> -Original Messages-
> From: "Yann Ylavic"
> Sent Time: 2018-01-05 18:41:15 (Friday)
> To: httpd-dev
> Cc:
> Subject: Re: Consistent update of invoking ap_map_http_request_error()
>
> Hi,
>
> On Fri, Jan 5, 2018 at 4:56 AM, 牛旭 wrote:
> > Our team researches on the consistent edits of Httpd during
> > evolution. And we have figured out several spots that may be missed
> > for consistent update.
>
> I'm not sure what consistent means here, consistent with?
>
> > They are both about invoking of function,
> > ap_map_http_request_error(). We suggest to escape call of function
> > and return the corresponding error code directly when this call is
> > under the control dependence where return value of ap_get_brigade()
> > does not equal to APR_SUCCESS.
>
> Looking at the recommendations and patches, why do you the below
> change is equivalent or consitent:
> - return ap_map_http_request_error(status, HTTP_BAD_REQUEST);
> + return HTTP_BAD_REQUEST;
> ?
>
> The purpose of ap_map_http_request_error() is to turn a system error
> (errno, apr_status_t) to an HTTP error code (4xx, 5xx), *and* forward
> internal errors like AP_FILTER_ERROR to the core handler.
> The above change pretty much defeats that, the "status" returned by
> the preceding ap_get_brigade() may vary and
> ap_map_http_request_error() takes care of this.
>
In this way, i guess the patches that i have discovered is against your
original design purpose. Which means the following recommendations are also not
feasible. Anyway, thanks for your reply.
> >
> > More recommendations and supporting patches are saved in attachments.
> > It is so kind of you to reply me about the correctness of our
> > suggestions.
>
> Could you elaborate a bit more about these recommendations? For now I
> don't see which consistency is talked about, sorry.
>
>
> Regards,
> Yann.