We have a base template with something like: {% compress js %}{% block js %}{% endblock %}{% endcompress %}. Any number of templates might extend this base template, and define their own {% block js %}...{% endblock %} content.

If any of those templates includes a template tag or variable that depends on context that is not available during offline generation, OfflineGenerationError exceptions will be raised.

The error page when this happens only highlights the {% compress js %} line of the base template, as a hint as to where the error might be occurring.

Then we basically need to search the entire project for every occurrence of {% block js %} and audit its content, looking for a dynamic variable or template tag.

Looking at the traceback, in compressor.templatetags.compress.render_offline, just before the OfflineGenerationError exception is raised, we first generate the key from the original content with key = get_offline_hexdigest(self.get_original_content(context)).

I think we should pass an extended message to OfflineGenerationError which includes not only the key, but also the original content that was used to generate the key. This will provide a MUCH stronger hint as to where we might find the offending dynamic content.