Re: Bulk build information

* On 2026-04-06 at 11:16 BST, Jonathan Perkin wrote:

>* On 2026-03-27 at 14:11 GMT, Thomas Klausner wrote:
>
>>On Fri, Mar 27, 2026 at 12:58:09PM +0100, Jonathan Perkin wrote:
>>>My question is what other information do people want added?  One thing I'd
>>>like to add if possible is a link to the change that might be responsible
>>>for the failure or most recent commit, though naturally that will get
>>>complicated.  Anything else?
>>
>>My wishlist: differences to the previous build: what failed before and
>>built now, what built before and failed now, what was unmasked from a
>>fixed build failure and still doesn't build.
>>
>>But that needs information about the previous build.
>
>I've now added most of this to the bob report.

The latest bob reports now also include links to the commits that are 
the likely causes of any new failures.  For example:

   https://reports.pkgci.org/Darwin/14.5/arm64/20260413T215911Z/report.html

shows the commits made to each PKGPATH since the previous build run for 
any new failures, and should help identify the causes of regressions.

Ideas for future enhancements include automatically cc'ing anyone on the 
commit list, as well as looking further afield than just PKGPATH for 
changes to any file in SCAN_DEPENDS so that e.g. changes to mk/ or bl3 
that end up breaking packages are also recorded.  Both would be pretty 
trivial to add, but I feel deserve a lot more thought and discussion.

Cheers,

-- 
Jonathan Perkin                    pkgsrc.smartos.org
Open Source Complete Cloud   www.tritondatacenter.com