The old solution didn't really work all that well: you look at a page,
then you load the raw copy, you edit it, it finds the buffer
containing the raw copy (great!), reloads, it gets redirected to the
original page (great!). Then you return to the edit buffer, make a
change and save: now it doesn't find the buffer containing the raw
copy (since the Elpher buffer is showing the original page after the
redirect!) and so a new Elpher buffer gets created, which is then
redirected to the original page (now you have two of them). And if you
make another edit, you have three, and so on, forever. This is lame.
`add-to-list' causes issues when used in conjunction with another hook. For
example, in my config I add a hook using the standard `add-hook' function, which
initializes (apparently) `gemini-mode-hook' with /just/ the function I add.
Subsequent calls to `add-hook' convert the singleton to a list and add to it,
but `add-to-list' throws an error because `gemini-mode-hook' isn't a list.
Emacs ... eh?
We can't push titan into the address of the current page because if we
do, once we create a new buffer, we can't follow the redirect: it
points the old current-page, which now contains the titan scheme.
Also, if we create a new buffer because no elpher buffer is showing
our editor's current page, put that new buffer in elpher mode and set
a current page. If we don't use elpher mode, or doesn't have a current
page, following the redirect runs into errors.
Martin Joerg
Case Duckworth
Alex Schroeder
Ashley Duckworth