aboutsummaryrefslogtreecommitdiffhomepage
path: root/BUGS
diff options
context:
space:
mode:
authorGravatar David Aspinall <da@inf.ed.ac.uk>2002-07-12 08:14:27 +0000
committerGravatar David Aspinall <da@inf.ed.ac.uk>2002-07-12 08:14:27 +0000
commit9f223fa9ed8fbc437dd60b446e1e4f82e88a1096 (patch)
tree4c9a6065fd8496056505a458efeda180e620f406 /BUGS
parentdc236b0cdbcec9ceae637fdb90f5aef46eecdb22 (diff)
Mention probs with tracing mode.
Diffstat (limited to 'BUGS')
-rw-r--r--BUGS15
1 files changed, 15 insertions, 0 deletions
diff --git a/BUGS b/BUGS
index eb39f2e4..80146e4b 100644
--- a/BUGS
+++ b/BUGS
@@ -20,6 +20,21 @@ versions, and those which only apply to particular versions.
If you have more than one theorem with the same name in a buffer,
their proof visibilities are controlled together.
+** Issues with tracing mode (Isabelle only)
+
+1. Large volumes of output can cause Emacs to hog CPU spending
+all its time processing the output (esp with fontifying and X-symbol
+decoding). It becomes difficult to use normal editing commands,
+even C-c C-c to interrupt the prover. Workaround: hitting C-g,
+the Emacs quit key, will interrupt the prover in this state.
+See manual for further description of this.
+
+2. Interrupt response may be lost in large volumes of output, when
+using pty communication. Symptom is interrupt on C-g, but PG thinks
+the prover is still busy. Workaround: hit return in the shell buffer,
+or set proof-shell-process-connection-type to nil to use piped
+communication.
+
** If the proof assistant goes into a loop displaying lots of information
It may be difficult or impossible to interrupt it, because Emacs