1 # $Id: TODO,v 8.2 1997/04/12 15:53:31 bostic Exp $ (Berkeley) $Date: 1997/04/12 15:53:31 $
3 ========================================================================
5 ========================================================================
6 2.0: Open mode is not yet implemented.
9 ========================================================================
11 ========================================================================
12 1.N When nvi edits files that don't have trailing newlines, it appends
13 one, regardless. This is required, by default, from POSIX.2.
15 1.N: If you run out of space in the recovery directory, the recovery
16 file is left in place.
18 2.0: It's known that it's possible to sync the backing files in the
19 wrong manner, leaving backup files that aren't recoverable. This
20 is going to be left alone until we have a logging version of DB,
21 which will hopefully fix this (or at least make it possible to
24 ========================================================================
26 ========================================================================
27 1.N: Make the search pattern and command history (what the '.' command
28 executes) remembered between windows, or configurable so that it
31 1.N: Change the screen scrolling to not eat user characters... i.e.
32 g/pattern/foo should not eat already entered chars.
34 1.N: The vi main command loop should use the general-purpose overflow
35 and underflow routines. In addition, the vi command loop uses
36 unsigned longs -- should probably be fixed as a 32-bit unsigned
37 type, and then check to make sure it's never used as as variable
40 1.N: Should "view" set a lock on the file?
42 1.N: Should "view" copy the file (or even open a recovery file)?
44 1.N: The strings found by searches should be highlighted until the next
47 1.N: Display a split vi screen for the :help command.
49 1.N: When getting a key for a continue screen, we should always read from
50 the terminal, not from a mapped key.
52 1.N: The sentence, paragraph and section movement commands don't match
53 historic practice in some boundary cases. This should be left
54 alone until POSIX 1003.2 makes up its mind.
56 1.N: The vs_sm_fill routine should scroll if possible, not always redraw.
58 1.N: Think about setting a dirty/inuse bits on the lines of the SMAP
59 structure. That way the message routines could steal lines and
60 refresh would continue to work, because it would know not to touch
61 the lines that were in use.
63 ========================================================================
65 ========================================================================
66 2.0: ^C isn't passed to the shell in the script windows as an interrupt
69 2.0: It would be nice to inverse video the replaced text during
70 interactive substitute.
72 2.0: The :args command should put the current file name out in reverse
73 video. This isn't going to be easy, currently only full lines can
74 be in reverse video, not just parts.
76 ========================================================================
78 ========================================================================
79 1.N In single-line screens, have to press 'q' twice when quitting out
80 of a ":set all" display.
82 ========================================================================
84 ========================================================================
85 1.N: We currently permit the user to change the lines, columns and term
86 edit options. Shouldn't that be illegal in a window interface?
88 ========================================================================
90 ========================================================================
91 2.0: There's a serious problem with error returns -- we need to separate
92 command failure from fatal error, consistently, over the entire source
93 tree. We need to rework all of vi to have three return values:
97 Right now we don't recognize fatal errors for what they are.
99 2.0: The complete list of POSIX.1 calls that can return EINTR are:
100 wait, waitpid, sleep, dup2, close, read, write,
101 fcntl(SETLCKW) tcsetattr, tcdrain
102 The problem is that technically, any system/library call can
103 return EINTR, so, while nvi blocks (most of?) the obvious ones,
104 someone may have to do a complete pass and block signals
107 2.0: The options hardtabs, lisp, optimize, redraw, and slowopen
108 are recognized, but not implemented.
111 2.0: Field editing shouldn't be hard to add to nvi:
113 Field editing file template:
115 field # row/column start row/column stop
116 label field # Label string
117 re field # Matching re string.
118 field # row/column start row/column stop
119 label field # Label string
120 re field # Matching re string.
122 <tab> moves to the next field
123 <bs> in column 0 moves to the previous field