Обсуждение: Re: [HACKERS] [COMMITTERS] pgsql: Tag refs/tags/REL_10_BETA1 was created
Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 5/16/17 18:14, pgsql@postgresql.org wrote:
>> Tag refs/tags/REL_10_BETA1 was created.
> Was this change in naming pattern intentional?
Yes, it was. Andrew Dunstan suggested[1] during the
two-part-version-number discussion that we should start including a "_"
after REL in tag and branch names for v10 and later, so that those names
would sort correctly compared to the tag/branch names for earlier branches
(at least when using C locale). I believe his main concern was some logic
in the buildfarm, but it seems like a good idea in general.
When we get to v100, we'll need some other hack to make it work ...
but I plan to be safely dead by then.
BTW, I now remember having wondered[2] if we should make any other changes
in version-number formatting while we're at it, like maybe "10beta1"
should be "10.beta1". It's a bit late to have remembered it for beta1,
but is anyone hot to change anything else about these labels?
regards, tom lane
[1] https://www.postgresql.org/message-id/57364C11.4040004@dunslane.net
[2] https://www.postgresql.org/message-id/20780.1463176901%40sss.pgh.pa.us
On 05/16/2017 10:37 PM, Tom Lane wrote: > Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes: >> On 5/16/17 18:14, pgsql@postgresql.org wrote: >>> Tag refs/tags/REL_10_BETA1 was created. >> Was this change in naming pattern intentional? > Yes, it was. Andrew Dunstan suggested[1] during the > two-part-version-number discussion that we should start including a "_" > after REL in tag and branch names for v10 and later, so that those names > would sort correctly compared to the tag/branch names for earlier branches > (at least when using C locale). I believe his main concern was some logic > in the buildfarm, but it seems like a good idea in general. > > When we get to v100, we'll need some other hack to make it work ... > but I plan to be safely dead by then. > Me too. Since posterity will be deprived of both of us let's note that the same hack will work, we'll just need two underscores. cheers andrew -- Andrew Dunstan https://www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On 5/16/17 22:37, Tom Lane wrote: > BTW, I now remember having wondered[2] if we should make any other changes > in version-number formatting while we're at it, like maybe "10beta1" > should be "10.beta1". That's not a naming format I've ever seen. I think the current format is fine. -- Peter Eisentraut http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On Wed, May 17, 2017 at 2:41 PM, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
On 5/16/17 22:37, Tom Lane wrote:
> BTW, I now remember having wondered[2] if we should make any other changes
> in version-number formatting while we're at it, like maybe "10beta1"
> should be "10.beta1".
That's not a naming format I've ever seen.
I think the current format is fine.
+1. I have also never seen that one, and think the current one is good.
On Wed, May 17, 2017 at 8:04 AM, Andrew Dunstan <andrew.dunstan@2ndquadrant.com> wrote: >> When we get to v100, we'll need some other hack to make it work ... >> but I plan to be safely dead by then. > > Me too. Since posterity will be deprived of both of us let's note that > the same hack will work, we'll just need two underscores. That cure sounds worse than the disease, but I guess we can leave the decision to posterity. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company