Fernando de Oliveira wrote: > This mail was written earlier. Lost again internet connection (after > several months, it has started again: twice in a day). > > I will send it as is. > > Bruce, have just seen [blfs-book] r11441. Liked it very much. > > Em 16-07-2013 13:10, Bruce Dubbs escreveu: > >> Thanks for the work on that. > > It's my pleasure. > >> I do think that we need to do some >> reorganization of the book for java. I think I will add a sub-chapter >> to Chapter 13 - Programming. The title will be 'Java' with sections >> 'About Java', 'OpenJDK/IcedTea', 'JUnit', and 'Apache-Ant' from Chapter >> 12. These are mutually dependent sections and differ quite a bit from >> other packages in the book. I'll put the binary install instructions in >> the About Java section. That should remove the circular dependency in >> the book now. > > I like these ideas. > > Some update. > > Tests reached same stuck point, is there now for more than two hours. > > Point where it is: > > $ tail OpenJDK-1.7.0.40-2.4.1-re-run-tests-2013.07.16-12h03m32s.log > Passed: java/awt/Frame/FrameSetSizeStressTest/FrameSetSizeStressTest.java > FAILED: java/awt/Frame/FrameSize/TestFrameSize.java > Passed: java/awt/Frame/HideMaximized/HideMaximized.java > Passed: java/awt/Frame/HugeFrame/HugeFrame.java > FAILED: java/awt/Frame/InvisibleOwner/InvisibleOwner.java > Passed: java/awt/Frame/LayoutOnMaximizeTest/LayoutOnMaximizeTest.java > FAILED: java/awt/Frame/ResizeAfterSetFont/ResizeAfterSetFont.java > FAILED: java/awt/Frame/WindowDragTest/WindowDragTest.java > FAILED: java/awt/FullScreen/FullScreenInsets/FullScreenInsets.java > Passed: > java/awt/FullScreen/NoResizeEventOnDMChangeTest/NoResizeEventOnDMChangeTest.java > > I will left tests running about 6 hours. Nothing done, after last line > above.
See what the test is after FullScreenInsets.java. That's where the problem is. Also check the load with top. If the system is idle, there is probably a hang. > For several months, had not run the tests. It seems that a new test > patch might be necessary. There are so many test failures we know since > long... In the book, we have: "The included version of jtreg is old, and > the test suite is also very dependent on the host system and the > environment that it is run in". > > But real word use seems to be ok (browser plugin, ant, JUnit, have also > installed maven, national program for income tax in my country is in > java, sites that need java working, these which I remember). I have > difficulty in finding the tests really necessary. > > I should be wrong in the following, please correct me, will not argue. > > The only advantage I see in the tests is to have fun trying to learn and > solve build problems, i.e., it is kind of educational. > > What is really the necessity of tests here? Would it be acceptable > commenting out the tests part just for the time being? > > Please, just trying to discuss and learn. My view is that we (BLFS) need to run the tests to have the confidence in our instructions. The user can do it or not, but few users should want to do a 130 SBU test suite. In LFS, I put in a note some time ago that the tests for automake were not recommended, basically because they take a long time -- over 30 SBUs. Some packages are just large -- libreoffice comes to mind. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
