New version adds mechanism for temporarily blocking (disabling) "Force Full Update" button when model reset is started. This is done with using new ProblemModel::featuresChanged() signal which emitted by standard problem reporter and cppcheck plugins at start/end of analysis.
Old version always draw "Force Full Update" button in enabled state which is confusing when update is already running.
Based on the D11081 revision.
Works as expected
| No Linters Available |
| No Unit Test Coverage |
LGTM in general, but please do see my remark.
| plugins/problemreporter/problemreportermodel.cpp | ||
|---|---|---|
| 111 | Kind of scary to have this regexp in this class here... What's the problem when .patch files are actually being updated? I guess we just get a runtime warning? If so, then I'd rather just remove this regexp here... Or find another way to detect these non-user documents more cleanly. | |
| plugins/problemreporter/problemreportermodel.cpp | ||
|---|---|---|
| 111 | The main problem with .patch files is that the BackgroundParser does not emit parseJobFinished signal for them. Therefore m_fullUpdateDocuments will be not-empty at the end of parsing :( Quick-and-dirty way to detect such filename without QRegularExpression may be somethig like: static const QString vcsDiffFilePrefix(QLatin1String("kdevelop_"));
static const QString vcsDiffFileSuffix(QLatin1String(".patch"));
auto fileName = url.fileName();
if (fileName.size() == (vcsDiffFilePrefix.size() + vcsDiffFileSuffix.size() + 6) &&
fileName.startsWith(vcsDiffFilePrefix) &&
fileName.endsWith(vcsDiffFilePrefix)) {
// our stuff ...
}Also I don't know when such .patch files are created - sometimes they are exists, sometimes - not. Do you know some some step-to-reproduce actions for force .patch files creation? This can help me to make some tests and find more beautiful solution. | |
| plugins/problemreporter/problemreportermodel.cpp | ||
|---|---|---|
| 111 | The regex itself isn't the problem. The problem is that we need to have this check/filter here altogether. m_fullUpdateDocuments should never contain these .patch files (or IOW the problem store should never contain it) to begin with. Do you agree? I don't have the full picture about how problem store gets filled, I'd need to look into that myself... | |