Fixed a bug in how deletions of entries in an auto-vivified hashref works.
[dbsrgits/DBM-Deep.git] / t_attic / TODO
CommitLineData
2120a181 1=head1 NAME
2
3Testing TODO
4
5=head1 PURPOSE
6
7This file is to detail the tests, in a general sense, that have yet to be
8written so that I don't forget them.
9
10=head1 MISSING TESTS
11
12=over 4
13
14=item * Readonly filehandles
15
16=over 4
17
18=item * Mutations on readonly filehandles
19
20This is to verify that the appropriate errors are thrown
21
22=item * Run an optimize on a readonly FH
23
24=back
25
26=item * _copy_value()
27
28For some reason, $c doesn't seem to be undefinable in _copy_value. Maybe this
29means that the bless()ing should occur iff C<!$c-E<gt>isa('DBM::Deep')>?
30
31=item * Splice
32
33=over 4
34
35=item * Undefined initial offset
36
37=item * splicing in a group that's equal to the target
38
39=back
40
41=item * Passing in a fh without a file_offset
42
43=item * Do I ever use print_at() without passing in offset?
44
45=item * How should the inode check for locking happen?
46
47=item * medium and large pack_sizes
48
49Need to make sure I only run the large pack_size test on 64-bit Perls
50
51=item * max_buckets check
52
53=item * get_classname() on a deleted sector
54
55How should this be triggered?!
56
57=item * Open a corrupted file that has a header, but not initial reference
58
59=item * Max out the number of transactions
60
61=item * Delete something in the head that has its own value in a transaction
62
63=item * Run an import within a transaction
64
65=over 4
66
67=item * Should all assignments with a non-scalar rvalue happen within a sub-transaction?
68
69=item * Does this mean that sub-transactions should just be done right now?
70
71It shouldn't be too hard to variablize which transaction is the base instead
72of hard-coding 0 . . .
73
74=back
75
76=item * Delete something within a transaction, then commit.
77
78Verify that the space is reusable by assigning more to the DB.
79
80=back
81
82=cut