fBasics / fUtilities -- getCall() // Martin's summer absence
Martin Maechler
maechler at stat.math.ethz.ch
Wed Jun 22 09:33:01 CEST 2011
Dear Rmetrics co-maintainers,
{i.e. mainly dear Diethelm and Yohan, right ?}
For the next (major) version of R 2.14.0, due in October,
I've introduced a 'getCall()' {S3} generic with a default
method, which is used in stats::update() {for model updating}.
Because of that, I've found that the CRAN versions of fExtremes and fGarch
now warn.. ===> I've wanted to "fix" this smartly,
namely, depending on if the package runs in
R version <= 2.13.x (where there's no getCall() in stats)
and R version >= 2.14.0, define it as a function or as a Method.
(--> see another e-mail just to you two).
For that, I've looked in the R-forge versions of all Rmetrics
packages, and found that
*both* packages
fUtilities and fBasics
define a getCall() function.. and I've wondered, when I saw that
the two packages have basically the same
fBasics/R/utils-getS4.R
fUtilities/R/getS4.R
and I see more fBasics/R/utils-*.R source files.
Do I guess correctly, that
fUtilities is deprecated already now (together with fCalendar and fSeries)?
(-- but then, why did you (Yohan) still add changes to fUtilities in
February 2010?)
With respect to that:
Is there somewhere written up ... hopefully as part of the
R-forge sources, the current Rmetrics
strategy of dependencies of packages,
so I can read it and look it up ..
(and also add to it the fact that we agreed on factoring out
'stabledist' from 'fBasics')
?
Thank you for feedback...
BTW: Note that I will spend the next 3 days in Vienna (with part of R core),
and then travel to Columbia and Ecuador ... for full three weeks ...
up to July 18,
for vacation (visiting, traveling) where I won't work, and only
have quite intermittent and short internet connection, probably.
Best regards,
Martin
More information about the Rmetrics-core
mailing list