* In addition to the /convert/ page which supports office file formats, we also put some time into improving interoperability for our existing apps. We're introducing the ability to export rich text documents as Markdown (via the [turndown](https://github.com/mixmark-io/turndown) library), to import trello's JSON format into our Kanban app (with some loss of attributes because we don't support all the same features), and to export form summaries as CSV files.
* In addition to the /convert/ page which supports office file formats, we also put some time into improving interoperability for our existing apps. We're introducing the ability to export rich text documents as Markdown (via the [turndown](https://github.com/mixmark-io/turndown) library), to import trello's JSON format into our Kanban app (with some loss of attributes because we don't support all the same features), and to export form summaries as CSV files.
* note: the currently merged implementation does not preserve all information for complex question/answer types such as polls, multi-line radios, or multi-line checkboxes. More improvements will be included in our next release.
* We've added another extension to our customized markdown renderer which replaces markdown images with a warning that CryptPad blocks remote content to prevent malicious users from tracking visitors to certain pages. Such images should already be blocked by our strict use of Content-Security-Policy headers, but this will provide a better indication why images are failing to load on isnstances that are correctly configured and a modest improvement to users' privacy on instances that aren't.
* We've added another extension to our customized markdown renderer which replaces markdown images with a warning that CryptPad blocks remote content to prevent malicious users from tracking visitors to certain pages. Such images should already be blocked by our strict use of Content-Security-Policy headers, but this will provide a better indication why images are failing to load on isnstances that are correctly configured and a modest improvement to users' privacy on instances that aren't.
* Up until now it was possible to include style tags in markdown documents, which some of our more advanced users used in order to customize the appearance of their rendered documents. Unfortunately, these styles were not applied strictly to the markdown preview window, but to the page as a whole, making it possible to break the platform's interface (for that pad) through the use of overly broad and powerful style rules. As of this release style tags are now treated as special elements, such that their contents are compiled as [LESS](https://lesscss.org/) within a scope that is only applied to the preview pane. This was intended as a bug fix, but it's included here as a _feature_ because advanced users might see it as such and use it to do neat things. We have no funding for further work in this direction, however, and presently have no intent of providing documentation about this behaviour.
* Up until now it was possible to include style tags in markdown documents, which some of our more advanced users used in order to customize the appearance of their rendered documents. Unfortunately, these styles were not applied strictly to the markdown preview window, but to the page as a whole, making it possible to break the platform's interface (for that pad) through the use of overly broad and powerful style rules. As of this release style tags are now treated as special elements, such that their contents are compiled as [LESS](https://lesscss.org/) within a scope that is only applied to the preview pane. This was intended as a bug fix, but it's included here as a _feature_ because advanced users might see it as such and use it to do neat things. We have no funding for further work in this direction, however, and presently have no intent of providing documentation about this behaviour.
* The checkup page uses some slightly nicer methods of displaying values returned by tests when the expected value of `true` is not returned. Some tests have been revised to return the problematic value instead of `false` when the test fails, since there were some cases where it was not clear why the test was failing, such as when a header was present but duplicated.
* The checkup page uses some slightly nicer methods of displaying values returned by tests when the expected value of `true` is not returned. Some tests have been revised to return the problematic value instead of `false` when the test fails, since there were some cases where it was not clear why the test was failing, such as when a header was present but duplicated.
ansuz
3 years ago