[Libwebsockets] Latest version

Srikar Vadlamani SrikarVadlamani at trafficware.com
Mon Feb 1 17:25:18 CET 2016

The lws_http_serve_file crash with libev is fixed, I just verified it. 

What does "lws_issue_raw: wsi 0x100ae700: LWS_SSL_CAPABLE_ERROR" signify?

My release cycles are much longer and most of the libwebsockets tests are done as part of application, hence it takes longer.

Your idea of stable branch and live fixes if good or for non-main stream features as you put it, may be an email notice for all users to do some sanity testing before each release.
That would weed out obvious bugs right after a release.


-----Original Message-----
From: Andy Green [mailto:andy.green at linaro.org] On Behalf Of Andy Green
Sent: Monday, February 01, 2016 10:08 AM
To: Srikar Vadlamani <SrikarVadlamani at trafficware.com>; libwebsockets at ml.libwebsockets.org
Subject: RE: [Libwebsockets] Latest version

On February 1, 2016 10:16:04 PM GMT+08:00, Srikar Vadlamani <SrikarVadlamani at trafficware.com> wrote:
>When is 1.7.0 coming?

Today it seems we should wait for the libuv port, if that doesn't take too long.  Libev / uv and multithreaded service needs testing.

Otherwise it should be good to go afaik.

I know some amount of people use master from bug reports. But unless there's some guy(s) actively testing it and shaking the tree and communicating, it's difficult to judge from silence how ready it is.  Silence sometimes means everything working and sometimes only that nobody hit the problem yet or hit the problem, shrugged and went back to 1.6 and got on with their life.  (For example your email is not giving me any info about your testing of it, just "tell me when you will tag something").

From my own testing HEAD continues to pass Coverity, Travis, Appveyor, Autobahn, attack.sh and the usual test server and client.  From users recently I know it's generally working on Windows and I'm waiting to hear about Codenomicon from the guy who has that.

Looking at 1.6 he similarly worked quite well but went out with bugs on non-mainstream features like uri argument processing and libev.  They were caught quite quickly after release but it seems many people only try it *after* release.  Maybe the answer is do our best and then use the recently-introduced stable branches to get ported fixes in realtime so they tend to get more perfect subsequently.


>-----Original Message-----
>From: Andy Green [mailto:andy.green at linaro.org] On Behalf Of Andy Green
>Sent: Sunday, January 31, 2016 4:24 AM
>To: Roger Light <roger at atchoo.org>; libwebsockets at ml.libwebsockets.org
>Subject: Re: [Libwebsockets] Latest version
>On 01/31/2016 06:02 PM, Roger Light wrote:
>> Hi Andy,
>> The website lists 1.6.0 as the latest version, you might want to
>update it.
>I updated it but with 1.7.0 coming it needs a bit more of an update to 
>explain the s6able branches, I'll do that at the time.
>> Cheers,
>> Roger
>> _______________________________________________
>> Libwebsockets mailing list
>> Libwebsockets at ml.libwebsockets.org
>> http://ml.libwebsockets.org/mailman/listinfo/libwebsockets
>Libwebsockets mailing list
>Libwebsockets at ml.libwebsockets.org

More information about the Libwebsockets mailing list