Buffers for yielding to Django are now 8KB and sizes are tracked by BytesIO/StringIO instead of a separate counter. That speeds up sending GML3.2 data by some 13%.
Trimming away whitespace decreases the size of payloads by some 9%, assuming no HTTP compression.
Buffers for yielding to Django are now 8KB and sizes are tracked by
BytesIO
/StringIO
instead of a separate counter. That speeds up sending GML3.2 data by some 13%.Trimming away whitespace decreases the size of payloads by some 9%, assuming no HTTP compression.
See commit messages for more details.