You are viewing a plain text version of this content. The canonical link for it is here.
Posted to rivet-dev@tcl.apache.org by Massimo Manghi <mx...@apache.org> on 2022/06/09 10:37:12 UTC
Rivet 3.2.2rc1
Rivet 3.2.2 rc1 is available from
http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
New features and bug fixes
+ a memory leak was fixed in Rivet_Var (src/mod_rivet_ng/rivetCore.c)
+ Installation and TEA scripts were updated
+ The form [::rivet::var all] accepts an optional flat list argument
of key-value pairs which are defaults for an application query variables
(yes, this argument must be interpreted as a dictionary). This form
generalizes the [::rivet::var get <varname> <default>] form.
+ Tests for this new form were added and also ::rivet::var tests in
general were improved
+ Configuration directive HonorHeaderOnlyReqs, though still supported,
was renamed as HonorHeadRequests. Speaking the HTTP related parlance I
think this a better wording of this directive (suggestions? Like
dropping support for this directive and honor HTTP head requests every
time?)
-- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org
Re: Rivet 3.2.2rc1
Posted by Brice Hamon <no...@gmail.com>.
Thanks Massimo.
Will test it asap.
Thanks for your support.
Brice.
On Thu, Jun 9, 2022 at 6:44 AM Massimo Manghi <mx...@apache.org> wrote:
> Rivet 3.2.2 rc1 is available from
>
> http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
>
> New features and bug fixes
>
> + a memory leak was fixed in Rivet_Var (src/mod_rivet_ng/rivetCore.c)
>
> + Installation and TEA scripts were updated
>
> + The form [::rivet::var all] accepts an optional flat list argument
> of key-value pairs which are defaults for an application query variables
> (yes, this argument must be interpreted as a dictionary). This form
> generalizes the [::rivet::var get <varname> <default>] form.
>
> + Tests for this new form were added and also ::rivet::var tests in
> general were improved
>
> + Configuration directive HonorHeaderOnlyReqs, though still supported,
> was renamed as HonorHeadRequests. Speaking the HTTP related parlance I
> think this a better wording of this directive (suggestions? Like
> dropping support for this directive and honor HTTP head requests every
> time?)
>
> -- Massimo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
> For additional commands, e-mail: rivet-dev-help@tcl.apache.org
>
>
Re: Rivet 3.2.2rc1
Posted by Massimo Manghi <mx...@apache.org>.
On 09/06/22 15:40, Harald Oehlmann wrote:
>>
>>
>> The switch has been around ever since and it's only function is to
>> force mod_rivet to return a "text/html" message avoiding actual script
>> execution.
>>
>> if (private->r->header_only &&
>> !private->running_conf->honor_head_requests)
>> {
>> TclWeb_SetHeaderType(DEFAULT_HEADER_TYPE, private->req);
>> TclWeb_PrintHeaders(private->req);
>> retval = OK;
>> goto sendcleanup;
>> }
>>
>>
>> Thus no meaningful information is returned if HonorHeadRequests is
>> false (default)
>>
>> -- Massimo
>
> I would call that a bug. Head should be handled correctly.
>
> Thank you for caring,
> Harald
>
>
In hindsight it's probably so. We are thinking quite alike, before I
read your message I was considering to open a bug in order to document
either the removal or the change of the default for boolean
honor_head_request
I remember the configuration directive that allowed to have real HEAD
responses was one of my earliest contributions to mod_rivet in 2007 or
2008. I needed to have meaningful data from a HEAD request but as a
newcomer I didn't want to change abruptly the original behavior of the
code in branch 1_0 (never released) assuming there was some good reason
for it. I have some faint memory that it had been chosen to shunt code
execution for HEAD requests for performance reasons
more thoughts?
-- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org
Re: Rivet 3.2.2rc1
Posted by Massimo Manghi <mx...@apache.org>.
I prepared a new tar archive available from
http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
As per Harald corrent remark, I'm assuming now that refusing to serve
HEAD HTTP requests is in principle wrong. I made the transitional step
of changing HonorHeadRequests default to 'true' while we consider to
drop support for this switch in future major releases
-- Massimo
On 09/06/22 15:40, Harald Oehlmann wrote:
> Am 09.06.2022 um 15:37 schrieb Massimo Manghi:
> [...]
> I would call that a bug. Head should be handled correctly.
>
> Thank you for caring,
> Harald
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org
Re: Rivet 3.2.2rc1
Posted by Harald Oehlmann <ha...@elmicron.de>.
Am 09.06.2022 um 15:37 schrieb Massimo Manghi:
> On 09/06/22 14:00, Harald Oehlmann wrote:
>>
>> Am 09.06.2022 um 12:37 schrieb Massimo Manghi:
>>> Rivet 3.2.2 rc1 is available from
>>>
>>> http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
>>>
>>> New features and bug fixes
>>>
>>> + a memory leak was fixed in Rivet_Var (src/mod_rivet_ng/rivetCore.c)
>>>
>>> + Installation and TEA scripts were updated
>>>
>>> + The form [::rivet::var all] accepts an optional flat list
>>> argument of key-value pairs which are defaults for an application
>>> query variables (yes, this argument must be interpreted as a
>>> dictionary). This form generalizes the [::rivet::var get <varname>
>>> <default>] form.
>>>
>>> + Tests for this new form were added and also ::rivet::var tests in
>>> general were improved
>>>
>>> + Configuration directive HonorHeaderOnlyReqs, though still
>>> supported, was renamed as HonorHeadRequests. Speaking the HTTP
>>> related parlance I think this a better wording of this directive
>>> (suggestions? Like dropping support for this directive and honor HTTP
>>> head requests every time?)
>>
>> Dear Massimo,
>>
>> thank you for the great and detail work.
>> I am in favor of all those changes.
>>
>> My thought about the head request: head requests are standard for http
>> since the beginning. I would just honor them always. They are part of
>> http and should be supported in a standard way.
>>
>> Thank you,
>> Harald
>>
>
>
> The switch has been around ever since and it's only function is to force
> mod_rivet to return a "text/html" message avoiding actual script execution.
>
> if (private->r->header_only &&
> !private->running_conf->honor_head_requests)
> {
> TclWeb_SetHeaderType(DEFAULT_HEADER_TYPE, private->req);
> TclWeb_PrintHeaders(private->req);
> retval = OK;
> goto sendcleanup;
> }
>
>
> Thus no meaningful information is returned if HonorHeadRequests is false
> (default)
>
> -- Massimo
I would call that a bug. Head should be handled correctly.
Thank you for caring,
Harald
--
ELMICRON Dr. Harald Oehlmann GmbH
Koesener Str. 85
06618 NAUMBURG - Germany
Phone: +49 3445 781120
Direct: +49 3445 781127
www.Elmicron.de
German legal references:
Geschaeftsfuehrer: Dr. Harald Oehlmann
UST Nr. / VAT ID No.: DE206105272
HRB 212803 Stendal
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org
Re: Rivet 3.2.2rc1
Posted by Massimo Manghi <mx...@apache.org>.
On 09/06/22 14:00, Harald Oehlmann wrote:
>
> Am 09.06.2022 um 12:37 schrieb Massimo Manghi:
>> Rivet 3.2.2 rc1 is available from
>>
>> http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
>>
>> New features and bug fixes
>>
>> + a memory leak was fixed in Rivet_Var (src/mod_rivet_ng/rivetCore.c)
>>
>> + Installation and TEA scripts were updated
>>
>> + The form [::rivet::var all] accepts an optional flat list argument
>> of key-value pairs which are defaults for an application query
>> variables (yes, this argument must be interpreted as a dictionary).
>> This form generalizes the [::rivet::var get <varname> <default>] form.
>>
>> + Tests for this new form were added and also ::rivet::var tests in
>> general were improved
>>
>> + Configuration directive HonorHeaderOnlyReqs, though still
>> supported, was renamed as HonorHeadRequests. Speaking the HTTP related
>> parlance I think this a better wording of this directive (suggestions?
>> Like dropping support for this directive and honor HTTP head requests
>> every time?)
>
> Dear Massimo,
>
> thank you for the great and detail work.
> I am in favor of all those changes.
>
> My thought about the head request: head requests are standard for http
> since the beginning. I would just honor them always. They are part of
> http and should be supported in a standard way.
>
> Thank you,
> Harald
>
The switch has been around ever since and it's only function is to force
mod_rivet to return a "text/html" message avoiding actual script execution.
if (private->r->header_only &&
!private->running_conf->honor_head_requests)
{
TclWeb_SetHeaderType(DEFAULT_HEADER_TYPE, private->req);
TclWeb_PrintHeaders(private->req);
retval = OK;
goto sendcleanup;
}
Thus no meaningful information is returned if HonorHeadRequests is false
(default)
-- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org
Re: Rivet 3.2.2rc1
Posted by Harald Oehlmann <ha...@elmicron.de>.
Am 09.06.2022 um 12:37 schrieb Massimo Manghi:
> Rivet 3.2.2 rc1 is available from
>
> http://www.rivetweb.org/~manghi/rivet/rivet-3.2.2rc1.tar.gz
>
> New features and bug fixes
>
> + a memory leak was fixed in Rivet_Var (src/mod_rivet_ng/rivetCore.c)
>
> + Installation and TEA scripts were updated
>
> + The form [::rivet::var all] accepts an optional flat list argument
> of key-value pairs which are defaults for an application query variables
> (yes, this argument must be interpreted as a dictionary). This form
> generalizes the [::rivet::var get <varname> <default>] form.
>
> + Tests for this new form were added and also ::rivet::var tests in
> general were improved
>
> + Configuration directive HonorHeaderOnlyReqs, though still supported,
> was renamed as HonorHeadRequests. Speaking the HTTP related parlance I
> think this a better wording of this directive (suggestions? Like
> dropping support for this directive and honor HTTP head requests every
> time?)
Dear Massimo,
thank you for the great and detail work.
I am in favor of all those changes.
My thought about the head request: head requests are standard for http
since the beginning. I would just honor them always. They are part of
http and should be supported in a standard way.
Thank you,
Harald
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: rivet-dev-help@tcl.apache.org