Do not include sparse_array.o in libssl

[huiyuexu]

sparse_array.o is not needed in libssl at 3.0.x version.

fix #22113

Checklist

• documentation is added or updated
• tests are added or updated

[tom-cosgrove-arm]

LGTM

CI fuzz failure is unrelated, and I've raised issue #22114 to cover the broken 3.0 branch

[huiyuexu]

How I can rebuild the CIFuzz? @tom-cosgrove-arm

[t8m]

This is for 3.1 and 3.0 branches only as sparse arrays are used by QUIC.

[t8m]

It is possible to do something for the master branch as well - i.e., make the sparse array added only when QUIC is enabled. However that would be a different patch.

[openssl-machine]

24 hours has passed since 'approval: done' was set, but as this PR has been updated in that time the label 'approval: ready to merge' is not being automatically set. Please review the updates and set the label manually.

[paulidale]

Do we even need this for QUIC?

[levitte]

Why would we need this for QUIC? I just grepped for ossl_sa_ and OPENSSL_SA through the source, and I can't see it used anywhere except for the property code (in libcrypto) and a test program.

[levitte]

Ah, found the spot where it might look like libssl uses sparse arrays. However, it builds fine after this diff too:

diff --git a/ssl/event_queue.c b/ssl/event_queue.c
index 49890a36b5..20f4020db5 100644
--- a/ssl/event_queue.c
+++ b/ssl/event_queue.c
@@ -9,7 +9,6 @@
 
 #include <stdlib.h>
 #include "internal/event_queue.h"
-#include "crypto/sparse_array.h"
 #include "ssl_local.h"
 
 struct ossl_event_queue_st {

[paulidale]

I saw that, it's just the header file and isn't actually used.

[huiyuexu]

Could this submission be merged?

[paulidale]

Our automation didn't send a reminder here. It's meant to.

Possibly because this PR is against the 3.0 branch not master?

[t8m]

So, should this be merged to master as well?

[mattcaswell]

So, should this be merged to master as well?

I think it should be. My approval counts for master/3.1 and 3.0. It sounds like there should be a follow on PR to remove the spurious header file inclusion.

[mattcaswell]

It sounds like there should be a follow on PR to remove the spurious header file inclusion.

#22174

[t8m]

Merged to master, 3.1, and 3.0 branches. Thank you for your contribution.