24
24
As of 10.03.2007, the xml package does not support namespaces yet, so you might wish
25
25
to exclude namespace tests. To do this, edit xmlconf/xmlconf.xml file and comment out
26
two lines at the bottom which reference 'eduni-ns10' and 'eduni-ns11' testsuites.
28
(The last lines should look like:
26
the lines that contain references &eduni-ns10; &eduni-ns11; and &eduni-nse;
30
--------------------------------------------
31
The following issues were encountered while testing the parser. Fortunately, none
32
of these change the category of any test, but in some cases cause incorrect error
33
message and/or postion to be reported.
35
1) xmltest/not-wf/sa/081.xml
36
xmltest/not-wf/sa/082.xml
37
xmltest/not-wf/sa/083.xml
38
xmltest/not-wf/sa/084.xml
40
All four reference an external entity with SystemID 'nul', which is a reserved
41
name under Windows (you won't be able to create such file). The archive contains
42
a file named 'nul.ent' that differs from entity's SystemID, so it won't resolve
43
anyway even in non-Windows.
44
This issue does not have any effect on FCL parser.
45
Additionally, tests 083.xml and 084.xml contain a reference to undefined notation.
46
This cause an extra validation error to be reported before the fatal error.
51
Both tests are missing ']' that should close the internal DTD subset.
57
All three have a NOTATION attribute declared on EMPTY element. This causes an extra
58
validation error to be reported before the fatal one.
60
4) ibm/xml-1.1/not-wf/p02/ibm02n66.ent
62
Presumably, missing '<' at start of CDATA. Does not change the diagnostic, though.
64
5) ibm/not-wf/p23/ibm23n05.xml
66
Contains encoding name 'ASCII' which is not supported by the parser. As a result, it aborts
67
before detecting the illegal XML declaration closing sequence.
69
6) ibm/not-wf/p72/ibm72n09.xml
71
Missing whitespace between 'ENTITY' and '%' at line 6 is detected before the bad tag closing
74
7) ibm/not-wf/p77/ibm77n01.ent
76
Invalid encoding name 'UTF8' is detected before the wrong token order.
78
8) sun/invalid/attr03.xml
79
sun/invalid/attr04.xml
80
sun/invalid/attr15.xml
82
Have a NOTATION attribute is declared on EMPTY element. Diagnostics incorrect.
84
9) ibm/invalid/p56/ibm56i11.xml
85
ibm/invalid/p56/ibm56i12.xml
86
ibm/invalid/p56/ibm56i14.xml
87
ibm/invalid/p56/ibm56i15.xml
89
Contain a reference to undeclared notation 'gif'. Diagnostics incorrect.
91
10) eduni/xml-1.1/052.xml
94
Intended to test handling of NEL and LSEP chars as element content whitespace, these
95
tests enclose NEL and LSEP within ordinary ascii chars ('abc_def') that are clearly not
96
a whitespace. A 'correct' error is therefore reported regardless of actual NEL/LSEP handling.
98
11) ibm/not-wf/p69/ibm69n06.xml
99
ibm/not-wf/p69/ibm69n07.xml
101
Designed to check parameter entity recursion, both tests contain PE references within entity
102
value declarations in internal DTD subset, which is a fatal error by itself.
104
12) ibm/not-wf/p21/ibm21n01.xml
106
Tests illegal CDEnd, but has an extra '[' in CDStart, which is detected earlier.
108
13) ibm/not-wf/p21/ibm21n02.xml
110
Tests illegal CDEnd, but has lowercase 'cdata' in CDStart, which is detected earlier.
112
14) ibm/xml-1.1/not-wf/p02/ibm02n58.xml
114
The first illegal character 0x99 is at position (2, 24), but another one at position (4,7) is
115
represented with malformed UTF-8 sequence (0xC1 0xA3, while correct one is 0xC2 0x99).
116
An 'xml-unaware' decoder can detect this before processing any 'normal' characters,
117
so diagnostics may be wrong.