> This is about a now-rather-old curses - [...]
> [...misbehaviour...]
It's a bug in curses; it assumes newly-malloc()ed memory is cleared in
a least one tiny respect (NONL in a new screen is uninitialized; in the
broken run, it ends up being -1 instead of 0). Setting
MALLOC_OPTIONS=AJVX (my usual debugging setting) makes the "working"
command misbehave too; it appears to "work", when it does, because
something else has already grabbed the non-zeroed malloc()ed block, so
curses gets new memory.
Unfortunately this is the 5.2 curses, so it's not entirely trivial to
run it under my undefined-memory checking SPARC emulator. I may try to
make it build there, or possibly on a work Linux machine where I can
use valgrind, to figure out whether there are other such issues.
This change makes the symptom go away for me:
--- a/lib/libcurses/screen.c
+++ b/lib/libcurses/screen.c
@@ -117,7 +117,7 @@ newterm(char *type, FILE *outfd, FILE *infd)
if ((type == NULL) && (sp = getenv("TERM")) == NULL)
return NULL;
- if ((new_screen = (SCREEN *) malloc(sizeof(SCREEN))) == NULL)
+ if ((new_screen = (SCREEN *) calloc(sizeof(SCREEN),1)) == NULL)
return NULL;
#ifdef DEBUG
It obviously is not the *right* fix, but it's good enough for me for
the moment.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML [email protected]
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
