MIME-Version: 1.0
References: <20250714194351.e9.nmisch@google.com>
In-Reply-To: <20250714194351.e9.nmisch@google.com>
From: Magnus Hagander <magnus@hagander.net>
Date: Mon, 14 Jul 2025 21:47:42 +0200
Message-ID: 
 <CABUevEwb9OWgQp1XDRe-hQB8cQS5kbKO1jVb-_E-zzu1jBZQkA@mail.gmail.com>
Subject: Re: List archives "Next" button new behavior
To: Noah Misch <noah@leadboat.com>
Cc: pgsql-www@postgresql.org
Content-Type: multipart/alternative; boundary="000000000000851c960639e8ef67"
Archived-At: 
 <https://www.postgresql.org/message-id/CABUevEwb9OWgQp1XDRe-hQB8cQS5kbKO1jVb-_E-zzu1jBZQkA%40mail.gmail.com>
Precedence: bulk

--000000000000851c960639e8ef67
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, Jul 14, 2025 at 9:43=E2=80=AFPM Noah Misch <noah@leadboat.com> wrot=
e:

> A few weeks ago, the "Next" button on pages like
> https://www.postgresql.org/list/pgsql-www/since/202507100000 changed
> behavior:
>
> Old behavior: "Next" would use the timestamp of the newest message as of
> when
> you last loaded the page.  "Next" would take you to a page containing tha=
t
> timestamp and all newer-timestamp messages.  For example, as I'm writing
> this,
> the newest message has timestamp 2025-07-10T14:03.  "Next" used to take m=
e
> to
> https://www.postgresql.org/list/pgsql-www/since/202507101403 and I'd then
> see
> the 14:03 message and later messages, if any.
>
> New behavior: As long as the latest message is from 2025-07-10, "Next"
> takes
> one to https://www.postgresql.org/list/pgsql-www/since/202507100000.
> Also,
> visiting https://www.postgresql.org/list/pgsql-www/since/202507101403
> redirects to 202507100000 (zeroing the HH:MM from the request).
>
> (I'm simplifying; there are more behavior nuances if one starts at a date
> far
> enough in the past for messages to fill the page.)
>
> Does anyone have background on this change?  I found no discussion of it.
>
> I liked the old behavior better, since it made the "Next" button
> convenient to
> see everything since I last caught up.  That said, I'll adapt if the
> change is
> intentional and/or favored.
>

Yes. The problem with the old more detailed URLs is that in practice it
generated an infinite number of URLs for AI bots to crawl, and they did. We
got a constant load of a few hundred bot hits per second just looking at
those indexes. Because AI is the future....

--=20
 Magnus Hagander
 Me: https://www.hagander.net/ <http://www.hagander.net/>
 Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>

--000000000000851c960639e8ef67
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote g=
mail_quote_container"><div dir=3D"ltr" class=3D"gmail_attr">On Mon, Jul 14,=
 2025 at 9:43=E2=80=AFPM Noah Misch &lt;<a href=3D"mailto:noah@leadboat.com=
">noah@leadboat.com</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quot=
e" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204)=
;padding-left:1ex">A few weeks ago, the &quot;Next&quot; button on pages li=
ke<br>
<a href=3D"https://www.postgresql.org/list/pgsql-www/since/202507100000" re=
l=3D"noreferrer" target=3D"_blank">https://www.postgresql.org/list/pgsql-ww=
w/since/202507100000</a> changed<br>
behavior:<br>
<br>
Old behavior: &quot;Next&quot; would use the timestamp of the newest messag=
e as of when<br>
you last loaded the page.=C2=A0 &quot;Next&quot; would take you to a page c=
ontaining that<br>
timestamp and all newer-timestamp messages.=C2=A0 For example, as I&#39;m w=
riting this,<br>
the newest message has timestamp 2025-07-10T14:03.=C2=A0 &quot;Next&quot; u=
sed to take me to<br>
<a href=3D"https://www.postgresql.org/list/pgsql-www/since/202507101403" re=
l=3D"noreferrer" target=3D"_blank">https://www.postgresql.org/list/pgsql-ww=
w/since/202507101403</a> and I&#39;d then see<br>
the 14:03 message and later messages, if any.<br>
<br>
New behavior: As long as the latest message is from 2025-07-10, &quot;Next&=
quot; takes<br>
one to <a href=3D"https://www.postgresql.org/list/pgsql-www/since/202507100=
000" rel=3D"noreferrer" target=3D"_blank">https://www.postgresql.org/list/p=
gsql-www/since/202507100000</a>.=C2=A0 Also,<br>
visiting <a href=3D"https://www.postgresql.org/list/pgsql-www/since/2025071=
01403" rel=3D"noreferrer" target=3D"_blank">https://www.postgresql.org/list=
/pgsql-www/since/202507101403</a><br>
redirects to 202507100000 (zeroing the HH:MM from the request).<br>
<br>
(I&#39;m simplifying; there are more behavior nuances if one starts at a da=
te far<br>
enough in the past for messages to fill the page.)<br>
<br>
Does anyone have background on this change?=C2=A0 I found no discussion of =
it.<br>
<br>
I liked the old behavior better, since it made the &quot;Next&quot; button =
convenient to<br>
see everything since I last caught up.=C2=A0 That said, I&#39;ll adapt if t=
he change is<br>
intentional and/or favored.<br></blockquote><div><br></div><div>Yes. The pr=
oblem with the old more detailed URLs is that in practice it generated an i=
nfinite number of URLs for AI bots to crawl, and they did. We got a constan=
t load of a few hundred bot hits per second just looking at those indexes. =
Because AI is the future....</div><div>=C2=A0</div></div><span class=3D"gma=
il_signature_prefix">-- </span><br><div dir=3D"ltr" class=3D"gmail_signatur=
e"><div dir=3D"ltr"><div>=C2=A0Magnus Hagander<br>=C2=A0Me: <a href=3D"http=
://www.hagander.net/" target=3D"_blank">https://www.hagander.net/</a><br>=
=C2=A0Work: <a href=3D"http://www.redpill-linpro.com/" target=3D"_blank">ht=
tps://www.redpill-linpro.com/</a></div></div></div></div>

--000000000000851c960639e8ef67--