Mail Archives: djgpp/1998/03/31/03:06:01
On Mon, 30 Mar 1998, Ralph Proctor wrote:
> After 18 lines of "patching file..." notes, it stopped and I saw this:
>
> Reversed (or previously applied) patch detected! Assume -R? [n]
This means that `patch' detected a file that seems to be already
patched. `patch' then suggests to reverse (undo) the patch--that's
what the -R switch does--in case this is what you want.
So it appears that you have some of the sources already changed.
Nothing to worry about, except that you need to look at the patches
which were rejected (they are in the files whose names end with a `#'
on plain DOS and with `.rej' on Windows 9X; `patch' prints the names
of these ``reject files'' whenever it doesn't apply some of the
patches). Look at the rejected hunks and compare them with the
sources you have, to see whether all of the rejected patches are
already in your sources. If not, edit the sources manually to bring
them into sync.
> Then it went to a conclusion, but on the way there were some discouraging
> remarks which I do not understand.
That is the punishment reserved for those who use a tool (in this
case, `patch') without reading its docs ;-).
> Now, maybe the problem is I did not answer the questions properly.
No, you really did the right thing there.
Btw, I suggest everybody to use `patch' with the -b switch. This
forces `patch' to always leave the original version of the file as
backup (with `.orig' extension on LFN platforms and with `~' as the
last file name character on non-LFN ones), so if worse comes to worst,
you still have the original file available.
Shawn, if your patching instructions don't include -b, I suggest
adding it.
- Raw text -