Fix GH-10611: fpm_env_init_main leaks environ

[nielsdos]

Fixes GH-10611

[devnexen]

@nielsdos what do you get with a ASAN build ? do you think your changes trigger the FPM failures ?

[nielsdos]

I did use an ASAN build, no failures here... I probably did something wrong though, let me check...

[nielsdos]

It was my fault. I put the wrong "which" argument and apparently had a stale build so that the test failures didn't trigger.
I changed it to FPM_CLEANUP_PARENT_EXIT_MAIN which I think is correct now and the memleak also doesn't trigger.
Thanks for notifying me!

[bwoebi]

@nielsdos Planning on merging that? :-)

[nielsdos]

@nielsdos Planning on merging that? :-)

I didn't get an approval from @bukka yet, but after that I can merge :)

[bwoebi]

@nielsdos If I'm not mistaken that actually also needs to run in the child processes, as fpm will fork without exec?

I'm currently still seeing:

[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "Direct leak of 232 byte(s) in 1 object(s) allocated from:"
[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "    #0 0x7f834082a330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)"
[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "    #1 0x5586267d0df9 in fpm_env_init_main /usr/local/src/php/sapi/fpm/fpm/fpm_env.c:282"
[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "    #2 0x5586267bdf04 in fpm_init /usr/local/src/php/sapi/fpm/fpm/fpm.c:62"
[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "    #3 0x5586267df8ff in main /usr/local/src/php/sapi/fpm/fpm/fpm_main.c:1822"
[13-Feb-2024 17:35:00] WARNING: [pool www] child 37438 said into stderr: "    #4 0x7f833fb0509a in __libc_start_main ../csu/libc-start.c:308"

[nielsdos]

Possibly, you could try that and see if it solves it. I could also look at it next week or so.