Re: make "HTTP/0.9" support opt-in ?

2018-07-02 Thread Daniel Stenberg
On Mon, 2 Jul 2018, Dennis Clarke wrote: It may be of value to create a curl/libcurl "policy" wherein the annual survey data may indicate the degree to which a feature is an opt-in or a default feature. In this way one would be able to say that there is less than a 5% usage rate for a given

Re: DEPRECATE.md

2018-07-02 Thread Rich Gray
Daniel Stenberg wrote: On Mon, 2 Jul 2018, Rich Gray wrote:    https://curl.haxx.se/dev/deprecate.html I wonder if for depreciation of build options like axTLS, you should arrange a big warning message should anyone try to build it.  This may get someone's attention when they might

Re: make "HTTP/0.9" support opt-in ?

2018-07-02 Thread Dennis Clarke
libcurl supports HTTP/0.9 by default, which might come as a surprise to users. Around 3% of users in the annual survey claim they use HTTP/0.9 with curl. It may be of value to create a curl/libcurl "policy" wherein the annual survey data may indicate the degree to which a feature is an

Re: DEPRECATE.md

2018-07-02 Thread Daniel Stenberg
On Mon, 2 Jul 2018, Rich Gray wrote:   https://curl.haxx.se/dev/deprecate.html I wonder if for depreciation of build options like axTLS, you should arrange a big warning message should anyone try to build it. This may get someone's attention when they might otherwise miss notification on

Re: DEPRECATE.md

2018-07-02 Thread Rich Gray
Daniel Stenberg wrote: FYI, I just added docs/DEPRECATE.md to the git repo, outlining the deprecation plans for axTLS and pipelining. This document is now available on the site here:   https://curl.haxx.se/dev/deprecate.html When these features have been deprecated, I think they should

make "HTTP/0.9" support opt-in ?

2018-07-02 Thread Daniel Stenberg
Hi, We have this bug [1] that shows a short "HTTP/0.9" response and how curl just then ignores the data it receives. HTTP/0.9 is the popular name for the never truly named HTTP version that existed before HTTP/1.0 was born. It has no response headers at all but instead it just sends data

Re: DEPRECATE.md

2018-07-02 Thread James Fuller
that is reasonable ... deprecate label is all I need. thx Jim On 2 July 2018 at 09:39, Daniel Stenberg wrote: > On Mon, 2 Jul 2018, James Fuller wrote: > >> Maybe we could raise an issue for each deprecated item (apologies for >> being a process nazi ... just want to be able to subscribe to

Re: DEPRECATE.md

2018-07-02 Thread Daniel Stenberg
On Mon, 2 Jul 2018, James Fuller wrote: Maybe we could raise an issue for each deprecated item (apologies for being a process nazi ... just want to be able to subscribe to them), Sure, but I think I prefer doing an issue/PR per step/action so that we don't need to have them linger around

Re: DEPRECATE.md

2018-07-02 Thread James Fuller
Maybe we could raise an issue for each deprecated item (apologies for being a process nazi ... just want to be able to subscribe to them), Jim On 2 July 2018 at 08:23, Daniel Stenberg wrote: > FYI, > > I just added docs/DEPRECATE.md to the git repo, outlining the deprecation > plans for axTLS

DEPRECATE.md

2018-07-02 Thread Daniel Stenberg
FYI, I just added docs/DEPRECATE.md to the git repo, outlining the deprecation plans for axTLS and pipelining. This document is now available on the site here: https://curl.haxx.se/dev/deprecate.html When these features have been deprecated, I think they should still be mentioned in this