|
2 | 2 | CSS4J RELEASE NOTES |
3 | 3 | =================== |
4 | 4 |
|
5 | | -Release 2.0.3 - June 5, 2020 |
| 5 | +Release 2.0.4 - July 8, 2020 |
6 | 6 | ---------------------------- |
7 | 7 |
|
8 | 8 | Highlights |
9 | 9 | ---------- |
10 | | - Since 1.0, the library allows the use of the 'advanced' attr() function that is |
11 | | -described in recent CSS specifications (although not yet implemented in major |
12 | | -web browsers). For most use cases, this is just another feature like others, but |
13 | | -depending on how you are using the library, this may represent a security risk, |
14 | | -due to the possibility of leaking attribute values by malicious CSS. And in the |
15 | | -future, the addition of new functions -like the proposed 'concat()'- to CSS may |
16 | | -represent a real security issue for all users. |
17 | | - |
18 | | - While the CSS Working Group is discussing about the problem, this css4j release |
19 | | -brings a few restrictions on how the attr() function can be used. attr() will be |
20 | | -invalid if it is not used in the 'content' property, in the following cases: |
21 | | - |
22 | | - . It is applied on the 'value' attribute of the 'input' element. |
23 | | - . It is applied on the 'link' or 'meta' elements. |
24 | | - . The attribute name contains 'nonce', 'pass', 'user', 'session', 'uid' and |
25 | | - other similar potentially sensitive names. |
26 | | - |
27 | | - A full fix should be applied once the CSSWG settles on the issue. |
28 | | - |
29 | | - Several other fixes/improvements are provided, and all users are encouraged to |
30 | | -upgrade. |
| 10 | + Bugfixes, and the nu.validator htmlparser artifact is used instead of the (very |
| 11 | +similar) nu.validator.htmlparser's. |
31 | 12 |
|
32 | 13 |
|
33 | 14 | Upgrading from 1.0 |
|
0 commit comments