Re: Beanshell and BSF performance
James Pharaoh <[email protected]> Fri, 26 Feb 2010 11:53:29 +0000
| Newsgroups | gmane.comp.jakarta.jmeter.devel |
|---|---|
| Message-ID | <[email protected]> |
2010/2/26 sebb <[email protected]>: > On 26/02/2010, James Pharaoh <[email protected]> wrote: >> I notice both these scripting facilities seem to recompile the script >> every time they are used. I'm pretty sure this shouldn't be necessary >> and that the performance could be increased significantly by compiling >> once. Does anyone know any good reason this is not done? > Because one would have to keep track of whether the script changed > during the test run, e.g. because the GUI used variable or function > references. Or indeed another script might change the script file > (unlikely, but there might be a use case for this). So in this case you could store scripts based on the hash code of the string, or the timestamp and/or hash code of the file. There could even be an option to always recompile, always reuse or always check. James