sdk: correct static keypair bytes in test sample transaction

[t-nelson]

Problem

way back when, we incorrectly converted a pkcs8-formatted ed25519 test keypair to raw bytes while porting ed25519 support from ring to ed25519_dalek. i noticed and yolo'd a "fix" into #32926, but nfc if the raw private key bytes were actually a valid private key.

Summary of Changes

• use the correct keypair bytes from the original pkcs8
• verify that our "sample" transaction is actually signed correctly

NOTE: i went to the trouble of tracking this down to ensure #32926 wasn't likely to break any existing filesystem keypairs we may have generated in the past.

[codecov]

Codecov Report

Merging #32930 (0dcddd8) into master (9d83bb2) will decrease coverage by 0.1%.
The diff coverage is 100.0%.

@@            Coverage Diff            @@
##           master   #32930     +/-   ##
=========================================
- Coverage    82.0%    82.0%   -0.1%     
=========================================
  Files         784      784             
  Lines      212449   212450      +1     
=========================================
- Hits       174270   174238     -32     
- Misses      38179    38212     +33     

[brooksprumo]

Lgtm

[joncinque]

Hehe that was a fun little dive into history. Looks good!