Hello guys, as again there is new series of lo it means more tests were added to the list of those that are run during the make phase (where normal software do only build).
For developers it helps to identify issues quickly because your build crashes if you try to compile your changes but for users that try to build the lo just for themselves the crashes can get really annoying. Even that is not much of a problem as distributions usually provide binaries that were once build so they do not run the tests for each various hw/cflags/march/etc etc, sadly I am working on Gentoo, which means I do get bazilion of bugs for failing tests that are not reproducable on most machines. Small list of bugs from "experimental" branch after two days (meaning not much people using it and those errors will just grow as it moves to stable): https://bugs.gentoo.org/show_bug.cgi?id=394659 https://bugs.gentoo.org/show_bug.cgi?id=394615 https://bugs.gentoo.org/show_bug.cgi?id=394767 So unless you can look at one of those traces and tell me that the build should be stopped because "it is obvious error in the code (not in the test, the app code) that needs to be fixed" I would want you guys to rethink running the tests for release build during the make phase and just allow me to run them in designated test phase where it is allowed to fail and wont stop user from installing the office suite. (Yes those errors are logged, yes I will provide them, and yes user will be able to use his office suite and not wait on some test fixes) This is quite critical issue for packaging in Gentoo and I would love to fix it some other way than implementing cppunit hack that would return true on each test so users will be able to install their package... Cheers Tom _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice