Bug #105651 for Apache-Template: Content Length and UTF-8 Characters

Fri Jul 03 14:53:25 2015 troymore [...] nbnet.nb.ca - Ticket created

Subject:	Content Length and UTF-8 Characters
Date:	Fri, 3 Jul 2015 15:53:15 -0300
To:	<bug-Apache-Template [...] rt.cpan.org>
From:	"Troy Morehouse" <troymore [...] nbnet.nb.ca>

It appears that when setting: TT2Headers length And the resulting template contains UTF8 characters that have an ordinal value greater than 255, the resulting content gets truncated by intermediate proxies and or the client. On Chrome, the console dhows there error: <http://stackoverflow.com/questions/29894154/chrome-neterr-incomplete-chunke d-encoding-error> net::ERR_INCOMPLETE_CHUNKED_ENCODING By Setting: TT2Headers none appears to fix the problem. In Template::Apache::Service, I see that the length of the content is calculated by the perl "length" function. This function computes character length, rather than byte length. A simple fix would be: use Encode (); $r->headers_out->add('Content-Length' => length(Encode::encode_utf8($$content))) if $all or $headers->{ length };

Download winmail.dat
application/ms-tnef 4.9k

Message body not shown because it is not plain text.

Fri Jul 03 16:32:35 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-07-03 14:53:25, troymore@nbnet.nb.ca wrote: Show quoted text

> It appears that when setting: > > > > TT2Headers length > > > > And the resulting template contains UTF8 characters that have an ordinal > value greater than 255, the resulting content gets truncated by > intermediate proxies and or the client. On Chrome, the console dhows there > error: > > > > > > <http://stackoverflow.com/questions/29894154/chrome-neterr-incomplete-chunke > d-encoding-error> net::ERR_INCOMPLETE_CHUNKED_ENCODING > > > > > By Setting: > > > > TT2Headers none > > > > appears to fix the problem. > > > > In Template::Apache::Service, I see that the length of the content is > calculated by the perl "length" function. This function computes character > length, rather than byte length. A simple fix would be: > > > > use Encode (); > > > > $r->headers_out->add('Content-Length' => > length(Encode::encode_utf8($$content))) > > if $all or $headers->{ length }; >

This will be wrong if $$content contains characters between 128 and 255 and no utf8 flag is set.

Fri Jul 03 16:32:35 2015 The RT System itself - Status changed from 'new' to 'open'

Fri Jul 03 16:44:59 2015 troymore [...] nbnet.nb.ca - Correspondence added

Subject:	RE: [rt.cpan.org #105651] Content Length and UTF-8 Characters
Date:	Fri, 3 Jul 2015 17:45:05 -0300
To:	<bug-Apache-Template [...] rt.cpan.org>
From:	"Troy Morehouse" <troymore [...] nbnet.nb.ca>

I suppose one could do this then: use utf8; use Encode (); $r->headers_out->add('Content-Length' => length(utf8::is_itf8($$content) ? Encode::encode_utf8($$content) : $$content) ) if $all or $headers->{ length }; Or { use bytes; $r->headers_out->add('Content-Length' => length($$content)) if $all or $headers->{ length }; } Or something to that nature (i.e. bytes::length(...)) Show quoted text

-----Original Message----- From: Slaven_Rezic via RT [mailto:bug-Apache-Template@rt.cpan.org] Sent: July 3, 2015 5:33 PM To: troymore@nbnet.nb.ca Subject: [rt.cpan.org #105651] Content Length and UTF-8 Characters <URL: https://rt.cpan.org/Ticket/Display.html?id=105651 > On 2015-07-03 14:53:25, troymore@nbnet.nb.ca wrote:

> It appears that when setting: > > > > TT2Headers length > > > > And the resulting template contains UTF8 characters that have an ordinal > value greater than 255, the resulting content gets truncated by > intermediate proxies and or the client. On Chrome, the console dhows there > error: > > > > > > <http://stackoverflow.com/questions/29894154/chrome-neterr-incomplete-chunke > d-encoding-error> net::ERR_INCOMPLETE_CHUNKED_ENCODING > > > > > By Setting: > > > > TT2Headers none > > > > appears to fix the problem. > > > > In Template::Apache::Service, I see that the length of the content is > calculated by the perl "length" function. This function computes character > length, rather than byte length. A simple fix would be: > > > > use Encode (); > > > > $r->headers_out->add('Content-Length' => > length(Encode::encode_utf8($$content))) > > if $all or $headers->{ length }; >

This will be wrong if $$content contains characters between 128 and 255 and no utf8 flag is set.