script-to-address command doesn't load from STDIN

[libbitcoin]

script-to-address command doesn't load from STDIN

[evoskuil]

Try this:

$ cat msig.script | bx script-to-address

[evoskuil]

I need to get around to porting the multisig and stealth examples to the BX wiki.

[evoskuil]

Oh, I see your msig.script file is encoded. STDIN from that file should be a single base16 token. script-decode converts it to a set of tokens, which script-to-address accepts. This works as expected:

c:\>bx script-encode 2 [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] 3 checkmultisig > msig.script
5221024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f21024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f21024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f53ae

c:\>type msig.script | bx script-decode
2 [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] [ 024745a36740327223a4dd3b9f42da6f03a5418d6ef18f1371e097f7dedc49b25f ] 3 checkmultisig

c:\>type msig.script | bx script-decode | bx script-to-address
37xnxAAtVbceQXeudE7edjmToaLPuh2bz3

[evoskuil]

The example is actually quite a bit more straightforward in BX, since there is no reason whatsoever to encode the script as base16. BX preserves the option to encode and decode script for the case where the script is being extracted from a tx or some other encoded source. But when fabricating scripts base16 is just a nuisance.

Additionally, BX doesn't overload parameter types. SX, very inconsistently, would inspect the data in certain parameters to determine the type. This complicated implementation and documentation of the commands. BX strongly types each parameter so there are no ambiguities, and I haven't found that this hinders usability.

[periodic1236]

I definitely prefer the design of BX and I'm glad you made the effort to standardize input/output processing.

As for your example on Windows, it actually illustrates the bug. Take another look at the two commands in my post. 37xnxAAtVbceQXeudE7edjmToaLPuh2bz3 is not the right address - it's the same thing that bx script-to-address 0 returns. Sorry for not making that clear.

[evoskuil]

OK, I see it. Now that I have a repro I'll take care of it.

[evoskuil]

The stdin=true attribute was missing from the metadata declaration for script-to-address (added below):

  <command symbol="script-to-address" formerly="scripthash" category="TRANSACTION" description="Create a BIP16 pay-to-script-hash address from an encoded script.">
    <argument name="TOKEN" stdin="true" limit="-1" type="string" description="The script. If not specified the script is read from STDIN."/>
  </command>

So the script was always empty if provided through STDIN.

[evoskuil]

Fixed by #217

[evoskuil]

@periodic1236 Thanks for writing this up!

[evoskuil]

The unit test suite is comprehensive inside of each command, but the coverage is quite light on the command line handing. This is mitigated in part by the fact that much of that code is generated, but if the metadata is incorrect ... :<.

[periodic1236]

Thanks! Nice to see it was a simple fix.