With that fixed it still did not load correctly and/or the tests did not
pass. So, for macOS these tests remain disabled.

Also:

GHA/macos: build for aarch64e. (recognized by Apple clang as of this
patch. llvm and gcc fall back to aarch64.)

Follow-up to 56d949d #1687

make test configs compatible with CMake and macOS

testclutch · 2025-06-19T16:18:27Z

Analysis of PR #17653 at 86202575:

Test 2056 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Note that this test has failed in 22 different CI jobs (the link just goes to one of them).

Test 2057 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Note that this test has failed in 20 different CI jobs (the link just goes to one of them).

Generated by Testclutch

https://github.com/curl/curl/actions/runs/15761470154/job/44428953936?pr=17653 ``` test 2056...[HTTP Negotiate authentication (stub krb5)] 2056: protocol FAILED! There was no content at all in the file log/16/server.input. Server glitch? Total curl failure? Returned: 2006 == Contents of files in the log/16/ dir after test 2056 === Start of file commands.log ../src/curl -q --output log/16/curl2056.out --include --trace-ascii log/16/trace2056 --trace-config all --trace-time --negotiate http://127.0.0.1:50648/2056 > log/16/stdout2056 2> log/16/stderr2056 === End of file commands.log === Start of file http_server.log 15:33:30.366826 ====> Client connect dyld[17773]: terminating because inserted dylib 'libstubgss.dylib' could not be loaded: tried: '/Users/runner/work/curl/curl/bld/tests/libtest/.libs/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/libstubgss.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'arm64e')), 'libstubgss.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OSlibstubgss.dylib' (no such file), 'libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/.libs/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/libstubgss.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'arm64e')), '/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file) dyld[17773]: tried: '/Users/runner/work/curl/curl/bld/tests/libtest/.libs/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/libstubgss.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'arm64e')), 'libstubgss.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OSlibstubgss.dylib' (no such file), 'libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/.libs/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libtest/libstubgss.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'arm64e')), '/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file), '/Users/runner/work/curl/curl/bld/tests/libstubgss.dylib' (no such file) ``` https://github.com/curl/curl/actions/runs/15761470154/job/44428953936?pr=17653#step:16:3183

https://github.com/curl/curl/actions/runs/15762896027/job/44433397386?pr=17653#step:16:3248 ``` test 2056...[HTTP Negotiate authentication (stub krb5)] 2056: protocol FAILED! There was no content at all in the file log/4/server.input. Server glitch? Total curl failure? Returned: 2006 == Contents of files in the log/4/ dir after test 2056 === Start of file commands.log ../src/curl -q --output log/4/curl2056.out --include --trace-ascii log/4/trace2056 --trace-config all --trace-time --negotiate http://127.0.0.1:49216/2056 > log/4/stdout2056 2> log/4/stderr2056 === End of file commands.log === Start of file http_server.log 16:59:26.108155 ====> Client connect 16:59:26.137818 accept_connection 3 returned 4 16:59:26.137876 accept_connection 3 returned 0 16:59:26.138003 Read 97 bytes 16:59:26.140198 Process 97 bytes request 16:59:26.140253 Got request: GET /verifiedserver HTTP/1.1 16:59:26.140325 Are-we-friendly question received 16:59:26.140589 Wrote request (97 bytes) input to log/4/server.input 16:59:26.140800 Identifying ourselves as friends 16:59:26.173804 Response sent (55 bytes) and written to log/4/server.response 16:59:26.173842 special request received, no persistency 16:59:26.237431 ====> Client disconnect 0 === End of file http_server.log === Start of file http_verify.log * Trying 127.0.0.1:49216... * Connected to 127.0.0.1 (127.0.0.1) port 49216 * using HTTP/1.x > GET /verifiedserver HTTP/1.1 > Host: 127.0.0.1:49216 > User-Agent: curl/8.15.0-DEV > Accept: */* > * Request completely sent off < HTTP/1.1 200 OK < Content-Length: 16 < { [16 bytes data] * Connection #0 to host 127.0.0.1 left intact === End of file http_verify.log === Start of file http_verify.out WE ROOLZ: 7432 === End of file http_verify.out === Start of file server.cmd Testnum 2056 === End of file server.cmd === Start of file server.response HTTP/1.1 200 OK Content-Length: 16 WE ROOLZ: 7432 === End of file server.response test 2057...[HTTP Negotiate authentication (stub NTLM)] 2057: protocol FAILED! There was no content at all in the file log/19/server.input. Server glitch? Total curl failure? Returned: 2006 == Contents of files in the log/19/ dir after test 2057 === Start of file commands.log ../src/curl -q --output log/19/curl2057.out --include --trace-ascii log/19/trace2057 --trace-config all --trace-time --negotiate http://127.0.0.1:49219/2057 > log/19/stdout2057 2> log/19/stderr2057 === End of file commands.log === Start of file http_server.log 16:59:26.133510 ====> Client connect 16:59:26.139663 accept_connection 3 returned 4 16:59:26.139725 accept_connection 3 returned 0 16:59:26.143086 Read 97 bytes 16:59:26.154577 Process 97 bytes request 16:59:26.173113 Got request: GET /verifiedserver HTTP/1.1 16:59:26.173263 Are-we-friendly question received 16:59:26.237855 Wrote request (97 bytes) input to log/19/server.input 16:59:26.238029 Identifying ourselves as friends 16:59:26.241446 Response sent (55 bytes) and written to log/19/server.response 16:59:26.306495 special request received, no persistency 16:59:26.306574 ====> Client disconnect 0 === End of file http_server.log === Start of file http_verify.log * Trying 127.0.0.1:49219... * Connected to 127.0.0.1 (127.0.0.1) port 49219 * using HTTP/1.x > GET /verifiedserver HTTP/1.1 > Host: 127.0.0.1:49219 > User-Agent: curl/8.15.0-DEV > Accept: */* > * Request completely sent off < HTTP/1.1 200 OK < Content-Length: 16 < { [16 bytes data] * Connection #0 to host 127.0.0.1 left intact === End of file http_verify.log === Start of file http_verify.out WE ROOLZ: 7416 === End of file http_verify.out === Start of file server.cmd Testnum 2057 === End of file server.cmd === Start of file server.response HTTP/1.1 200 OK Content-Length: 16 WE ROOLZ: 7416 === End of file server.response ```

This reverts commit f650be4. There is already a GSS cmake job, right below this one.

This reverts commit 862815f.

Used by test 2056 and 2057, in a way that's Linux- & autotools-specific. This patch builds it for all Unix, syncing cmake with autotools. Adapt the two tests to find the library in CMake builds as well. Tested OK on Linux. (CI does not test this. The corresponding jobs build in debug mode, while the `LD_PRELOAD` feature is locked to non-debug.) On macOS it didn't load without building everything for aarch64e arch: "../bld/tests/libtest/libstubgss.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'arm64e'))" With that fixed it still did not load correctly and/or the tests did not pass. So, for macOS these tests remain disabled. Also: - GHA/macos: build for aarch64e. (recognized by Apple clang as of this patch. llvm and gcc fall back to aarch64.) Follow-up to 56d949d curl#1687 Closes curl#17653

CMake builds by default don't include a triplet in the `curl -V` output, but a CMake-specific OS string, which is usually capitalized or stylized, e.g. "Linux", or "FreeBSD". Make the regexp expression case-insensitive to handle this. Follow-up to 171b623 #17653 Closes #17756

Replace the `libstubgss.so`-based overload solution with one built into libcurl at compile-time. The previous, `LD_PRELOAD`-based, solution was non-portable, allowlisted for Linux, BSD and Solaris. It also required non-debug builds, which turned out to be an accidental condition: 7d342c7. It also required a curl tool built against a shared libcurl. Detecting this condition wasn't always accurate, e.g. with certain cmake configurations. The overload solution also didn't work on macOS, though it theoretically should have: - #17653 - #2394 Experiments on making the overload solution work in more envs: - #17759 That revealed that it also did not work on NetBSD, in CI. The replacement solution is overloading the necessary GSS-API functions for test 2056 and 2057 at compile time. It requires a debug-enabled curl build (due to its insecure nature). This makes these tests run on all platforms. Including most GSS jobs in CI, that are running tests. (the exception is old-linux, non-debug jobs, where it felt overkill to enable debug for this.) The refactored GSS stub code needs to overload less than before because it's free to use the official GSS API. (This didn't work with the overload solution on Alpine for example). It can also use libcurl functions, allowing to replace `snprintf()` with `msnprintf()`. OS/400 is also overloading GSS API functions. I haven't tested how this works after this PR. In theory it should, because this PR doesn't rely on preprocessor overrides. Note that for future GSS tests, it may be necessary to stub these GSS API functions: `gss_inquire_context()`, `gss_unwrap()`, `gss_wrap()`. They are on codepaths not (yet) touched by tests. Also: - stub-gss: check for token buffer overrun. - stub-gss: replace size macros with `sizeof()`. - GHA: enable debug for some jobs with GSS. - GHA/linux: ignore results for 2056 and 2057 in the valgrind job. They leak the same way as seen with 2077 and 2078. Ref: 7020ba7 #17462 Ref: 1467597 #14430 - GHA/linux: fix to ignore `gss_import_name()` leaks in valgrind builds. only. - lib/vauth/krb5_gssapi: reduce variable scope. - lib/vauth/spnego_gssapi: reduce variable scope. - tests/libtest: drop code and build logic dealing with `libstubgss`. - runtests: - drop `ld_preload` feature. - drop special handling of `LD_PRELOAD` env in tests. - drop logic dealing with shared curl tool detection. - drop `LD_PRELOAD` envs from tests. Follow-up to 56d949d #1687 Closes #17752

vszakats added the cmake label Jun 17, 2025

github-actions bot added build tests labels Jun 17, 2025

vszakats added 3 commits June 19, 2025 17:25

move filelist to Makefile.inc

41c1a04

add to cmake

8afedb4

try enable subgss for cmake and macos 1

8620257

vszakats force-pushed the cm-libtests-gssstub branch from 857414f to 8620257 Compare June 19, 2025 15:30

unity off for stubgss

84e1bbe

vszakats added 2 commits June 19, 2025 18:19

debug 1

74a2565

github-actions bot added the CI Continuous Integration label Jun 19, 2025

vszakats force-pushed the cm-libtests-gssstub branch from 9db8002 to 071f3bd Compare June 19, 2025 16:40

blind try 2

d505f55

vszakats force-pushed the cm-libtests-gssstub branch from 071f3bd to d505f55 Compare June 19, 2025 16:56

vszakats added 8 commits June 19, 2025 19:14

gha file revert

1ef7116

try linux gss job

f650be4

does not seem to be loading on macOS, either way the test fails 2

4e42d21

Revert "try linux gss job"

6e0734d

This reverts commit f650be4. There is already a GSS cmake job, right below this one.

try gss tests again, requires non-debug builds

862815f

runtests.pl drop debug line

25b8425

Revert "try gss tests again, requires non-debug builds"

c5f4d9a

This reverts commit 862815f.

vszakats force-pushed the cm-libtests-gssstub branch from 05596b4 to c5f4d9a Compare June 19, 2025 18:22

drop more debug lines

2d2b69a

vszakats closed this in 171b623 Jun 19, 2025

vszakats deleted the cm-libtests-gssstub branch June 19, 2025 18:42

vszakats mentioned this pull request Jun 26, 2025

runtests: fix LD_PRELOAD detection for cmake-built curl binaries #17756

Closed

vszakats mentioned this pull request Jun 27, 2025

tests: move GSS-API dynamic stub into debug-mode libcurl #17752

Closed

6 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

cmake: build `stubgss` library for libtests to match autotools #17653

cmake: build `stubgss` library for libtests to match autotools #17653

Uh oh!

vszakats commented Jun 17, 2025 •

edited

Loading

Uh oh!

testclutch commented Jun 19, 2025

Uh oh!

Uh oh!

Uh oh!

cmake: build stubgss library for libtests to match autotools #17653

cmake: build stubgss library for libtests to match autotools #17653

Uh oh!

Conversation

vszakats commented Jun 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

testclutch commented Jun 19, 2025

Generated by Testclutch

Uh oh!

Uh oh!

cmake: build `stubgss` library for libtests to match autotools #17653

cmake: build `stubgss` library for libtests to match autotools #17653

vszakats commented Jun 17, 2025 •

edited

Loading