You can clone with
HTTPS or Subversion.
Originally on 2011-09-08
It would probably be better for this to work like the LaTeX CV format, which
can display over 1,000 records at a time. Certainly, a limit of 1,000 would be
more suitable than the 100 limit as the point of the bibtex format is not to
read the result but cut-and-paste to a file.
The limit is inside search_engine.py, but the effect is seen in the output format. There is machinery in search_engine meant to raise (remove?) this limit for xml output formats; perhaps that is helpful here.