"Jeffrey C Honig" wrote:
When running under mh-e, I set the MHCONTEXT environment variable to
/dev/null. This prevents my mh-e context from confusing my command line
context.
With 1.1 this breaks things.
We could allow MHCONTEXT=/dev/null with the following trivial patch.
Google reveals a script or two that uses this trick. I'm inclined to
apply it (although it seems a bit ugly) -- objections?
Index: sbr/context_read.c
===================================================================
RCS file: /cvsroot/nmh/nmh/sbr/context_read.c,v
retrieving revision 1.7
diff -u -r1.7 context_read.c
--- sbr/context_read.c 13 Nov 2005 23:51:18 -0000 1.7
+++ sbr/context_read.c 14 Nov 2005 22:19:28 -0000
@@ -120,8 +120,11 @@
if ((cp = getenv ("MHCONTEXT")) == (char *)0 || *cp == '\0')
cp = context;
- /* context is NULL if context_foil() was called to disable use of context
*/
- if (!cp) {
+ /* context is NULL if context_foil() was called to disable use of context
+ * We also support users setting explicitly setting MHCONTEXT to /dev/null.
+ * (if this wasn't specialcased then the locking would be liable to fail)
+ */
+ if (!cp || (strcmp(cp,"/dev/null") == 0)) {
ctxpath = NULL;
return;
}
-- PMM
_______________________________________________
Nmh-workers mailing list
Nmh-workers@nongnu.org
http://lists.nongnu.org/mailman/listinfo/nmh-workers