On 10/11/20 10:25 PM, Jonathan S. Katz wrote:
> On 10/11/20 8:08 PM, Nikolay Samokhvalov wrote:
>> Not sure if this page has some incoming links from somewhere, but being
>> on the page
>> https://www.postgresql.org/docs/release/13.0/ I naturally replaced the
>> version with 9.1 (omitting the minor part) -- and found something
>> unexpected:
>>
>> https://www.postgresql.org/docs/release/9.1/ -- has "E.99.1. Migration
>> to Version 9.0.1", which is, obviously, not what is expected.
>>
>> Moreover, the right sidebar has links to non-existnent versions such as
>> 9.9: https://www.postgresql.org/docs/release/9.9/.
>
> What, you don't remember PostgreSQL 9.9? People were so excited for it! ;)
>
>> If I entered the full version, with minor part
>> (https://www.postgresql.org/docs/release/9.2.0/) it would be okay, of
>> course.
>
> I seem to recall a similar thread (which I can't seem to find) about an
> error, i.e. one would put in 9.1 and it would 404. Basically, we have to
> deal with three difference number schemes:
>
> - Pre-6.0
> - Pre-10.0
> - Post-10.0
>
> where each of their own quirks. I think the probability of people link
> hacking the release notes to the older versions is low, but even so, we
> should either have, e.g. 9.1 renders the actual 9.1.0, or we should error.
>
> I propose we do the former. Please see attached patch.
I touched up the comment on the patch and pushed it. Commit is here[1].
I'm sad to see PostgreSQL 9.9 go by the wayside, but glad that we now
have proper behavior if someone should try to link to, e.g., "9.1".
Thanks for reporting!
Jonathan
[1]
https://git.postgresql.org/gitweb/?p=pgweb.git;a=commitdiff;h=f923e095352f1be55d0c3658e55472271ed4d8b3