search

USGS-R / usgs-r.github.com

Splash page for USGS-R organization
4 stars 7 forks source link

adding geoknife to package repo #1

Closed jordansread closed 10 years ago

jordansread commented 10 years ago

Not sure why some of Dave's packages are NeedsComp = TRUE on here...

jordansread commented 10 years ago

@dlorenz-usgs yes, I think that must be the case. Do you know why your source packages were set to "NeedsCompilation: No"?

If you look at the diff, we are talking about restrend, rloadest, USGSwsQW, and USGSwsStats. Are those all packages that use FORTRAN or C?

ldecicco-USGS commented 10 years ago

Those are the packages that have fortran. Now the question is, why did your write_PACKAGES command change thoses?

dlorenz-usgs commented 10 years ago

Laura, I assume that when I last updated the packages that the write_PACKAGES function. I assumed that it checked the directory structure and set NeedsCompilation to yes if it found the src folder. I used the write_PACKAGES function from tools with these arguments: verbose=TRUE, subdirs=TRUE, type="win.binary" or "source", depending on what is appropriate, or "mac.binary" if I was able to build it (I think all are current, but were updated later). I could force the issue with a line in the DESCRIPTION file, but that is not supposed to be needed. Other ideas? Dave

On Tue, Jul 29, 2014 at 10:30 AM, Laura DeCicco notifications@github.com wrote:

Those are the packages that have fortran. Now the question is, why did your write_PACKAGES command change thoses?

— Reply to this email directly or view it on GitHub https://github.com/USGS-R/usgs-r.github.com/pull/1#issuecomment-50492746 .

jordansread commented 10 years ago

I think this is just a difference between my mac version of tools (v3.1.0) and your Windows versions (probably v3.1.1?)

It seems to install fine w/ "no" or "yes". "yes" makes more sense to me because it does have to compile. My machines seem to ignore the argument when compiling from source. Only tested rloadest.

dlorenz-usgs commented 10 years ago

Jordan, I am using R 3.1.0. But you are probably correct, if the src directory is in the source tarball, then that code gets compiled and NeedsCompilation is not used. But then what use is that field? Dave

On Tue, Jul 29, 2014 at 12:16 PM, Jordan S Read notifications@github.com wrote:

I think this is just a difference between my mac version of tools (v3.1.0) and your Windows versions (probably v3.1.1?)

It seems to install fine w/ "no" or "yes". "yes" makes more sense to me because it does have to compile. My machines seem to ignore the argument when compiling from source. Only tested rloadest.

— Reply to this email directly or view it on GitHub https://github.com/USGS-R/usgs-r.github.com/pull/1#issuecomment-50507787 .

ldecicco-USGS commented 10 years ago

I'm still not sure on the difference, but I don't think anything will be broken. If so, we can revert, I'm going to do this merge. I'll try installing a couple of the packages on both my windows and mac to see how it goes.

jordansread commented 10 years ago

I am guessing that the NeedsComp is a field that is in the process of being depreciated. I couldn't find any explanation on the web about this, so that is just a guess.