1 | | A proper WXR importer would be a good first step here, and to build one we need to have an XML_Tag_Processor in core just like we have HTML_Tag_Processor. Brownie points if it supports stream processing. WXR carries post data where site URLs need to be rewritten in the block attributes and inside the HTML markup. Some of these posts may be larger than the available memory size, hence the need for stream processing. All of that also applies to the full site transfer. |
| 1 | A proper WXR importer would be a good first step here. Brownie points if it supports stream processing. WXR carries post data where site URLs need to be rewritten in the block attributes and inside the HTML markup. Some of these posts may be larger than the available memory size, hence the need for stream processing. All of that also applies to the full site transfer. |
| 2 | |
| 3 | The only blocker I can see for that WXR importer is lack of an XML_Tag_Processor in core. We do have HTML_Tag_Processor, but it can't be used to process XML documents. Luckily, @dmsnell is working on the XML machinery. Once it is available, it will unlock a lot of exploratory avenues here. |