On 08/04/2017 20:56, Paul Rogers wrote:
As of your request, it makes no sense that the maintainer of the java
page publish his test results. A point does not make statistics...
Also, I do not have always the time (or the will) to sort out all the
test failures. Last time I run the tests, I had consistently more than
50 errors, all in the same area, so I think it was more of a problem
with my configuration than with the package. And all in all, it was
just 50 out of more than 6,000. So what would be the value of
publishing those tests results?
Pierre, I have neither the inclination nor skill to chase down and
evaluate the failures myself. Let me be clear, the only thing I'm
asking for is these four lines from the version chosen for the book
be CAPTURED at the time it is the current release, and presented in
the book for reference:
jdk8u152-b01 Passed Failed* Not Run** Error Summary
jdk 5085 16 880 1 5982
hotspot 658 43 17 4 722
langtools 3109 0 0 4 3113
Certainly these numbers now 50 updates since the version I will build
are of little to no use to me. Of course, this is not the first place
in the book where we may be told, "YMMV". I think we are in agreement
that having run the tests, if we were to get numbers approximately like
those, then *probably* the build went OK and we can move on.
I think what is important is not the number of failures (or pass).
On the page the book links to, you can get a listing of the tests
with their status by clicking on the numbers under "Summary".
You can then compare to an identical listing on the host, and see
if the same tests have failed...
I have open a ticket to try to improve wording. I'll make a comment
about including what you suggest.
Regards
Pierre
--
http://lists.linuxfromscratch.org/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
