curl / Mailing Lists / curl-library / Single Mail
Buy commercial curl support from WolfSSL. We help you work out your issues, debug your libcurl applications, use the API, port to new platforms, add new features and more. With a team lead by the curl founder himself.

Re: noproxy breakage

From: Timothe Litt <litt_at_acm.org>
Date: Thu, 27 Oct 2022 18:07:57 -0400


On 27-Oct-22 17:42, Randall via curl-library wrote:
> On October 27, 2022 5:30 PM, Daniel wrote:
>> A regression in the noproxy filter functionality in 7.86.0 has been
> suggested to be
>> reason enough for a patch release:
>>
>> https://github.com/curl/curl/issues/9813
>>
>> We don't yet have any stated policy for how to judge when a bug is reason
>> enough for a patch release but maybe this is the time to try to forumlate a
>> guideline?
> As a starting point, a common set of guidelines for justifying patch
> releases include the following non-conflicting requirements:
>
> 1. Is there a high rated security advisory for which a fix is available?
> 2. Is there a data corruption during processing for which a fix is
> available?
> 3. Is there a stack corruption that causes the software to prematurely
> terminate for which a fix is available?
> 4. Does the fix not change any API, which would exclude a patch-level
> release.
> 5. Is the fix of general applicability instead of one environment, OS
> release, or platform?
> 6. Is there a malware attack vector that can be closed with an available
> patch?
>
> Just my off-the-cuff dump of what I have used on other projects for patch
> release justification.
>
> -Randall
>
This is a good list.  Some additional considerations:

  * What is the scope/risk of the patch?
  * What else is pending?  Sometimes it's worth waiting for more
    experience (with the release and the patch) to avoid multiple patch
    releases, especially close together.
  * What is the best delivery method - a patch or a full release?
  * Is there a work-around that has acceptable impact to the users?
    (now, and when the fix is released)
  * Is this a regression, or a newly-discovered issue?  The hurdle for
    the latter triggering a release is higher (but not infinite).
  * Is the patch a work-around, or the final solution?
  * What's impact on the project? (How many reports?  Is it blocking
    adoption of the release?)
  * How many users are actually impacted?  Are any critical to the
    project, or severely impacted?

Attempts to have a mechanical scorecard for a decision are rarely
successful, but thinking about the issues is worthwhile.

Timothe Litt
ACM Distinguished Engineer
--------------------------
This communication may not represent the ACM or my employer's views,
if any, on the matters discussed.


-- 
Unsubscribe: https://lists.haxx.se/listinfo/curl-library
Etiquette:   https://curl.se/mail/etiquette.html
Received on 2022-10-28