danw [Tue, 19 Jan 1999 03:25:44 +0000 (03:25 +0000)]
Make the flow of control more obvious by removing a lot of code that
isn't doing anything useful since we removed lastname and
firstname_lastname aliases.
danw [Mon, 18 Jan 1999 21:44:34 +0000 (21:44 +0000)]
- Rewrite address output code to use more obvious variable names,
no global variables, and more comments.
- Check list table in addition to user table when starting
(fixes [1242] in moira)
- Save time by only loading the machines that are used as POP servers,
instead of all of them.
(none of these change the output at all)
danw [Tue, 29 Dec 1998 02:11:27 +0000 (02:11 +0000)]
add a "-q" (quote output) option so mrtest-parsing scripts can tell
newlines and commas in the content apart from newlines and commas as
delimiters. Not that there are any mrtest-parsing scripts or anything.
danw [Sun, 20 Dec 1998 21:26:33 +0000 (21:26 +0000)]
make department and xdept 50-characters long to match the Data
Warehouse. (We had previously been getting abbreviated names like
"Elec Eng & Comp Sci", but the Warehouse has names like "Electrical
Engineering & Computer Science, Dept Of", which we don't want to just
truncate)
danw [Sun, 20 Dec 1998 21:23:18 +0000 (21:23 +0000)]
fix FixCase() to only capitalize the letter after an apostrophe if the
apostrophe is the second character in the word. (eg, "O'BRIEN" ->
"O'Brien", but "PRESIDENT'S OFFICE" -> "President's Office", not
"President'S Office".)
danw [Mon, 14 Dec 1998 18:53:33 +0000 (18:53 +0000)]
Allow hostname to start with a digit in hostname_check, but disallow it
in access_host/access_ahal. (So you can make hostnames start with a number
only if you're on the query acl.)
danw [Tue, 10 Nov 1998 19:08:04 +0000 (19:08 +0000)]
Check the length of a new user's full name against USERS_FULLNAME_SIZE
in case that's shorter than USERS_FIRST_SIZE + USERS_MIDDLE_SIZE +
USERS_LAST_SIZE (which it is at the moment)
danw [Tue, 10 Nov 1998 19:06:31 +0000 (19:06 +0000)]
use "login = :argv[0]" instead of "login LIKE :argv[0]" in access_login:
we haven't done regexp conversion on the string yet, and if we're in
access_login, the user doesn't have bits to use regexps on this query
anyway
kcr [Fri, 30 Oct 1998 18:10:35 +0000 (18:10 +0000)]
Have chsh look up the fmodtime instead of the modtime in the user table.
Made uglier by the fact that you have to look up the fmodtime and the shell
with two different queries. Should be revisited if danw ever adds gush.
kcr [Fri, 30 Oct 1998 18:08:20 +0000 (18:08 +0000)]
Change the uush (update user shell) query to change the fmodtime field
instead of the modtime field in the users table so that deactivations can
more accurately depend on modtime.
danw [Mon, 19 Oct 1998 20:18:58 +0000 (20:18 +0000)]
Add a cast to (unsigned char) in a sprintf argument. Otherwise,
default promotions cause "\377" = 255 = -128 to be written as
"\37777777600" instead of "\377", causing a buffer overrun.
Don't try to build afssync or incremental if we don't find AFS, and don't
try to build afssync if that directory is missing (as it will be in the
public release)
danw [Tue, 25 Aug 1998 15:58:18 +0000 (15:58 +0000)]
rewrite FORCED_USERNAME message since if we're returning it from SWRD,
(1) we won't have the user's full name and account class available, and
(2) the user will already have seen FOUND, so it's weird to say "I've
found you in the database" a second time.
danw [Tue, 25 Aug 1998 15:52:48 +0000 (15:52 +0000)]
find_usernames returns NULL either for "can't suggest a username" or
"an error occurred". make sure it sets or unsets errno appropriately and
check errno as well as the return value when calling it. fixes [1185] in
moira / [1519] in moira-admin
danw [Thu, 13 Aug 1998 18:20:51 +0000 (18:20 +0000)]
remove encrypted ID support: if we do load in someone who already has
an account with an encrypted ID field, encrypting their current MIT ID
and comparing against it will fail anyway, since the encrypted ID
would be their SSN or old MIT ID. So the code is useless.