Hello Boris. I know that opening a Nexus JIRA is the usual way to get responses.
BTW also consider scraping the SCM URLs from the POM files and contact the upstream Repos, the maven -src archives are often pruned down and not builtable (if present at all). So it does not hurt to archive them, but don’t expect them to help you in all cases. Sadly Tag parsing and SCM scraping is not the most reliable thing (we do it for consumed dependencies), but with some manual overwrite it’s manageable at small scale. Maybe you would get help if you provide such a Registry as a github project or ask OSSIndex for cooperation. Gruss Bernd -- http://bernd.eckenfels.net ________________________________ Von: Boris Baldassari <boris.baldass...@gmail.com> Gesendet: Montag, Juni 14, 2021 5:25 PM An: dev@maven.apache.org Betreff: Re: Software Heritage connector to Maven repositories Hi Frederik, Thanks for the kind answer and pointers. Yes, we know that non-consensual mirroring, as well as scrapping, is explicitly forbidden. Hence the question here. :-) It should be noted however that our case is a bit specific: we want to get only *some* types of artefacts (poms and src jars), but we want *all* of them. So it's a special case of partial mirroring. Furthermore we will only access them *once* for the archiving so the "re-use" feature of a proxy, mirror or similar is not needed. Maybe a repo maintainer could provide some more wisdom? Thanks, cheers! -- boris On 14/06/2021 13:41, Frederik Boster wrote: > I am not in any way affiliated with Apache or Sonatype. So take my opinion > with a grain of salt :) > > Trying to mirror the entire Maven Central repository will unfortunately get > you automatically banned. > To circumvent that I would suggest you setup your own Maven Central mirror > first. [1] > > [1] > https://maven.apache.org/guides/mini/guide-mirror-settings.html#creating-your-own-mirror > > On Mon, Jun 14, 2021, 12:12 Boris Baldassari <boris.baldass...@gmail.com> > wrote: > >> Hiho good people, >> >> I am currently developing a Maven repositories connector for the >> Software Heritage Foundation [1]. In a nutshell, the SWH aims to >> archive all existing source code in the world, and provides useful >> publicly available services and related tools (unique IDs/DOIs, search, >> datasets, graph tools..). It's all open-source, and many large forges >> and software systems have already been archived (GitHub, GitLab, npm, >> pypi, debian packages, CRAN..) [2]. Now we would like to archive the >> Maven ecosystem. >> >> [1] https://www.softwareheritage.org/ >> [2] https://archive.softwareheritage.org/ >> >> I'm reaching out to ask for wisdom and start a discussion about how this >> could be achieved without impacting anybody, i.e. neither Maven >> repositories maintainers nor the users. Our plan for now is to use the >> maven indexer indexes for the listing, and then download poms and source >> jars, in a way that we see as the most efficient and fair. We of course >> respect all rate-limiting policies (and http error codes), and we are >> polite and patient (although tenacious). >> >> So, here are my questions: >> >> * Who should we talk to to achieve that? i.e. are there maven repository >> maintainers on the list, or do you know of a better place to ask? >> >> * Although we believe the above mentioned process is the most efficient >> and fair one, maybe there is a better way to list, and archive artefact >> sources? Any feedback or mere thoughts are welcome. >> >> >> Thanks in advance, have a wonderful day! -- Boris Baldassari Castalia Solutions -- Elegant Software Engineering Web: http://castalia.solutions Tel: +33 6 48 03 82 89 --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org