I’m writing this mail before I submit an inappropriate ticket against
Rails 1.1.6. This could be answered with “Well, don’t do that then!”,
but I encountered something peculiar.
In a Rails application directory, without a network connection, I did
$ rake rails:freeze:edge
which results in the following commands executed:
rm -rf vendor/rails
mkdir -p vendor/rails
after which svn export from “http://dev.rubyonrails.org/svn/rails/trunk”
Because of no network connection, that obviously fails with:
svn: PROPFIND of ‘/svn/rails’: Could not resolve hostname
and rake is aborted.
Problem: after that every subsequent
fails with the error:
no such file to load –
and there is no way (that I could find) to repair this situation.
The strange thing is that this file wasn’t available in the beginning,
but it wasn’t a problem then. The only difference is that in the
beginning vendor/rails/ was not present, but after the first rake task
it is created.
Possible solution: just like a pre-check in the rake task is done
whether svn is present, the rake task could check whether
“http://dev.rubyonrails.org/svn/rails/trunk” is accessible. That avoids
deletion of directories without having it filled by the task.
Is this worth submitting a ticket, or is this something that should stay
as it is?