[Libreoffice-qa] daily vs. 4.-0+ bibisect repos [was: Trouble bibisecting - 'com::sun::star::uno::RuntimeException']

[bjoern]

Hi Mirosław,

Im CC'ing Jibel on this, so he might comment on the things wrt to the Canonical
QA lab repo.

@Jibel: You find the full post at
http://nabble.documentfoundation.org/Libreoffice-qa-Trouble-bibisecting-com-sun-star-uno-RuntimeException-tp4079718p4079667.html

On Thu, Oct 24, 2013 at 10:39:24PM +0200, Mirosław Zalewski wrote:
> First and foremost, we have two repositories from two men. These not
> only serve different purpose, but also are set up differently, which
> means that user must perform different steps depending on which one he
> has. Unfortunately, these differences are never laid out
> systematically. I learned them by trial and error.
> I think that "versions" table could use another column with name of man
> who created archive in question. Then we can reference to them as
> "Bjoern archive" and "Jean archive".

Thats true. I set up the old bibisect repo and the 2013-10-12 actually pick up
exactly were the pre-4.0 build left. As those two are build in the same way
they can (and should) be joined:

 https://bugs.freedesktop.org/show_bug.cgi?id=70652

I build these on my home machine (which then heats the room with 500W for four
days). These bibisect repos were never intended to be (and will never be
intended to be) updated daily or even monthly, but once for every major
release.

The "Jean archive" is set up by jibel (in CC) and shall provide builds
continuously. As this doesnt allow for repacking the repo grows huge.

In the long run, Norbert (shm_get) and Cloph plan to create bibisect repos from
the daily build. Unfortunately, while this has been decided to be the way to go
even already in Berlin, it always fell victim to other topics needing emergency
attention.

> If one is to seriously bibisect, he has to download two packages:
> 1. Bjoern 4.0, which covers everything from 2011-08-06 (about midway
> between branching 3.4 and 3.5 from master) to 2012-12-08 (around
> branching 4.0 from master).
> 
> 2. Either Bjoern 2013-10-12 or Jean 4.0+. They both cover roughly the
> same area, since 2012-12-08 to 2013-09-17 (later date is subject to
> change). The difference between them is density of builds - Jean
> archive is far denser, which means that each build covers fewer
> commits, which means that finding patch that introduced regression is
> easier for developer with bisect log. But we pay for this with almost 4
> times bigger package.

Yes, for bibisecting, currently "Bjoern 4.0" and "Bjoern 2013-10-12" are the
way to go. The Jean archive might be helpful once the Bjoern 2013-10-12 repo
grows too old -- OTOH hopefully, by then we have bibisect-from-dailies in
place.

> In Bjoern packages (well, "4.0" that is; I did not checked
> "2012-10-12", but assume it is set up the same way) our main concern is
> HEAD set up at oldest commit. This way `git rev-parse HEAD` will
> produce ID of oldest commit. If we blindly follow instructions at wiki,
> we will end up assigning "latest" tag to oldest possible commit.
> Our second problem is that we have empty git working tree.
> Both of them can be fixed by issuing:
>     git checkout master

Yes, my repos are intentionally checked out at an empty root commit, as
everything else would make them bigger without need.
 
> If anyone can update wiki page with information provided in this
> message, please do. 

Oh, please do yourself -- chinese whispers is a fun game, but its not good for
technical documentation. ;)

If you dont want that, please at least add a link to this Mail on some archive
there, as its a good discussion of the topic.

Best,

Bjoern