Software regression

A software regression is a type of <a href="/facts/Software_bug/I1uSXySv">software bug</a> where a feature that has worked before stops working. This may happen after changes are applied to the software's <a href="/facts/Source_code/o9J6Jv23">source code</a>, including the addition of new <a href="/facts/Software_feature/9JRKJJxa">features</a> and bug fixes. They may also be introduced by changes to the environment in which the software is running, such as system upgrades, <a href="/facts/Patch_(computing)/sIHVBPls">system patching</a> or a change to <a href="/facts/Daylight_saving_time/bEWvB1Rt">daylight saving time</a>. A software performance regression is a situation where the software still functions correctly, but performs more slowly or uses more memory or resources than before. Various types of software regressions have been identified in practice, including the following:

<ul><li>Local – a change introduces a new bug in the changed module or component.</li>
<li>Remote – a change in one part of the software breaks functionality in another module or component.</li>
<li>Unmasked – a change unmasks an already existing bug that had no effect before the change.</li></ul>
Regressions are often caused by <a href="/facts/Hotfix/K10d8qco">encompassed bug fixes</a> included in <a href="/facts/Software_patch/sIHVBPls">software patches</a>. One approach to avoiding this kind of problem is <a href="/facts/Regression_testing/37GwvbAS">regression testing</a>. A properly designed <a href="/facts/Test_plan/8buKdVWj">test plan</a> aims at preventing this possibility before releasing any software. <a href="/facts/Automated_testing/UAVXbSTV">Automated testing</a> and well-written <a href="/facts/Test_case/TVWkj0AQ">test cases</a> can reduce the likelihood of a regression.

Software regression open-in-new

Software regression