70322ae3 |
1 | TODO list for agedu |
2 | =================== |
3 | |
4 | Before it's non-embarrassingly releasable: |
5 | |
f2e52893 |
6 | - cross-Unix portability: |
7 | + use autoconf |
8 | * configure use of stat64 |
9 | * configure use of /proc/net/tcp |
10 | * configure use of /dev/random |
11 | * configure use of Linux syscall magic replacing readdir |
12 | + later glibcs have fdopendir, hooray! So we can use that |
13 | too, if it's available and O_NOATIME is too. |
14 | * what do we do elsewhere about _GNU_SOURCE? |
15 | |
5a29503d |
16 | - man page, --version. |
9d0b9596 |
17 | |
cfe942fb |
18 | Future possibilities: |
70322ae3 |
19 | |
1e8d78b9 |
20 | - IPv6 support in the HTTP server |
cfe942fb |
21 | * of course, Linux magic auth can still work in this context; we |
22 | merely have to be prepared to open one of /proc/net/tcp or |
23 | /proc/net/tcp6 as appropriate. |
1e8d78b9 |
24 | |
70322ae3 |
25 | - run-time configuration in the HTTP server |
26 | * I think this probably works by having a configuration form, or |
27 | a link pointing to one, somewhere on the report page. If you |
28 | want to reconfigure anything, you fill in and submit the form; |
29 | the web server receives HTTP GET with parameters and a |
30 | referer, adjusts its internal configuration, and returns an |
31 | HTTP redirect back to the referring page - which it then |
32 | re-renders in accordance with the change. |
33 | * All the same options should have their starting states |
34 | configurable on the command line too. |
35 | |
70322ae3 |
36 | - curses-ish equivalent of the web output |
37 | + try using xterm 256-colour mode. Can (n)curses handle that? If |
38 | not, try doing it manually. |
f2e52893 |
39 | + I think my current best idea is to bypass ncurses and go |
40 | straight to terminfo: generate lines of attribute-interleaved |
41 | text and display them, so we only really need the sequences |
42 | "go here and display stuff", "scroll up", "scroll down". |
f2e52893 |
43 | + Infrastructure work before doing any of this would be to split |
44 | html.c into two: one part to prepare an abstract data |
45 | structure describing an HTML-like report (in particular, all |
46 | the index lookups, percentage calculation, vector arithmetic |
47 | and line sorting), and another part to generate the literal |
48 | HTML. Then the former can be reused to produce very similar |
49 | reports in coloured plain text. |
70322ae3 |
50 | |
f2e52893 |
51 | - http://msdn.microsoft.com/en-us/library/ms724290.aspx suggest |
52 | modern Windowses support atime-equivalents, so a Windows port is |
373a02e5 |
53 | possible in principle. |
54 | + For a full Windows port, would need to modify the current |
55 | structure a lot, to abstract away (at least) memory-mapping of |
56 | files, details of disk scan procedure, networking for httpd. |
57 | Unclear what the right UI would be on Windows, too; |
58 | command-line exactly as now might be considered just a |
59 | _little_ unfriendly. Or perhaps not. |
60 | + Alternatively, a much easier approach would be to write a |
61 | Windows version of just the --scan-dump mode, which does a |
62 | filesystem scan via the Windows API and generates a valid |
63 | agedu dump file on standard output. Then one would simply feed |
64 | that over the network connection of one's choice to the rest |
65 | of agedu running on Unix as usual. |