In 1.3.1 (and previous 1.3 versions) the bookmarkable URLs are all generated
with trailing /. Early 1.3 betas did not have this.
While this encodes and decodes correctly within the framework, it tends to
generate URLs that look funny and causes wackiness with
QueryStringUrlCodingStrategy where the URL has a /? before all the
parameters. I am trying to mount a page to a filename with an extension so
that the browser will recognize this (having a .pdf generated), however the
/ is making it look like a directory instead of a filename.
Is the trailing / necessary?
It would seem that this can be removed simply by fixing line 136 of
AbstractRequestTargetUrlCodingStrategy:
url.append(key).append("/").append(escapedValue).append("/");
to remove the final .append("/").
Any reason why this should not be done?
Mark.
