W3C home > Mailing lists > Public > public-qa-dev@w3.org > November 2006

Re: Branches or not

From: olivier Thereaux <ot@w3.org>
Date: Thu, 16 Nov 2006 00:00:57 +0900
Message-Id: <7CCA5377-87B6-4BA1-98D8-C4BB660820F8@w3.org>
Cc: QA Dev <public-qa-dev@w3.org>
To: Sierk Bornemann <sierkb@gmx.de>

Hi Sierk,

On Nov 15, 2006, at 23:09 , Sierk Bornemann wrote:

> I have taken a CVS snapshot this early morning/night from  
> validator-0_7-branch, just right after the last CVS messages  
> reached me by mail, you said per mail, that all is done and the  
> website was updated. So, I thought, no more files will be changed  
> concerning 0.7.4.

And they weren't, there were no commits in the branch or on the tag,  
after that.

- mail hubs were slow today. I received some of my commit mails very  
late. It's possible you experienced that too.
- if you are using the pserver authentication, it is possible that  
the commits reached the (slave) pserver CVS server a while after they  
were sent to the master CVS server.

Either or both of the above may have caused the discrepancies you  
observed, but as I said, I am certain I did not make any change to  
the branch or tag after updating the online site (and I sent you mail  
after that).

What this seems to show is that:
* last minute changes, just before the release, should be avoided
* there are apparent delays for the tags and commits to be propagated.

So I propose that for the next releases, packages be made from the  
tarballs, which are the authoritative record of the release. I (or  
whoever takes care of the releases after me) will be sending mail to  
the list, saying that the code is stable and about to be released (or  
already released), and giving pointer to the tarballs.

That should save everyone a lot of trouble. What do you think?

Received on Wednesday, 15 November 2006 15:01:09 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:36:26 UTC