aboutsummaryrefslogtreecommitdiffhomepage
path: root/generic
diff options
context:
space:
mode:
authorGravatar David Aspinall <da@inf.ed.ac.uk>2000-02-29 08:17:17 +0000
committerGravatar David Aspinall <da@inf.ed.ac.uk>2000-02-29 08:17:17 +0000
commitbcd950491a466afa67c9b638686ad71c854045d2 (patch)
treebb81b9cb92ed9cc26e6d99fb36f57451f6df43f7 /generic
parent191d2381383576373368bd0712355e668d9a2ea6 (diff)
Forced process-connection-type always to nil, after all. May not be an issue with non-mule FSF Emacs (that was something different).
Diffstat (limited to 'generic')
-rw-r--r--generic/proof-shell.el13
1 files changed, 4 insertions, 9 deletions
diff --git a/generic/proof-shell.el b/generic/proof-shell.el
index 532e6c65..91fa4398 100644
--- a/generic/proof-shell.el
+++ b/generic/proof-shell.el
@@ -227,16 +227,11 @@ Does nothing if proof assistant is already running."
;; use a pipe, not a pty, by setting process-connection-type=nil.
;; It seems to work, but has known disadvantages.
;; (I've forgotten what they are, size problems?).
- ;; Unfortunately, this appears to conflict with *non*-mule
- ;; versions of FSF Emacs which don't work properly
- ;; with piped communication. (Anyone know why?)
- ;; So non-mule FSF loses with Solaris.
(process-connection-type
- ;; FIXME: test for Solaris here to use pty's on Linux
- (if (or (string-match "XEmacs" emacs-version)
- (boundp 'mule-version))
- nil
- process-connection-type))) ; default
+ ;; FIXME: might be good to test for Solaris here to use
+ ;; pty's on Linux as not problematic.
+ nil))
+
;; An improvement here might be to catch failure of
;; make-comint and then kill off the buffer. Then we