Interface testing is required at work, and the message editor is laborious and time consuming, so consider using the Web_submit_data function post method for message editing. In the message editor, there are three main problems, one of which takes two days to find out the problem, here to share with you.
Question 1: We support message header style separate messages and parallel messages in only one XML format. Therefore, after the message is edited in only one XML format, the LR execution succeeds, but the relevant log is not found in the system. When troubleshooting the problem, the message was written several times, and the configuration of the Web_submit_data function was changed , and the parameterized variables were checked several times .... Just this morning, early to the company, open LR, re-wrote two scripts to troubleshoot problems when I put "name= XML "," Value= ... " after" Name=xmldata ", execute the script, the LR console prompts successfully, the system foreground successfully displays the log . Then tried the "Name=xml1", "Name=xmlqaz", "name=xmlaaa" .... System Front desk can not appear log, now think: Only a script in XML format, name can only be equal to XMLData to succeed?
Issue 2: After editing the message header, report style separate script, after the implementation of LR, the system front desk does not show the log. Re-edit the script, click Insert-->new step-->add step-->submit data, Record mode, Encoding Type, Encode "at" sign as ASCII option does not tick, after configuration, execute the script again, the System foreground query interface to find the log. thinking: In html/http mode, if you do not select, the LR script takes precedence over HTML mode, does the script with multiple XML formats work more in HTML mode?
Issue 3: After the problem 2 is resolved, the log number that appears in the foreground query interface is inconsistent with the log number printed by the LR console. Finally, enter the parameterized parameter type interface, the Update value option is changed from each iteration to once, the script is executed again, and the system foreground log number matches the LR console print log number. Then, change the once back to the original each iteration, the system foreground log number and the LR console print log number consistent.
So far, the problem of writing the LR interface test script has been solved, and it took nearly 3 days to solve all the problems. What I want to share is helpful to the partners who are struggling on the road.
LoadRunner interface Script Web_submit_data problems encountered during writing and sharing