Ar an t-ochtú lá de mí Lúnasa, scríobh Mats Lidell:
There is a problem with running the mule-tests from the build
slaves. The problem can't be repeated when I run the same build
manually. When I run manually the mule tests completes, not 100% but
they are not aborted.
I think you can observe the problem in all builds, even the green
ones. (This is another strange thing that the buildbot sometimes
signals this as an error and sometimes not.)
Any ideas what this can come from and how to fix it?
Basically, the korea-util.el in core has bugs that the korea-util.el in the
packages does not. The former is picked up by the buildbot, the latter when
running manually--this is one of the load-path shadows the build has been
warning us about for years.
We currently need korea-util.el in core because
#'setup-korean-environment-internal anda #'exit-korean-environment are
necessary for it to be possible to switch to and from the Korean language
environment. My suggestion would be to move these (and the
korean-key-bindings variable, necessary for them) to korean.el, and remove
korea-util.el from core. And do similar things for the other *-util.el mule
files that are shadowed by package files.
‘Iodine deficiency was endemic in parts of the UK until, through what has been
described as “an unplanned and accidental public health triumph”, iodine was
added to cattle feed to improve milk production in the 1930s.’
(EN Pearce, Lancet, June 2011)
XEmacs-Beta mailing list