Re: [bitbake-devel] Improving npm(sw) fetcher & integration within Bitbake

[Jasper Orschulko <Jasper.Orschulko@iris-sensing.com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi Stefan,

would you (and any other interested party) be interested in a
conference call with Martin and I? We are greatly interested in your
current prototypes, how they perform for the various usecases and
if/how we can help you with the implementation.

- -- 
With best regards

Jasper Orschulko
DevOps Engineer

Tel. +49 30 58 58 14 265
Fax +49 30 58 58 14 999
Jasper.Orschulko@iris-sensing.com

• • • • • • • • • • • • • • • • • • • • • • • • • •

iris-GmbH
infrared & intelligent sensors
Schnellerstraße 1-5 | 12439 Berlin

https://iris-sensing.com/





On Mon, 2021-11-08 at 09:01 +0100, Stefan Herbrechtsmeier wrote:
> Am 06.11.2021 um 17:58 schrieb Mike Crowe:
> > On Thursday 04 November 2021 at 14:09:42 +0100, Alexander Kanavin
> > wrote:
> > > To the best of my knowledge, no other companies at the moment are
> > > using
> > > Yocto to integrate npm-based items into a product.
> > 
> > We make light use of npm in the production of the rootfs for our
> > products.
> > The upgrade to Dunfell was a bit painful, and the slowness is
> > annoying too,
> > but I think that we've ended up with things being better in the end
> > because
> > we can now be sure that all the sources are captured correctly.
> 
> We have a similar problem.
> 
> > We did run into a few bugs and some of our fixes for those have
> > landed. The
> > npm fetcher seems to be fighting with the usual way that Bitbake
> > expects
> > fetchers to work that I don't really understand enough (from either
> > side)
> > to know how to fix. (e.g.
> > https://bugzilla.yoctoproject.org/show_bug.cgi?id=14383 which
> > doesn't
> > directly affect us, but the underlying cause meant that our usual
> > method
> > for capture sources needed some extra workarounds.)
> 
> The npmsw fetcher is the only one that fetch multiple sources from
> one 
> and translate it into multiple fetch commands.
> 
> I think this is a wrong approach and makes more problems as it solve.
> I 
> suggest to move the logic to the recipetool and direct use the npm 
> fetcher per npm package (dependency). Even the npm fetcher could be 
> avoid if we doesn't require an autorev feature for an npm package.
> 
> This solution has the advantage that you could manipulate the 
> dependencies inside a recipe and you could replace a npm package with
> an 
> other oe package (DEPENDS / RDEPENDS).
> 
> The disadvantage is a big recipe instead of a big foreign
> configuration 
> file and it is impossible to use a foreign configuration file direct 
> from a git repository. But I think this use case doesn't match with
> the 
> oe requirements and we should improve the recipe generation so that
> this 
> could be done automatically.
> 
> Regards
>    Stefan
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEE4WyPMIC5Ap4+Ooo1Ygqew07VMNUFAmGJG5oACgkQYgqew07V
MNVTDgf/d6eY/AlnnHg3OkHP1GfKoRDmMKNuS1sjbx7xMyGsZBpWRrPGPwpLWa6T
rdS3qcGyVnMBbHeTRDp2yuYxxFCrBqyiSGpHfQwmUL9u94Si3ujTdKZKpQSYquT3
FDZW4WDILP6LCyfVFnLdLyLyVJd62X2tHd8jfG8c+70xQ5Ibh0CDpEslHvqWb/nS
30pKJ3FX/VnMit6BIY4ndiidW+UX/rnu5+w2AeqEk9H9t1ZcjCJyEP3TgF+f3icN
6S6um5EG+A9YJ7qeuWVDwxGyDkWod7TkXMTxYAdhud8Sfo0gkx3cJ1zLr10szGIj
bSpEn6im5cccoeaFtaWpiLUmuwpEzg==
=uNna
-----END PGP SIGNATURE-----