[Koha-bugs] [Bug 8437] Large database backups and large exports from export.pl fail under plack

[bugzilla-daemon at bugs.koha-community.org]

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=8437

--- Comment #34 from David Cook <dcook at prosentient.com.au> ---
(In reply to David Cook from comment #33)
> As Dobrica mentioned, it does appear that Starman is buffering the whole
> response before sending it. I am guessing this is due to our usage of
> Plack::App::CGIBin... but I'll investigate further.

If I understand correctly, this will never work for us, so long as we're using
Plack::App::CGIBin to serve export.pl. 

At https://metacpan.org/release/Plack/source/lib/Plack/App/CGIBin.pm#L47,
Plack::App::CGIBin uses Plack::App::WrapCGI. 

Plack::App::WrapCGI uses CGI::Emulate::PSGI at
https://metacpan.org/source/Plack::App::WrapCGI#L87.

CGI::Emulate::PSGI writes the response to a temporary file before returning the
response (as per
https://metacpan.org/release/CGI-Emulate-PSGI/source/lib/CGI/Emulate/PSGI.pm#L19).

If you look at CGI::Emulate::PSGI::Streaming
(https://metacpan.org/release/CGI-Emulate-PSGI-Streaming/source/lib/CGI/Emulate/PSGI/Streaming.pm),
you can see how that handler returns a closure rather than a tuple (as noted at
http://www.catalystframework.org/calendar/2013/10).

In order to do a streaming response, you need to leverage that $responder
coderef in the closure.

Of course, this should be doable. I have an idea germinating.

-- 
You are receiving this mail because:
You are watching all bug changes.