Fix memory leak for compression with merge chunks

[cevian]

The RelationInitIndexAccessInfo call leaks cache memory and seems to be unnecessary.

[codecov]

Codecov Report

Merging #5063 (1a39a94) into main (13ac2fc) will increase coverage by 27.22%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##             main    #5063       +/-   ##
===========================================
+ Coverage   62.38%   89.60%   +27.22%     
===========================================
  Files         227      227               
  Lines       45501    51616     +6115     
===========================================
+ Hits        28385    46252    +17867     
+ Misses      17116     5364    -11752     

Impacted Files	Coverage Δ
tsl/src/compression/compression.c	96.32% <ø> (+0.51%)	⬆️
tsl/test/src/test_chunk_stats.c	92.59% <0.00%> (-7.41%)	⬇️
src/uuid.c	78.57% <0.00%> (-6.05%)	⬇️
src/histogram.c	83.83% <0.00%> (-4.80%)	⬇️
tsl/src/fdw/fdw_utils.c	80.76% <0.00%> (-4.65%)	⬇️
src/time_utils.c	93.72% <0.00%> (-3.78%)	⬇️
src/time_bucket.c	95.81% <0.00%> (-2.76%)	⬇️
src/hypertable_cache.c	93.25% <0.00%> (-2.64%)	⬇️
tsl/src/dist_backup.c	92.30% <0.00%> (-2.52%)	⬇️
src/loader/bgw_message_queue.c	86.36% <0.00%> (-2.46%)	⬇️
... and 203 more

---

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 13ac2fc...1a39a94. Read the comment docs.

[akuzm]

Yeah, not sure why we're calling it directly, it's already called by relation_open -> RelationIdGetRelation -> RelationBuildDesc.

[antekresic]

Yeah, this was my bad, was trying to fill it manually because of some other issue I was dealing with and it got left in when I fixed that issue.

[SachinSetiya]

I rebased on top of the latest main (So that we can see sanitizer test results)