search

ropensci / neotoma

Programmatic R interface to the Neotoma Paleoecological Database.
https://docs.ropensci.org/neotoma
Other
30 stars 16 forks source link

Error in fetch(key) : internal error -3 in R_decompress1 #120

Closed SimonGoring closed 10 years ago

SimonGoring commented 10 years ago

We keep getting that error on other people's computer when they try to use the help.

lazy-load database '' is corrupt.

gavinsimpson commented 10 years ago

I wonder if this is related to the broken manual and the inability to build the PDF of the manual under R CMD check. I have also seen errors, but I don't think it was this one, if the user installed a new version of the package whilst using the older version in an R session.

Perhaps we try sort out what is causing the broken manual build and see if that helps?

SimonGoring commented 10 years ago

I have always had problems building the manual, it's been persistent, and I assumed it had something to do with my version of MikTeX, but maybe this is new.

gavinsimpson commented 10 years ago

It's not building on Linux either. Will dig out the error messages in morning but IIRC it was suggesting missing \item{} somewhere. Looks like a real problem, probably a syntax error someone in the Rd files.

I'll take a look in the morning by rendering all the .Re files to text which should show the error.

gavinsimpson commented 10 years ago

Well, the rendered Rd files parse OK in the Emacs + ESS previewer. I think what is causing the problems with building the PDF manual via LaTeX is that you are using a \tabular{} environment with the \describe{} environment in several value sections. You can't do that; you can only include \items in a \describe{}. So that is something we need to iron out and I suspect then the PDF manual will work OK.

And the changes in #121 and #122 now mean the manual will build, on Linux at least.

Now that those are in the tree I'll see if I can replicate this issue.

SimonGoring commented 10 years ago

Awesome. I just synced from upstream (ropensci) and I'm going to push a few small changes in the code based on bugs we encountered in the class on Wednesday and today, but none affect the help.

gavinsimpson commented 10 years ago

Ok, help seems to be working for me. I suspect that what is happening is that users are using devtools to install neotoma from within an R session and they are then loading neotoma via library() into the same session. But the act of installing the package has altered the man pages in their compiled representation being access by the current R session and hence the error.

If you can reproduce this, can you try to install neotoma using devtools, then exit that R session, restart R and load neotoma and see if the problem persists?

SimonGoring commented 10 years ago

I just tried to reproduce this on two different platforms and couldn't, so it looks like it's solved.