Yeah, looks like I need to look at this problem for real. Your fix
should be fine - in fact, it'll work in all cases, it's just not as fast
as possible if a real send_fd can be done.
I'll get something working soon and release it.
[EMAIL PROTECTED] (JR Mayberry) wrote:
>I hack/fixed it by taking out the check to see if its the last module in a
>chain..
>
>For me, that is always a safe assumption..
>
>Let me know if theres ever a real fix..
>
>
>----- Original Message -----
>From: "Ken Williams" <[EMAIL PROTECTED]>
>To: "JR Mayberry" <[EMAIL PROTECTED]>
>Cc: <[EMAIL PROTECTED]>
>Sent: Friday, April 06, 2001 1:10 AM
>Subject: Re: Apache::Compress and Apache::Filter
>
>
>> Hi JR,
>>
>> I've been avoiding this bug to my peril. Does the following patch fix it?
>>
>> =================================================================
>> --- Filter.pm 2000/12/20 03:43:44 1.16
>> +++ Filter.pm 2001/04/06 05:05:24
>> @@ -120,8 +120,8 @@
>>
>> sub send_fd {
>> my $self = shift;
>> - if ($self->is_last_filter) {
>> + if ($self->is_last_filter and !&Universal::isa($_[0], ref $self)) {
>> $self->SUPER::send_fd(@_);
>> } else {
>> my $fd = shift;
>> =================================================================
>>
>> The bug occurs when the client doesn't support gzip, and you're using
>> Apache::Filter, and Apache::Compress is the final filter in the chain.
>>
>>
>> [EMAIL PROTECTED] (JR Mayberry) wrote:
>> >Does anyone know anything about the above combo, and getting an error
>> >message:
>> >Bad filehandle at Filter.pm line 123
>> >
>> >when using a client that doesnt support gzip..(specifically 'ab', apache
>> >bench)
>>
>>
>> ------------------- -------------------
>> Ken Williams Last Bastion of Euclidity
>> [EMAIL PROTECTED] The Math Forum
>
>
------------------- -------------------
Ken Williams Last Bastion of Euclidity
[EMAIL PROTECTED] The Math Forum