| 1 |
On Sat, Jul 10, 2004 at 12:26:58AM -0400, Bryan D. Stine wrote: |
| 2 |
> On Saturday July 10, 2004 12:04 am, Jeff Smelser wrote: |
| 3 |
> > Nope.. Again, I want to see this 30 day poiicy.. And I also want to know |
| 4 |
> > what this has to do with ebuilds out side kde, that are past 30 days.. |
| 5 |
> > Mysql, being one. |
| 6 |
> http://www.gentoo.org/doc/en/policy.xml#doc_chap4, the part about moving from |
| 7 |
> ~ARCH to ARCH. "An indication of the package's stability would be no verified |
| 8 |
> or unresolved bug report for a month after the version's introduction." There |
| 9 |
> may be other factors involved in stability, like dependencies and such. |
| 10 |
Furthermore: |
| 11 |
"It is up to the maintainer of the package to deem which versions are |
| 12 |
stable or if development versions should be in package.mask or left in |
| 13 |
~arch." |
| 14 |
|
| 15 |
As one of the maintainers of MySQL, amongst other things, my general |
| 16 |
policy is that I'd consider the month in ~arch as a baseline, depending |
| 17 |
BOTH the importance of the package, how much the package gets used, and |
| 18 |
also the scope of changes between versions (I'm the kind of guy that |
| 19 |
will read diffs between versions sometimes, to verify that bugs are |
| 20 |
fixed). |
| 21 |
|
| 22 |
As an example of the shorter times, there are some packages I maintain |
| 23 |
where I am the upstream developer, and it was a tool I wrote for my own |
| 24 |
daily usage that others find useful, and for those they usually languish |
| 25 |
for only two weeks in ~arch, BECAUSE I use them heavily every work day. |
| 26 |
Another example here is when I can see the diff is extremely tiny, |
| 27 |
and non-invasive in every way possible, or fixes an important/major |
| 28 |
bug in the last arch version. |
| 29 |
|
| 30 |
On the opposite side, I keep some of the packages I maintain in ~arch |
| 31 |
for a lot longer, because I'm not yet confident enough they are in a |
| 32 |
stable enough state to inflict upon users of arch. MySQL is a prime |
| 33 |
example of this, where 4.0.19 was partially a screwup upstream, with |
| 34 |
a nasty bug that affected users of InnoDB in production (yes, it bit |
| 35 |
me) so I wanted to be damn sure that 4.0.20 was an improvement over |
| 36 |
4.0.18. |
| 37 |
|
| 38 |
-- |
| 39 |
Robin Hugh Johnson |
| 40 |
E-Mail : robbat2@××××××××××××××.net |
| 41 |
Home Page : http://www.orbis-terrarum.net/?l=people.robbat2 |
| 42 |
ICQ# : 30269588 or 41961639 |
| 43 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |
Archives