Package textbuffer provides a way to buffer text data based on a number of writes. Sometimes there is a need to use a buffer with predictable split criteria so that data stays logically solid even if only a single chunk gets its way on the final target writer. Splitting text buffers by size is opposed to such behavior as text lines cut an unpredictable way. An example of such need is maintaining a log lines buffer before writing into in a rotated file every number of writes. This package is safe for parallel execution.