Change Details

It makes little sense to start some index computations on non text content. So the orchestrator idea must be pushed a little further. Having a text content orchestrator is that push. So the idea is that once the mimetype indexing is done, send text contents to that orchestrator. Then the content orchestrator works as the initial orchestrator (cf. T576). Adapt the current indexer to do that.

It makes little sense to start some index computations on non text content. So the orchestrator idea must be pushed a little further. Once a content mimetype indexing is done, we can filter the textual ones and send those to the 'text orchestrator'. Then the content orchestrator works as the initial orchestrator (cf. T576) but only with sha1s whose content is text. Adapt the current indexer to do that.

It makes little sense to start some index computations on non text content. So the orchestrator idea must be pushed a little further. Once a content mimetype indexing is done, Having awe can filter the text content orchestrator is that pushual ones and send those to the 'text orchestrator'. So the idea is that once the mimetype indexing is done, send text contents to that orchestrator. Then the content orchestrator works as the initial orchestrator (cf. T576) but only with sha1s whose content is text. Adapt the current indexer to do that.