Thank for your replying. > The recent Tramp 2.1.18, which is integrated into the upcoming Emacs > 23.2, has already changed this (although due to another problem). The > code looks like this: > > --8<---------------cut here---------------start------------->8--- > --- tramp.el.original 2009-07-08 22:22:51.000000000 +0900 > +++ tramp.el 2010-04-22 11:24:19.918277500 +0900 > @@ -4095,13 +4095,15 @@ > v 5 "Decoding remote file %s with function %s..." > filename loc-dec) > (funcall loc-dec (point-min) (point-max)) > - (let ((coding-system-for-write 'binary)) > + (let (file-name-handler-alist > + (coding-system-for-write 'binary)) > (write-region (point-min) (point-max) tmpfile))) > > ;; If tramp-decoding-function is not defined for this > ;; method, we invoke tramp-decoding-command instead. > (let ((tmpfile2 (tramp-compat-make-temp-file filename))) > - (let ((coding-system-for-write 'binary)) > + (let (file-name-handler-alist > + (coding-system-for-write 'binary)) > (write-region (point-min) (point-max) tmpfile2)) > (tramp-message > v 5 "Decoding remote file %s with command %s..." > --8<---------------cut here---------------end--------------->8--- > > Could you, please, check, whether it works also for you?
I checked that it works properly now. The above change is better than mine, I think. >> And, I use `plink' usually. As you know, inline transfer of `plink' is >> very, very slow. (Of course, I sometimes use `pscp' if file size is >> large.) >> So I made `tramp-handle-file-local-copy' and `tramp-handle-write-region' >> enable inline transfer compressing. The speed is approximately 1.5 - 4 >> times than the original. (In some cases, approximately 10 times.) >> Please check "tramp.el.jka-compr-inhibit+compress.diff". >> If it is valid, please apply it. > > Nice idea. However, Tramp is copylefted by the FSF. As a result, > non-trivial contributions could be accepted only in case the author > (you) has signed the corresponding legal FSF papers. Have you done this > already? Or would you be willing to do? I have done it for Emacs already. It is valid for tramp, isn't it? If invalid, I will do it for tramp again. > Thanks, and best regards, Michael. > _______________________________________________ Tramp-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/tramp-devel
