Fixed oca generation bug creating invalid JSON

[wadeking98]

Previously the JSON files generated by the gen_ocabundlesjson.sh script would be invalid due to a trailing comma. This caused BC Wallet to fail to properly handle the JSON. I've added functionality to not add a comma on the last entries in the json file.

[jleach]

I've added a few comments. I think because its fairly long and a little complicated to read breaking it into a few functions might make it more readable or maintainable:

https://linuxize.com/post/bash-functions/

[jleach]

Use printf instead of echo for better control over formatting and escaping. For example, instead of echo -e "{" >${OCAIDSJSON}, you can use printf '{\n' > ${OCAIDSJSON}.

[jleach]

Use jq instead of manually constructing JSON strings. jq is a command-line tool for processing JSON data, and it can be used to construct complex JSON objects and arrays. For example, instead of manually constructing the JSON object for each bundle, you can use jq to construct the object.

 jq -n --arg id "$id" --arg bundle_path "$BUNDLE_PATH" '{($id): {path: $bundle_path}}' >> ${OCAIDSJSON}

[swcurran]

That’s a lot to eliminate the last line “,” :-). Painful…but good stuff.

And I thought most JSON processors handle an extra comma at the end of the list.

[jleach]

Use readarray instead of declare -a to read an array from a command output. For example, instead of declare -a schema_files; schema_files=(OCABundles/schema/*), you can use readarray -t schema_files < <(find OCABundles/schema -type f).

schema_files=()
readarray -t schema_files < <(find OCABundles/schema -type f)

[cvarjao]

@amanji , @swcurran , can we get this reviewed/merged?

[swcurran]

Sorry about that. It keeps getting bumped from my list. I’ll try for tomorrow EOD.

[swcurran]

Looks good — just the two typos and links to be changed, and we can merge this.

[swcurran]

Sorry for the long delay — had it on my list, but just not getting to it.

[swcurran]

@jleach — are your suggestions required for merging? Can’t merge until you say OK.

FYI - a suggestion on a probably easier way to make this change, and a reasonable approach in any Bash scripts. The entire JSON output could have been passed through another sed script with a simple substitution regex to remove the trailing comma on the second last line. Bash pipelines are easy to do.