As mentioned in a previous message, when you network-mount a folder that is
owned by a different UID, it gets flagged as "read-only" regardless of whether
you actually can write to it. I have a one-line patch which turns off this
behavior (included below). I was thinking it would be good to have a
command-line option which activates or deactivates the UID check. But this
would require some way to make the folder_read subroutine aware of
command-line parameters. There seems to be a number of ways to do this--
1) making the parameters global variables or
2) passing relevant values as parameters to the folder_read function
are two possibilities that came to mind. What do people here think of this?
So currently, I patch my kernel to never check UID's -- is there anything I
have overlooked that makes this a particularly Bad Idea?
Also, I powered up an old computer for the first time in a while and it took
me almost an hour to figure out why mail wasn't going out (I had it configured
to use a SMTP server I don't use anymore) I have another one-line patch below
which adds some extra diagnostic information (for -verbose) which might save
someone in a similar situation a little time.
--- nmh-1.0.4/sbr/folder_read.c Fri Apr 30 14:08:34 1999
+++ nmh-1.0.4-new/sbr/folder_read.c Sun Jan 21 04:41:42 2001
@@ -53,7 +53,7 @@
mp->numsel = 0;
mp->nummsg = 0;
- if (access (name, W_OK) == -1 || st.st_uid != getuid())
+ if (access (name, W_OK) == -1)
set_readonly (mp);
prefix_len = strlen(BACKUP_PREFIX);
--- nmh-1.0.4/uip/post.c Fri Mar 3 02:24:42 2000
+++ nmh-1.0.4-new/uip/post.c Sun Jan 21 04:29:47 2001
@@ -1409,6 +1409,7 @@
bccque ? "Blind" : "Sighted");
else
printf (" -- Posting for All Recipients --\n");
+ printf("Connecting to %s\n", serversw ? serversw : servers);
}
sigon ();
