#findbugs

How to write a customized gradle task to not to ignore Findbugs violations but fail after the analysis is completed

I want to write such a gradle task (using the Findbugs plugin) which fails if any Findbugs violations are found but only after completing the analysis. If I do ignoreFailures=true the task wont fail at all and if i make it false the task fails as soon as the first issue is found. I want the task to perform a complete analysis and fail only after its done if any violations are found.

Answer

The argument and debates regarding writing quality code have mostly rested in the favor of writing clean code. Cleanliness fosters quality. Clean code is straighforward, direct, reads like a story and makes it very understandable.

As a rule that we have at

Problem Statement : I have application specific API, with poor documentation and not so good API design. Combination of these two asks for careful development with good knowledge of API internals, which may cause unpredictable behavior if misused. The learning curv for API is very steap and cost of experienced developers even for simple tasks is high. This is a common problems across numerous home grown applications and ill designed enterprise applications. Either the bug density or cost of development/maintenance tends to be very high. I see two solution that are

If the code written by new comer, gets continuously analyzed by an expert, the learning will be much faster and quality of code will improve.

All code written as of now with the API, needs to be reviewed thoroughly by an expert to seek known problems.

Though obvious solution is not practical if expert is a human being. Alternative is to have a static code analyzer which will have all possible "MIS"use cases for API. It will be very easy to feed complete code base and get direct list of bugs existing. Also IDE plugins are present which will point out wrong usage while developer is writing the code. This will remove all known misuses of API and newbie learning will be ever faster. But none of the code analyzers in market will know how this custom API is to be used. Here comes Static code analyzers like PMD or FindBugs. These tools provide a way to write your custom rules, which will be used along with standard rules for the base language or platform. I see development of such tools to reduce the cost of maintenance of legacy products and increase the quality of new code being developed on top of these. Resources:

how to write a FindBugs Detector

how to write a PMD Rule

PMD IDE Integration