Fix package path computation in ClasspathScanner #2531
Conversation
Prior to this commit no trailing `/` character was appended to the computed package path. Now, except for the default package `""`, a `/` is appended to package path. This leads to corrected and documented behavior even if two modules start with the same name elements. Fixes #2500
Codecov Report
@@ Coverage Diff @@
## main #2531 +/- ##
============================================
+ Coverage 90.56% 90.57% +0.01%
- Complexity 4683 4684 +1
============================================
Files 407 407
Lines 11568 11571 +3
Branches 923 923
============================================
+ Hits 10476 10480 +4
Misses 814 814
+ Partials 278 277 -1
Continue to review full report at Codecov.
|
| } | ||
| { | ||
| var classes = classpathScanner.scanForClassesInPackage("com.greetings.test", allClasses); | ||
| var classNames = classes.stream().map(Class::getName).collect(Collectors.toList()); |
There was a problem hiding this comment.
nit: I'd use a static import for toList().
There was a problem hiding this comment.
Let's wait until we switched to Java 16 -- and replace all those with Stream::toList
There was a problem hiding this comment.
All? All 44 usages found in test files.
Good questions. I'll replace this JAR file and the related Btw. do we know how to reproduce/modify |
sormuras
force-pushed
the
issues/2500-scan-for-classes-in-modules
branch
2 times, most recently
from
40b8377
to
a7ec3a4
Compare
Prior to this commit no trailing `/` character was appended to the computed package path. Now, except for the default package `""`, a `/` is appended to package path. This leads to corrected and documented behavior even if two modules start with the same name elements. Fixes #2500
sormuras
force-pushed
the
issues/2500-scan-for-classes-in-modules
branch
from
a7ec3a4
to
6cf0011
Compare
I don't know. Sounds like it may be problematic. 🤷♂️ |
Thanks! 👍 |
Fix `getRootUrisForPackage()` in class `ClasspathScanner` by looking for two "wordings" of a package name. For example, package `foo.bar` is expanded to `foo/bar` and `foo/bar/`. The latter allows find package `foo.bar` in a module called `foo.bar`, and not `foo`. This commit also ensures, that there's no regression (compared to #2531) by launching test runs using the standalone artifact with `--select-package` and `--scan-class-path`. Fixes #2500 Fixes #2600 Fixes #2612
Fix `getRootUrisForPackage()` in class `ClasspathScanner` by looking for two "wordings" of a package name. For example, package `foo.bar` is expanded to `foo/bar` and `foo/bar/`. The latter allows find package `foo.bar` in a module called `foo.bar`, and not `foo`. This commit also ensures, that there's no regression (compared to #2531) by launching test runs using the standalone artifact with `--select-package` and `--scan-class-path`. Fixes #2500 Fixes #2600 Fixes #2612
Fix `getRootUrisForPackage()` in class `ClasspathScanner` by looking for two "wordings" of a package name. For example, package `foo.bar` is expanded to `foo/bar` and `foo/bar/`. The latter allows find package `foo.bar` in a module called `foo.bar`, and not `foo`. This commit also ensures, that there's no regression (compared to junit-team#2531) by launching test runs using the standalone artifact with `--select-package` and `--scan-class-path`. Fixes junit-team#2500 Fixes junit-team#2600 Fixes junit-team#2612
Overview
Prior to this commit no trailing
/character was appended to the computed package path.Now, except for the default package
"", a/is appended to package path. This leads to corrected and documented behavior even if two modules start with the same name elements.Fixes #2500
I hereby agree to the terms of the JUnit Contributor License Agreement.
Definition of Done
@APIannotations