This commit was generated by cvs2svn to track changes on a CVS vendor
[unix-history] / usr.sbin / sendmail / src / TODO
CommitLineData
15637ed4
RG
1(Version 4.1 of 7/25/83)
2
3-- get rid of MAXMAILERS, just allocate them as you go; merge the stats
4 structure in with the MAILERS structure; change the longs in
5 said structure into u_longs; print the file in ASCII, lock the
6 file before reading/printing.
7
8-- clean up sysexits; get rid of EX_BASE, most of sysexits.c is duplicated
9 in deliver.c
10
11-- mailq; split it out, by default, just # of messages, also add sort
12 by machine?
13
14-- flag so that if it has to forward mail it returns a note giving the
15 relay address?
16
17 ****** Finish properly implementing SMTP:
18 - check correct name in HELO exchange (?)
19
20 ***** Add an accounting package. [acct.c]
21
22 ***** When processing the queue, batch messages on the basis of
23 receiving host rather than message -- so that if connections
24 are expensive to complete, we get the most for our money.
25 Unfortunately, this is a major change, and will probably
26 wait for the next version.
27
28 **** Put in some more performance hooks.
29
30 *** Integrate a null name server for future hackers. [names.c]
31
32 *** Have "Operating System Error" give more info (e.g.,
33 "cannot fork"). [It really seems like this is ok now -- but
34 this is an old delivermail gripe -- check it out more?]
35
36 ** Per-mailer timeouts. Invalidates old .cf files.
37
38 ** Be able to generate a "response" (edit headers, etc.)
39 given a message to respond to.
40
41 * Have VRFY act as a limited name server (?).
42
43 * Implement SEND and related commands in SMTP.
44
45 * Extend full name matching. This should actually be done in a
46 complete name server. Should the name server be integrated
47 into sendmail (via a portal?) or should I "assume" that it
48 has been called by the time I am called?