Code review comment for lp:~allenap/maas/remove-squashfs

> So let me get this straight... the reason of the removal is because its
> causing too many issues (bugs).
>
> What are those?
>
> 1. One bug report I see linked is server side and is *not* An ssue with this
> feature.
> 2. Another report is an issue of not having updated a file with the release we
> want to import which is not even related to the script maas-import-squashfs
> but also affects maas-import-pxe-files and maas-import-ephemerals. and this
> bug would have still being there even of the feature wouldnt have existed.
> 3. Legitimate maas-import-squashfs.

Removing squashfs support means we can close two of these bugs, and
the third becomes a trivial fix.

>
> So from above..  how in the world this feature is attracting a
> disproportionate number of bugs if only 1 of them is related.... with a fix
> that should only take a couple lines of code???

For the size and benefit of the feature - i.e. a ~1 minute speedier
install - it's attracting a disproportionately high number of
bugs. That's my impression - I haven't tried to quantify it - but I'm
not alone in thinking this.

> So i think t is not fair to argue that this causes more issues than
> anything else.

I'm not saying that. I'm saying that it's more trouble than it's worth
right now. We don't *need* it, so it's better to move it out so that
we can work on - and SRU - a less-featureful but more stable product.

What is broken is the lack of tests. "Untested code is broken code" is
a principle that the Launchpad teams have all tried to follow for
years. It would be good to reintroduce this feature later on, or FPI,
but not in its current form.

>
> While it does lack tests.. so that's other import script that we have also
> seen broken in worse ways. This scripts is rock solid outside Maas and not
> because there's one minor bug it isn't. Otherwise other things wouldn't be
> rock solid for same or even worse bugs found in released versions of Maas.

Bugs filed is not a measure of quality.