Test fix - expect (null) instead of $(XX), and $(XX) instead of $(N) in some other places

[GBuella]

Hopefully helps with #77

---

This change is 

[codecov-io]

Codecov Report

Merging #79 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master      #79   +/-   ##
=======================================
  Coverage   91.95%   91.95%           
=======================================
  Files          11       11           
  Lines         957      957           
=======================================
  Hits          880      880           
  Misses         77       77

Flag	Coverage Δ
#pmemfile_tests	86.35% <ø> (ø)	⬆️
#regular_tests	90.8% <ø> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8405e65...3d19c90. Read the comment docs.

[krzycz]

---

Reviewed 2 of 2 files at r1.
Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[GBuella]

@elboulangero , could you try if this fixes the tests in #77 on your debians?

[elboulangero]

Hi @GBuella, thanks for pinging me.

Here's a build in Debian Jessie, test 25 and 28 still fail.

      Start 25: fork_logging
25/40 Test #25: fork_logging ...........................***Failed    0.17 sec
Executing: LD_PRELOAD=/build/syscall-intercept-0.1/obj-x86_64-linux-gnu/libsyscall_intercept.so.0.1.0
		INTERCEPT_ALL_OBJS=1
		/build/syscall-intercept-0.1/obj-x86_64-linux-gnu/test/fork_logging /build/syscall-intercept-0.1/test/fork_logging.c .log.logging .log.2.logging

...

      Start 28: hook_clone
28/40 Test #28: hook_clone .............................***Failed    0.17 sec
Executing: LD_PRELOAD=/build/syscall-intercept-0.1/obj-x86_64-linux-gnu/test/libhook_test_clone_preload.so
		INTERCEPT_ALL_OBJS=1
		/build/syscall-intercept-0.1/obj-x86_64-linux-gnu/test/fork_logging /build/syscall-intercept-0.1/test/fork_logging.c .log.hook_clone .log.2.hook_clone

...

Total Test time (real) =   2.05 sec

The following tests FAILED:
	 25 - fork_logging (Failed)
	 28 - hook_clone (Failed)

I attach the log files, tell me if you need more info.

log.2.hook_clone.txt
log.2.logging.txt
log.hook.txt
log.hook_clone.txt
log.logging.txt

[elboulangero]

Debian Stretch still succeeds

[elboulangero]

Debian Buster fails, test 25 and 28 fail exactly as on Jessie (same diff in the logs), and additionally the test 40 - syscall_format_logging fails. For some reason I didn't find a .log.2 file for this test, so I can only match syscall_format.log.match against .log.syscall_format_logging, which is not super helpful. Should I run some magic command to generate this .log.2.* file?

[GBuella]

@elboulangero I updated the patch, could you check it again?
I tried matching it with the logs you provided, there should be some improvement now.
I realized the mmap syscall's loggig format was changed, but some log lines in match files were never updated, since we didn't try it on such debian instances.
Thanks anyways!

[elboulangero]

@GBuella Thanks for the update! Now the tests succeed in Debian Jessie!

I still have a failure for Debian Buster, the 3 tests. From what I've seen

• Test 28 - hook_clone fails because of the lines wait4(4294967295, 0x0, 0x0, 0x0) = ?. Notice that the number 4294967295 is in fact -1. For Debian Jessie, this line is indeed wait4(-1, ...).
• Test 25 - fork_logging seems to have the same issue.
• Test 40 - syscall_format_logging I don't know what's going on here.

Files for Buster attached.

---

log.2.hook_clone.txt
log.2.logging.txt
log.hook_clone.txt
log.logging.txt
log.syscall_format_logging.txt

[marcinslusarz]

I think we should merge this as is and work on other issues in a new bug report.