Re: PEAR package retention policy
| From: | Helgi Þormar Þorbjörnsson | Date: | Thu, 05 Apr 2007 17:26:59 +0000 |
| Subject: | Re: PEAR package retention policy | ||
| References: | 1 | Groups: | php.pear.dev |
| Request: | Send a blank email to pear-dev+get-46145@lists.php.net to get a copy of this message | ||
Sounds good to me, thumbs up.
- Helgi
On Thu, 05 Apr 2007 00:44:14 +0400, Alexey Borzov wrote:
> Hi,
>
> As suggested by Greg in "Importing packages into CVS" thread, I'm
> posting a list of changes that should be done to PEAR guidelines and
> pearweb.
>
> == Problem ==
>
> Currently PEAR website allows package maintainers to delete packages and
> package releases without any limitations whatsoever. This means that a
> maintainer (or someone having his password) may delete a
> several-years-old package with dozens of releases or delete any release
> of said package.
>
> Such actions will affect all packages or applications depending on the
> deleted package or requiring a specific deleted release of the package.
>
> == Proposed solution ==
>
> In a nutshell, the proposed policy can be described as "Whatever was
> released though pear.php.net, stays released".
>
> The ability to delete a package release shall be limited to releases not
> older than two weeks (*). Package maintainers are advised to use this
> feature in the following cases only:
> - Broken package definition file
> - Missing required dependencies
> - Wrong dependencies
> - Missing files
> - etc.
> - Newly introduced severe bugs in stable releases (**) - Severe
> violations of PEAR's coding standards and regulations
> - Incorrect names in public API
> - Incorrect package version
>
> The ability to delete a package shall be limited to packages with no
> releases.
>
> = On "moved" packages =
>
> As some developers may wish to move their development away from PEAR's
> infrastructure, there should be a means to announce this move (as
> opposed to declaring their packages unsupported), closing the
> bug-tracker and giving a prominent link to a new location.
>
>
> (*) The actual period is discussable. (**) Releases of stability level
> lower than "stable" are done for testing purposes and are expected to
> contain bugs.
--
There are no stupid questions, only stupid people