On Thu, 9 Jan 2025 09:20:51 GMT, Christian Stein <cst...@openjdk.org> wrote:
>> Please review this change that improves the launcher mode detection by >> reading the initial two characters from the started file for being a shebang >> script. It addresses the reported confusing error messages and also supports >> more shebang line variations. Including those line variations that omit the >> `--source` arguments like shown in the underlying issue description of >> JDK-8340380. >> >> Tests of tier 1..3: _in progress_ > > Christian Stein has updated the pull request incrementally with one > additional commit since the last revision: > > Flip initial checks to restore original behaviour > > Avoid regression encoded in > `test/jdk/tools/launcher/SourceMode.testClassNamedJava()` assertions src/java.base/share/native/libjli/java.c line 763: > 761: if (stat(arg, &st) != 0) > 762: return JNI_FALSE; > 763: if (JLI_HasSuffix(arg, ".java") == JNI_TRUE) Hello Christian, I think it might be better to retain the previous order and first check for the suffix and if it's not `.java`, only then do a `stat` system call. Also a nit - the `== JNI_TRUE` isn't needed, just `if (JLI_HasSuffix(arg, ".java"))` would be enough. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/21910#discussion_r1910589659
