A new automated solution to improve global testing efficiency based on crawler technology

With the continuous expansion of the computer industry in the global market, the globalization of software products becomes more and more important. However, in the software globalization process, there are many heavy work is to find and verify the Product interface translation of the string, so as to ensure that the SOFTWARE product interface friendly and consistent with the custom of multinational language and culture. This is a big challenge for software globalization. We now provide a crawler-based automated solution to quickly find and locate translated strings in the Web application interface to reduce manual work and improve the efficiency of the global testing effort. The implementation steps for this automated solution are mainly as follows:

A. Use web crawler technology to explore URLs on Web applications and coexist as a list of URLs.

B. Use the browser proxy to load each URL corresponding to the content of the Web page, using the algorithm to find the translated string in the page.

C. Once you find the corresponding string, highlight the area of the translated string in the page, and screen the file to provide the local tester for the corresponding language to verify.

I. Globalization testing and its testing difficulties

With the growing of the computer industry in the international global market, the software products need not only strong function, but also adapt to the language and culture of the global customers in order to stand out in the competitive market, so the globalization test is becoming more and more important in the software development test. The globalization test mainly includes GVT (globalization verification test) and TVT (translation verification test), the former mainly for the use of software products across a number of cultural regions, And do not have to take account of linguistic and regional differences, requires that its functionality and design not only support a single language and region, but also ensure that a set of Unicode-enabled languages and related data in a particular area are properly entered, displayed, and exported, and that the latter is aimed at the reasonableness of the translation of strings displayed in the product, Whether it can be displayed correctly. The solution we designed is primarily to improve the efficiency of the latter's TVT testing.

At present, the difficulties encountered in the TVT testing process are mainly two points:

Locate the position of the string being measured in the product: Product functionality is increasingly complex, if the developer does not provide enough product information, testers need to learn to understand the product to find the divination of word string in the product of the page, which not only brings challenges to testers, but also spend more time cost, the test coverage is not optimistic.

TVT test work is tedious: if the test case is complex, or sometimes in order to save costs, we TSE (TVT Support Engineer) through the test case steps, directly help intercept the corresponding translation language of the product interface, and then sent to the local translation test People to make it easier for them to test more easily. But TVT test language may have 20, or even more, each test case corresponding to each language needs screenshots, in the case of more test cases, if the use of manual screenshots, will consume more manpower and time. It can be seen that this part of the work more repetitive and boring, but also may affect the enthusiasm of everyone's work.

Second, the automation of the solution related technology introduction

The main function of this automation solution is to automatically find the interface of the string in the product through the string information, automatically intercept the interface and save, the final local translation tester only need to test the screenshot directly, to avoid the difficulty and the manpower cost caused by the complex and repetitive test. The key technology of this solution is to find all the pages under the website product through the web crawler technology, and select the pages containing the divination of word string, so this scheme is only for web-based software products.

Web crawler technology is also called Web spider, it is a set of rules to automatically extract the Web page of the technology, each to a Web page to read the content of the Web page, to find other links in the Web page, and then through these link addresses to find the next page, so that the cycle continues until the site all the pages are crawling up to date, So it can be used to explore the structure of WEB application. Web crawler always start from a certain starting point, this is called the entry address, that is, just provide a WEB application portal address, with the crawler technology can search all the pages under this application.

IBM has a web crawler technology software PRODUCT called AppScan Standard, this solution can directly apply the technology in this product. AppScan Standard is the industry's excellent Web application security Testing tool, is the Web site and other Web applications under security attacks, to check whether the site has security vulnerabilities, so it has to analyze the site structure of the crawler technology, and we designed the solution need is this key technology , the request to use the Web page is sent using the HTTP protocol, send and return content is a unified HTML language, then the HTML language analysis, find the parameters and links inside, record and continue to send, and finally find the site under the numerous pages and directories, this function in the AppScan called "Explore "is to analyze and understand the structure of the unknown site and record it."

Design process and implementation of automation solution

The TVT manual testing process for web-based products is as follows:

TVT begins with a document called PII (program intergrated information) by the development team that contains strings that are visible to some users in the software product, similar to Figure 1, which is made up of Key=value, which is the type of software product in which users can See the string. Through the PII file, we can determine the development group needs to do a translation test string, the actual test involved in the file may be dozens of to hundreds of, the number of strings required to test the number of words also have thousands of to tens of thousands of. If the string inside the red box is the one that needs to be tested.

Figure 1. PII Documentation