File: BUGS

package info (click to toggle)
curves 0.8.7
  • links: PTS
  • area: main
  • in suites: slink
  • size: 704 kB
  • ctags: 1,001
  • sloc: cpp: 6,197; ansic: 519; makefile: 270; sh: 192; fortran: 149
file content (30 lines) | stat: -rw-r--r-- 1,507 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
#
# This is a file containing known bugs.  While I try to fix errors
# when they are encountered, we all know that unexplained problems
# SOMETIMES appear.  We want to document these as well as anticipated
# problems that don't affect usability at the moment.  Eventually,
# this file will be replaced with the bug database that is planned as
# the next project. :-)
#  

- When we mmap files, we are mapping one extra byte so that I get a
  null termination at the end of file.  Why?  I'm using strcspn to
  find the end of line, but there is no provision for strcspn to check
  length limitations.  The call should be replaced.  As it turns out,
  Linux appears to zero-fill the mmap'd memory, but this is not
  documented and I've not checked the source code to see if it is
  guaranteed.  Other systems may choose to handle this differently, so
  there are no guarantees that this scheme will continue to work.

- Preference reading does not do any error checking.  We need to
  report errors in the preferences file so the user may correct them.

- Error reporting is weak overall.  There are many asserts in the
  code, but little is done to catch problems.  I've been good about
  this in several other projects.  I'm not sure why I'm being lazy
  now.

- The cvs status command does not report about files we have not yet
  seen.  This is rather annoying when we want to know if the directory
  needs updating.  We need to check to see if the newer versions of
  CVS fix this.  We have version 1.5.