Fix ts_get_now_internal to use transaction time

[pmwkaa]

Related Issue: #2167

Not really sure how to test it, since we not exported this function. If necessary I guess I could make a test c function to wrap it up and then check that the result of it does not change when executed inside an active transaction more then once.

[svenklemm]

i think the current behaviour is intentional and not a bug

[codecov]

Codecov Report

Merging #2235 into master will increase coverage by 0.33%.
The diff coverage is 93.52%.

@@            Coverage Diff             @@
##           master    #2235      +/-   ##
==========================================
+ Coverage   89.94%   90.28%   +0.33%     
==========================================
  Files         211      212       +1     
  Lines       34262    34471     +209     
==========================================
+ Hits        30818    31123     +305     
+ Misses       3444     3348      -96     

Impacted Files	Coverage Δ
src/chunk.h	100.00% <ø> (ø)
src/compat.h	100.00% <ø> (ø)
src/dimension.c	95.38% <ø> (-0.10%)	⬇️
src/utils.h	100.00% <ø> (ø)
tsl/src/compression/simple8b_rle.h	97.37% <ø> (ø)
tsl/src/fdw/fdw.c	88.54% <ø> (+1.80%)	⬆️
tsl/src/init.c	88.88% <ø> (ø)
tsl/src/license.c	81.81% <ø> (ø)
tsl/src/nodes/gapfill/gapfill.c	100.00% <ø> (ø)
tsl/src/remote/txn.c	87.63% <46.66%> (-1.17%)	⬇️
... and 56 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 4538fc6...c240bff. Read the comment docs.

[WireBaron]

This seems more correct than the current implementation, but given Sven's objection we should probably discuss this during a group meeting.

[erimatnor]

FWIW, this now-replacement is used in two places in the continuous aggregate code: first to set the modification time for invalidations in the insert trigger, then in the materializer to calculate the end of the range to materialize. IMO, transaction time makes sense in both of these cases.

[pmwkaa]

@svenklemm any thoughts, suggestions?

[svenklemm]

since caggs seems to be only user of this function i guess its no problem to change it to whatever makes most sense for current caggs