types: add RE_ prefix to ARRAY_SIZE()

[alfredh]

1. avoid potential naming conflicts
2. add prefix to make the code more readable

[alfredh]

the CI failed with a new error:

test: test_sipsess_100rel_answer_not_allowed: test failed (Invalid argument [22])

this is not related to this change. Is it related to the PRACK commit yesterday ?

@cspiel1 could you please take a look ?

[cspiel1]

Could no reproduce this so far.

Where did you see the line?

test: test_sipsess_100rel_answer_not_allowed: test failed (Invalid argument [22])

On my dev host in home office with Ubuntu 22.04 and openssl 3.0.2. I still get this with valgrind

test 2 -- test_aes_gcm
==1302671== Conditional jump or move depends on uninitialised value(s)
==1302671==    at 0x4FB52A4: ??? (in /usr/lib/x86_64-linux-gnu/libcrypto.so.3)
==1302671==    by 0x4FB5581: ??? (in /usr/lib/x86_64-linux-gnu/libcrypto.so.3)
==1302671==    by 0x4EB60F4: EVP_DecryptFinal_ex (in /usr/lib/x86_64-linux-gnu/libcrypto.so.3)
==1302671==    by 0x4908C0D: aes_authenticate (aes.c:240)
==1302671==    by 0x11F4A5: test_aes_gcm (aes.c:341)
==1302671==    by 0x16661B: test_unit (test.c:510)
==1302671==    by 0x166EBC: test_reg (test.c:728)
==1302671==    by 0x1401F3: main (main.c:233)
==1302671== 

On my dev host at Commend office (also Ubuntu 22.04 and openssl 3.0.2) didn't had this. This is a different problem but hinders me to investigate the 100rel test. We will open a PR that adds a valgrind suppression file.

edit: As long as the conditional jump warning doesn't occur on github we don't need a suppression file in the repository.

[maximilianfridrich]

This is definitely due to my changes in baresip/retest#167. It seems to be a timing issue in retest, I will investigate it. Sorry for causing the tests to be unstable.

[maximilianfridrich]

This re PR could fix it: baresip/retest#177. Also, my comment on the retest issue baresip/retest#176 explains why I think this is happening.