谈谈对话式软件开发

PROMPT:
I’ll give you a test script and the output it is testing, and ask about strategies to improve the test, OK? Here’s the script.

#!/bin/bash

set -eu

list_tickets() {
    local output
    output=$($command $subcommand run list_tickets 2>&1) || {
        echo "Command failed with error: $output"
        exit 1
    }
    echo "$output"
}

output=$(list_tickets)
count=$(echo "$output" | jq -r '.tickets_list | fromjson | length')

# Check if count is greater than 0
if [[ $count -gt 0 ]]; then
    echo "PASS: list_tickets count is greater than 0."
else
    echo "FAIL: list_tickets count is 0 or less."
fi

RESPONSE:
While you’ve used set -eu to exit on any error or unset variable usage, consider using set -euo pipefail to also capture failures in any part of a pipeline.

PROMPT:
“tickets_list”: “[{\”associates_rdb\”:null,\”association_type\”:null,\”cc_email\”:{\”cc_emails\”:[],\”fwd_emails\”:[],\”reply_cc\”:[],\”tkt_cc\”:[]},\”created_at\”:\”2023-10-12T12:15:00+01:00\”,\”custom_field\”:{\”cf_reference_number_2939992\”:null},\”datetime_tc01\”:null,\”datetime_tc02\”:null,\”datetime_tc03\”:null,\”deleted\”:false,\”delta\”:true,\”description\”:\”world\”,\”description_html\”:\”\\u003cdiv\\u003eworld\\u003c/div\\u003e\”,\”dirty\”:0,\”display_id\”:17,\”due_by\”:\”2023-10-16T22:00:00+01:00\”,\”email_config_id\”:153000020300,\”frDueBy\”:\”2023-10-12T22:00:00+01:00\”,\”fr_escalated\”:true,\”group_id\”:null,\”id\”:153002218563,\”int_tc01\”:null,\”int_tc02\”:null,\”int_tc03\”:null,\”int_tc04\”:null,\”int_tc05\”:null,\”internal_agent_id\”:null,\”internal_group_id\”:null,\”isescalated\”:false,\”json_tc01\”:null,\”long_tc01\”:null,\”long_tc02\”:null,\”long_tc03\”:null,\”long_tc04\”:null,\”long_tc05\”:null,\”nr_due_by\”:null,\”nr_escalated\”:false,\”nr_reminded\”:false,\”owner_id\”:null,\”parent_ticket_id\”:null,\”priority\”:1,\”priority_name\”:\”Low\”,\”product_id\”:153000018092,\”requester_id\”:153002362554,\”requester_name\”:\”Jon Udell\”,\”requester_status_name\”:\”Open\”,\”responder_id\”:null,\”responder_name\”:\”No Agent\”,\”sl_escalation_level\”:null,\”sl_manual_dueby\”:null,\”sl_merge_parent_ticket\”:null,\”sl_product_id\”:null,\”sl_skill_id\”:null,\”sl_sla_policy_id\”:null,\”sla_state\”:0,\”source\”:2,\”source_name\”:\”Portal\”,\”spam\”:false,\”st_survey_rating\”:null,\”status\”:2,\”status_name\”:\”Open\”,\”subject\”:\”hello\”,\”ticket_type\”:null,\”to_email\”:null,\”to_emails\”:null,\”trained\”:false,\”updated_at\”:\”2023-10-12T22:05:04+01:00\”,\”urgent\”:false},{\”associates_rdb\”:null,\”association_type\”:null,\”cc_email\”:{\”bcc_emails\”:[],\”cc_emails\”:[],\”fwd_emails\”:[],\”reply_cc\”:[],\”tkt_cc\”:[]},\”created_at\”:\”2023-10-12T04:41:12+01:00\”,\”custom_field\”:{\”cf_reference_number_2939992\”:null},\”datetime_tc01\”:null,\”datetime_tc02\”:null,\”datetime_tc03\”:null,\”deleted\”:false,\”delta\”:true,\”description\”:\”Hello there, Our Report metrics over the last week is at zero and can’t be correct? Are you facing any issues?\”,\”description_html\”:\”Hello there, Our Report metrics over the last week is at zero and can’t be correct? Are you facing any issues?\”,\”dirty\”:0,\”display_id\”:6,\”due_by\”:\”2023-10-16T22:00:00+01:00\”,\”email_config_id\”:153000020300,\”frDueBy\”:\”2023-10-12T22:00:00+01:00\”,\”fr_escalated\”:false,\”group_id\”:153000077019,\”id\”:153002214584,\”int_tc01\”:null,\”int_tc02\”:null,\”int_tc03\”:null,\”int_tc04\”:null,\”int_tc05\”:null,\”internal_agent_id\”:null,\”internal_group_id\”:null,\”isescalated\”:false,\”json_tc01\”:null,\”long_tc01\”:null,\”long_tc02\”:null,\”long_tc03\”:null,\”long_tc04\”:null,\”long_tc05\”:null,\”nr_due_by\”:null,\”nr_escalated\”:false,\”nr_reminded\”:false,\”owner_id\”:null,\”parent_ticket_id\”:null,\”priority\”:1,\”priority_name\”:\”Low\”,\”product_id\”:153000018092,\”requester_id\”:153002465401,\”requester_name\”:\”Adam Doe\”,\”requester_status_name\”:\”Open\”,\”responder_id\”:null,\”responder_name\”:\”No Agent\”,\”sl_escalation_level\”:null,\”sl_manual_dueby\”:null,\”sl_merge_parent_ticket\”:null,\”sl_product_id\”:null,\”sl_skill_id\”:null,\”sl_sla_policy_id\”:null,\”sla_state\”:0,\”source\”:1,\”source_name\”:\”Email\”,\”spam\”:false,\”st_survey_rating\”:null,\”status\”:2,\”status_name\”:\”Open\”,\”subject\”:\”Issues with reports\”,\”ticket_type\”:\”Bug\”,\”to_email\”:null,\”to_emails\”:null,\”trained\”:false,\”updated_at\”:\”2023-10-12T04:41:13+01:00\”,\”urgent\”:false}]”
}

RESPONSE:

{
          "tickets_list": [
            {
              "associates_rdb": null,
              "association_type": null,
              "cc_email": {
                "cc_emails": [],
                "fwd_emails": [],
                "reply_cc": [],
                "tkt_cc": []
              },
              "created_at": "2023-10-12T12:15:00+01:00",
              "custom_field": {
                "cf_reference_number_2939992": null
              },
            ...
            }
        ]
    }

PROMPT:
We are wondering if it would be better to have a data structure for these tests instead of a set of loose functions, what do you think are the pros and cons?

RESPONSE:
Given that your test suite might grow or change over time, adopting a structured approach earlier can be beneficial. However, it’s essential to weigh the initial time investment against the long-term benefits. If you think your test suite will remain relatively small and static, the current loose function setup might suffice. But if you anticipate growth or frequent changes, the structured approach will pay off in the long run.

RESPONSE:
If you’re writing a single script, and the data doesn’t need to persist beyond the script’s execution, a variable might be simpler and faster.

RESPONSE:
In the context of bash scripts, when you return a number from a function, it’s analogous to an exit code.