Fw: [HEADS-UP] QA BotMails fully automated

[Ion-Mihai Tetcu]

Begin forwarded message:

Date: Thu, 14 Aug 2008 21:52:20 +0300
From: Ion-Mihai Tetcu <itetcu at FreeBSD.org>
To: Ion-Mihai Tetcu <itetcu at FreeBSD.org>
Cc: ports-developers at FreeBSD.org
Subject: Re: [HEADS-UP] QA BotMails fully automated


On Thu, 14 Aug 2008 13:32:26 +0300
Ion-Mihai Tetcu <itetcu at FreeBSD.org> wrote:

> Hi,
> 
> 
> being poked by Edwin this morning, I switched on automated
> commit-triggered BotMails. So blame him! ;-)
> 
> This means I don't get to check each mail first, so there may be false
> positives, of which I had only 6 so far, 3 of them being for fetch
> failures.
> 
> Also please, please drop me an email some 24 hours before you commit
> sweeping changes or changes to MK/*, etc. else you'll probably end up
> with your inbox filled with BotMails and also the lists will get
> spammed.
> (QA Tindy processes the commits in the order they appear in the commit
> mail not in dependency order so on big related commits transient
> errors might pop up; this will be fixed once I'll switch to tinderbox
> v.3 and implement queue scheduling/reordering in dependency order).
> 
> Please bounce back to me any mail you think it's a false positive or
> such (and please make sure you include the original mail's X-Mailer:
> header so that I know if the problem is already fixed in my scripts or
> not).
> 
> If things really run wild ping me on IRC (itetcu, ITetcu or IonBot on
> freenode and efnet).
> 
> 
> In other news:
> 
> Maintainers are now CC'ed on all BotMails.
> 
> For now, there are no exceptions (people who want their BotMails not
> being CC'ed to the lists); please drop me an email if you are in this
> category and I'll try to re-implement this "feature".
> 
> I'm BCC'ed to keep an aye on things while Mark is BCC'ed for the stats
> he does.

And I switched on automated mails for the rest of them
(regular QA builds and builds for ports triggered because they are
dependencies of an other port).

Currently there will be one mail/port/week for non-commit-builds.
This might change --> some weighting of (days since last sent, number
of build retries).

Three things annoy me and will annoy you also, no doubt :-(
- I don't see any way to detect ports with distfiles that need to be
fetched manually. I'm considering doing a patch to b.p.m and affected
ports to introduce a MANUAL_FETCH var for them; opinions?
- ports that need kernel sources don't fail in a consistent manner (in
the same target, with the same message). I'm considering doing a patch
to b.p.m and affected ports to introduce a NEED_KERNEL_SOURCES var for
them; opinions?
- the fail patters get the "right" reason in about 30% or something
(hence the subject of the BotMail will usually be wrong).


-- 
IOnut - Un^d^dregistered ;) FreeBSD "user"
  "Intellectual Property" is   nowhere near as valuable   as "Intellect"
FreeBSD committer -> itetcu at FreeBSD.org, PGP Key ID 057E9F8B493A297B


-- 
IOnut - Un^d^dregistered ;) FreeBSD "user"
  "Intellectual Property" is   nowhere near as valuable   as "Intellect"
FreeBSD committer -> itetcu at FreeBSD.org, PGP Key ID 057E9F8B493A297B
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20080815/ad2a78a2/signature.pgp