100% agree. But I am relaxed if it only gets applied to the binary
installer in the interim.
andy
At 13:03 04/11/98 +0100, Jan Vroonhof wrote:
This is the Efs-on-NT bugfix?. In fact I think that in that case it
should just be applied to the efs package (hopefully along with
something along the lines of my
efs-depends-no-longer-on-read-passwd-and-comint patch)
With a all due respect to the efs maintainers, I think these guidelines
1. Never apply patches to packages that have maintainers.
2. He we have a package system, we can fix it later after the release.
are generally OK. However they should not be absolute
a. We have a clear bug, it is critical for the package to work. We
have a working patch, it's been submitted to the maintainer for
inclusion in the mainline version.
b. The EFS maintainers currently are busy. Michael is refering to
argument 2.
c. Argument 2. IMHO does not apply here.
Why?
1. EFS is not just any package. By design EFS is more important for
the package tools than any other package, even xemacs-base! You
"cannot"[1] upgrade to the fixed package if the package system is
broken.
2. I think that at important points in XEmacs history such as new
version release one should make sure all packages are in a working
state as this is typically the time people will be
installing/upgrading packages too.
[1] Ok, you can do it by hand. Unfortunately currently you still have to
fetch EFS by hand.
---------------------------------------------------------------------------
" .sigs are like your face - rarely seen by you and uglier than you think"
Dr Andy Piper, Technical Architect, Parallax Solutions Ltd
mail: andyp(a)parallax.co.uk web:
www.parallax.co.uk/~andyp