On Thu, Aug 29, 2019 at 06:03:46PM +0100, Daniel P. Berrangé wrote:
On Thu, Aug 29, 2019 at 04:03:33PM +0000, Cedric Bosdonnat wrote:
> Hello,
>
> Here is an updated version of the previous Storage backend support matrix in the
doc.
> As discussed with Michal, the matrix is now part of hvsupport.html with a
> paragraph and link from storage.html.
>
> This has now been turned into a series to fix the links in TOC of the storage
> page
>
> Cédric Bosdonnat (2):
> Fix hvsupport toc
> Generate status of the backend implementation in hvsupport.html
>
> .gitignore | 1 +
> docs/Makefile.am | 12 +++++--
> docs/apibuild.py | 2 ++
> docs/hvsupport.pl | 4 ++-
> docs/storage.html.in | 13 ++++++++
> docs/storagebackendstatus.py | 63 ++++++++++++++++++++++++++++++++++++
I can't really say I'm a fan of implementing this via a second
python script, as opposed to just putting the logic into the
existing hvsupport.pl script directly.
I can totally understand though if the idea of writing Perl code
was not appealing :-)
FWIW, as the person responsible for much of the Perl code in libvirt,
I think we should make it an explicit TODO item to convert all our
Perl code into Python 3 code.
Please do make it explicit by proposing a patch against the HACKING page
with human-readable justification (i.e. not just ASCII art of animals)
Given that we're likely to incorporate some new languages like
Rust
and/or Go code in libvirt, we definitely want to try to cull some
and or or?
RE: your rewrite of virt-host-validate in Go, I feel like getting that
merged (even in a separate repo) is a precedent for further sunsetting
in case 'or' becomes the preferred conjunction there.
Jano
other languages we use.
Perl is the obvious one. The autotools mix
of autoconf/shell/m4/make languages are the other clear candidates,
to replace with meson's domain specific language.