Uniform the quoting criteria when dealing with transcripts: - transcripts are stored as they are, no quoting in the data we store (and therefore HTML should be avoided) - transcripts are displayed quoted, therefore disabling any effect from HTML eventually present in the content - bring the logics to build transcript content in the api to avoid duplication - avoid sending the transcript content to transcript-new-2, as we can generate it server side
Note: storing transcripts as a pre-formatted chunk of text is probably not the best approach, as any changes to the formatting cannot be applied to pre-existing data without a delicate upgrade. This new approach will make <br> tags inside older transcripts visible...
Chat reform: - use consistently package_id to model belonging of a chat object to a particular package instead of the context_id - set context_id to the package_id, as was already the case in practice, for permission checking - make sure package_id exists and set it to the context_id in case it doesn't to handle legacy installations