Nicolas R.

There is pending merge request for this issue via #43 but this is the original ticket, keeping it opened until we have not fixed it

No problems @demerphq I will revisit it an extra time :-) and submit and updated PR

At first glance this seems to be by design, look how the BLOCK directive is implemented template($block) - https://github.com/abw/Template2/blob/master/lib/Template/Directive.pm#L69 anon_block($block) - https://github.com/abw/Template2/blob/master/lib/Template/Directive.pm#L100

this is another leak like #152 we probably want to use the same technique as #201 to detect & fix it

1a0ce8c06d07498bf6a8d4b3a20ee61de0d73210 is adding a test to detect that leak and is catching a leak from a Regexp... ``` > RELEASE_TESTING=1 perl -Ilib t/zz-process-leak.t 1..4 # Searching for leak using Test::LeakTrace......

The leak detected by Test::LeakTrace comes from the singleton factory object stored in Template::Grammar Using the script by itself I cannot notice a memory leak using Stash::XS or Stash ```perl...

sounds a good idea @petdance would you like to submit a patch for these changes?

I think the website is generated from the pods from this repo the sentence you were pointing at for example is here: https://github.com/abw/Template2/blob/master/lib/Template/Manual/Config.pod#L202 We would then have to refresh the...

Why not using a different keyword like `copy_dir`? This way we would reduce the risk of introducing a regression on existing usages of `copy` and make the intent clear, that...

It could be nice to use Wallflower to generate a static version of the site. Using a GitHub workflow we could then use GitHub pages to host the static version...