R4R: Implement generate-only option for commands that create txs

[alessio]

The new CLI --generate-only flag builds an unsigned transaction and writes it to STDOUT. Likewise, REST clients can now append generate_only=true to a request's query arguments list and expect a JSON response carrying the unsigned transaction.

Closes: #966

• Linked to github-issue with discussion and accepted design OR link to spec that describes this work.
• Wrote tests
• Updated relevant documentation (docs/)
• Added entries in PENDING.md with issue #
• rereviewed Files changed in the github PR explorer

---

For Admin Use:

• Added appropriate labels to PR (ex. wip, ready-for-review, docs)
• Reviewers Assigned
• Squashed all commits, uses message "Merge pull request #XYZ: [title]" (coding standards)

[codecov]

Codecov Report

Merging #2165 into develop will decrease coverage by 0.27%.
The diff coverage is 5.26%.

@@             Coverage Diff             @@
##           develop    #2165      +/-   ##
===========================================
- Coverage    64.23%   63.95%   -0.28%     
===========================================
  Files          140      140              
  Lines         8575     8612      +37     
===========================================
  Hits          5508     5508              
- Misses        2690     2727      +37     
  Partials       377      377

[cwgoes]

I think this is a bit redundant with "offline" (pubkey-only) keys in gaiacli - did you look at that code?

This is definitely something we want regardless - maybe offline keys should automatically trigger --generate-only?

[cwgoes]

A bit of repeated code, also see comment re: offline keys.

[fedekunze]

utACK. Just add the test that's missing to CLI to test the msg.Signature length

[alexanderbez]

tested ACK. Left a minor note on the pending log.

Also, during testing, I found that I can generate txs without an amount. Is this desired/expected behavior? I would assume this logic should be identical to actually generating a real tx (except without signing and sending)...in other words, it should go through the same validation checks. No?

[alexanderbez]

tested ACK. Thanks!

[jackzampolin]

I don't know if we need validation on transaction generation. I think users should be able to generate w/o all required data and then fill in manually if desired. cc @alexanderbez Anyone else have thoughts here?

[alexanderbez]

@jackzampolin it provides more consistent behavior, but I poised it as a question more or less and not something that is necessary. I could be swayed the other way.

In other words, when I see --generate-only, my immediate thought is, "This does the exact same thing as without generate-only except signing and broadcasting". Hence my initial question 😄

[alessio]

Seems like we're all set and ready then

[alessio]

PR on sign command too is almost ready

[alessio]

@cwgoes are you happy with the changes?

[cwgoes]

utACK, thanks @alessio