"(" and ")" are treated literally when inside [], no need to escape them.
LoaderUtil::parseFeed():
htmlToPlainText(): make the regex pattern non-greedy otherwise it may
match the whole subject string.
make && ctest
| No Linters Available |
| No Unit Test Coverage |
| Buildable 20658 | |
| Build 20676: arc lint + arc unit |
Some of these regex patterns are too complex, (some parts I don't even know what they do e.g. "(?)" in loaderutil.cpp, line 64); so I made sure that they are all valid by calling QRegularExpression::isValid() on them.
A file-static function is not exported outside that cpp file, so that's quite expected.
| src/loaderutil.cpp | ||
|---|---|---|
| 69 ↗ | (On Diff #72795) | Indeed that (?) looks invalid. kregexpeditor says so, at least. @mlaurent in commit dda15b0aaddec01 you moved code around, but you also replaced (?:HREF) with (?), while the rest of this regexp didn't change. Any reason for that? I think that either HREF should be added back, or (?) should be removed. |
| src/tools.cpp | ||
| 159 ↗ | (On Diff #72795) | Any reason for the added ? (for non-greedy-match)? I can't think of a case where it would make a difference, here. |
| src/loaderutil.cpp | ||
|---|---|---|
| 69 ↗ | (On Diff #72795) | The pattern is valid according to QRegularExpression::isValid(), and it seems to match sort of on word boundaries like \b and \B... weird :) |
| src/tools.cpp | ||
| 159 ↗ | (On Diff #72795) | Greedy matching would match: <h1 style="display: block" some title in the the badly formatted header</h1> so, it'll remove the whole string from < to the last >. |
| src/loaderutil.cpp | ||
|---|---|---|
| 69 ↗ | (On Diff #72795) | Perhaps I made an error on regexp. I extracted code for testing it (testing current bug) so if all is ok... |
| src/loaderutil.cpp | ||
|---|---|---|
| 69 ↗ | (On Diff #72795) | Yes, all unit tests still pass. |