On 04/24/2018 07:57 PM, Marion et Christophe JAILLET wrote: > > Le 24/04/2018 à 11:24, Ruediger Pluem a écrit : >> >> On 04/21/2018 08:36 PM, Marion et Christophe JAILLET wrote: >>> man pages generated by java 8 without Jakarta Oro and without >>> -Xbootclasspath/p look fine. >>> >>> But the one generated by java 9 and 11 are unusable! >>> So unless fixed, I'm -1 on removing -Xbootclasspath/p because Java 9 and up >>> break or man page generation. >> I am confused now. For Java 8 I understand that it doesn't matter if we use >> -Xbootclasspath/p. >> Java 9 doesn't work at all with -Xbootclasspath/p added and the result >> without is wrong. >> So my conclusion would be: Remove -Xbootclasspath/p and investigate why Java >> 9 is failing (might be a bug >> or another option that needs to be set for Java 9). >> Do I miss something? >> Thanks for testing, btw. >> >> Regards >> >> Rüdiger >> > My point is to leave -Xbootclasspath/p as long as we can't have a working > java 9 version. > This would prevent someone from generating some invalid man files and commit > them (or even release them) if no-one spots > it. At first, I didn't notice the issue with the man page, and apparently > Eric neither. > I see it as a guardrail. > > Of course, we should investigate, have Java 9 work, and "allow others to use > it", but only WHEN we know it works. > > My only concern is to prevent unaware people from silently generating invalid > files. > Testing the Java version and stopping if >= 9 is an option. Leaving > -Xbootclasspath/p does the same, but is is "out of > the box" :)
I haven't seen it from this perspective. Makes perfect sense. Thanks for explaining. Regards Rüdiger --------------------------------------------------------------------- To unsubscribe, e-mail: docs-unsubscr...@httpd.apache.org For additional commands, e-mail: docs-h...@httpd.apache.org
