Annotation of hypermess/hypermail/KNOWN_BUGS, revision 1.5
1.3 kahan 1:
2: This is a list of know problems with the current version of Hypermail
3:
4: =================================================
5: Problem: Converting existing hypermail archives
6: to use a different HTML file suffix:
7:
8: If you are going to change the filename suffix used from the default
9: ".html" to something else, you must deal with the existing files first.
10: Otherwise, hypermail does not know they are there and starts numbering
11: files from the begining.
1.1 kahan 12:
1.3 kahan 13: This means that you need to either start out fresh, regenerate the
14: archive from an existing mailbox or convert all existing files and
15: the links within the existing archive files to use the proper extension.
16:
17: =================================================
18: Problem: File Locking On NFS Mounted Directories:
19:
20: File locking doesn't work all that well over NFS. For that reason
21: you are encouraged to run Hypermail on the same system that the
22: disk is on. This is due to possible timing issues within NFS.
23: Hypermail based network locking is not expected to work at this time.
24:
25: =================================================
26: Problem: Message Dates:
27:
28: Dates on messages prior to 1970 are not processed correctly.
29:
30: =================================================
31: Problem: Not parsing correctly.
32:
33: The following line is not parsing correctly.
34:
35: news:fr.comp.os.ms-windows.win95 post news:34EF975C.4097FF89@club-internet.fr
36:
37: =================================================
38: Problem: Embedded HTML
39:
40: If HTML file included in message it is just putting it there
41: without concern of BASE, TITLE, HEAD, BODY ...
42:
43: =================================================
44: Problem: Blank name not interpreted correctly.
45:
46: Puts out a <STRONG></STRONG> if name is blank.
47:
48: =================================================
49: Problem: Memory leaks
50:
51: Leaking < 30 bytes per message processed.
52:
53: =================================================
54: Problem: Empty container elements
55:
56: Putting out empty container elements <P>. Should be <P> </P>
57:
58: =================================================
59: Problem: No verification of link syntax
60:
61: Not verifying a valid HTML link syntax before converting somthing
62: with a http://.
63:
1.4 kahan 64: =================================================
65: Problem: indextable format problem
66:
67: The message list column widths for "subject" "author" and "date" are sized
68: based on the content. Subjects and dates are fairly predictably riddled
69: with white space on which long lines can wrap.
70:
71: The "author" string is not -- in particular, I have a person with a long
72: pseudonym (having a descriptive pseudo-middle name, used to distinguish
73: between several people with the same first and last name), and a long
74: machine.domain identifier -- coming out to 52 (!) characters total. His
75: MUA does not give a "real name" string that hypermail could use.
76:
77: In a typical browser display, this forces the 'author' column to about
78: 2/3rds of the display width, and the subject and date get crammed into the
79: remaining 1/3rd.
80:
81: The "hm_usetable" resource only affects headers and footers, and not the
82: bulk of the index display itself. Is there an existing configuration
83: choice that could help me?
84:
1.5 ! fumi 85: =================================================
! 86: Problem: corrupt mailbox handing.
! 87:
! 88: If the parser never finds the end-boundary in the alternative section it
! 89: leaves the parser in an undefined state, it still treats "^From_"
! 90: (underscore to display space) lines as the beginning of a new mail. The
! 91: problem I think is more related to that the "alternative parser" never really
! 92: finishes its work until it finds the end-boundary and when the following
! 93: "From " line starts a new mail, something bad happens. Hypermail should
! 94: process corrupt messages a bit better.
! 95:
! 96: =================================================
! 97: Problem: Building on DEC Alpha systems.
! 98:
! 99: Make sure and use the GCC compiler and not the native DEC compiler
! 100: when building on a DEC Alpha using TRU64 (OSF1).
! 101:
1.4 kahan 102:
Webmaster