Re: New blog - who dis?

On Mon, Sep 4, 2023 at 2:47 PM Andreas 'ads' Scherbaum <ads(at)pgug(dot)de> wrote:
>
>
>
> On Mon, Sep 4, 2023 at 1:00 PM Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
>>
>> Hello,
>>
>> On 2023-Sep-04, Andreas 'ads' Scherbaum wrote:
>>
>> > I plan to migrate my blog to a new software platform, which
>> > will also change the URLs which appear in the RSS feed. There
>> > is no convenient way to keep the old URLs in place.
>> >
>> > Most importantly, this will affect Planet PostgreSQL, which
>> > suddenly might see about 150 "new" blog postings.
>> >
>> > Is there a recommended way how to deal with such a move?
>>
>> Each post in the blog has a "guid" unique identifier, which is usually
>> the same as the URL, but some platforms let you set up something
>> different. If you can "migrate" your posts to the new platform while
>> keeping the GUIDs, that would be best -- they would not be seen as new
>> posts. The actual URLs don't actually matter.
>
>
> The guid in my case is the full URL of the posting, including the domain.
> I would need to break and fix quite a few things to port this guid over to
> the new system, and I can easily miss something before going live.

You wouldn't need to keep the URL for the new posts, only the GUIDs.
That is, new posts could have GUIDs in a new format, old posts could
just use the old URL in the GUID and the new URL in the, well, URL.

> I'd rather not go down this path.

Strictly speaking, per the RSS requirements, you have to. Not donig
so will cause reposts for anybody *else* who is tracking your RSS feed
as well, not just Planet PostgreSQL.

>> If your platform doesn't let you do this, I think PlanetPostgres would
>> mark the new posts as hidden anyway, because of the volume (but pester
>> everyone along the way). That way only future new posts (actually new
>> posts) would be syndicated, but everything would appear duplicate in the
>> admin interface.
>
>
> Will it work if I disconnect the old blog from Planet, then move the software
> and apply the blog again? Will this ingest all previous postings on the
> feed, or just the new ones?

This will ingest all previous postings that are in the RSS feed. So
one way you could do is perhaps put a cutoff on your side so that
posts made before date <x> are simply not included in the RSS?

Now, planet has a few safeguards against this, so it may not be a huge
problem there -- but per above, there are also others...

* No posts older than 7 days will get posted to *twitter*. They only
go in the planet RSS feed(s).
* The planet RSS feeds contain 30 items. The homepage as well. At this
point you can see this goes back to Aug 24, so not very far. That
means that any entries older than that will be ingested into the
system, but they won't actually be shown to anybody.
* The feed passed through to www.postgresql.org further restricts this
to just the past 10

So this would indicate that if you have a period of say 2 weeks of no
postings, *planet* won't notice. Others might.

Another thing is that if a poll finds 5 or more posts at the same time
it will automatically mark them as hidden and send an email about
something probably went wrong. This will not affect a new blog, but if
you just modify the contents of the RSS this is I think likely to
happen. Which would again save planet, but not any other consumers of
your feed.

//Magnus