> In case you're wondering what's happening with sane-1.0.1: I'm
> basically ready with the (hopefully) last pre-release but I currently
> am experiencing difficulties contacting the ftp server (hosted by
> redhat). As soon as that problem has been resolved, I'll update the
> ftp site and send a new announcement (in the meantime, please send me
> patches that solve serious bugs only).
On April 6 I sent a patch to sane-devel, which fixed a few bugs in the
SCSI support for FreeBSD.
However, now it seems that in a misdirected effort to clean up my
changes before generating and sending the patch I seriously botched
it...
Here is the faulty bit:
--- sanei/sanei_scsi.c.org Mon Apr 5 14:26:35 1999
+++ sanei/sanei_scsi.c Mon Apr 5 14:58:17 1999
@@ -1689,7 +1689,7 @@
char* data_buf;
size_t data_len;
- if (fd < 0) {
+ if (fd < 0 || fd > CAM_MAXDEVS || cam_devices[fd] != NULL) {
fprintf(stderr, "attempt to reference invalid unit %d\n", fd);
return SANE_STATUS_INVAL;
}
A quick visual inspection of the changed line should be enough to
realize that the "!= NULL" should be "== NULL". I.e. it should read:
"if the faked fd is out-of bounds or there is no such open device"
That is not quite what the above evaluates to :(
I sincerely apologize of being sloppy, and hope that a correction
will make it in time for the release.
/Mikko
P.S. I *do* consider this serious, as it completely hoses
SANE under FreeBSD.
Mikko Tyo"la"ja"rvi________________________________mikko@securitydynamics.com
SecurityDynamics
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com