Re: [ML] Re: Hosting of gnustep.org

[H. Nikolaus Schaller]

> Am 19.01.2021 um 17:49 schrieb Ivan Vučica <ivan@vucica.net>:
> 
> On Tue, Jan 19, 2021 at 7:18 AM H. Nikolaus Schaller <hns@goldelico.com> 
> wrote:
>> 
>> 
>>> Am 19.01.2021 um 00:36 schrieb Ivan Vučica <ivan@vucica.net>:
>>> 
>>> Some options:
>>> 
>>> - Should we turn the index into a set of static pages generated from
>>> MySQL data, which we then check in?
>>> - If we turn each entry into a single .md file, we can perhaps
>>> leverage https://github.com/allejo/jekyll-toc
>> 
>> Hm.
>> 
>> The page is highly interactive (searching/filtering/jumping) and accepts
>> proposing updates. It is even counting access and building a TOP10 list.
>> 
>> How should that work with static content?
> 
> Hypothetically speaking, as I don't want to put that much effort and
> I'd rather let it run as-is:

Me too.

> 
> As we stand, there aren't so many items. The list of them can be
> fetched all at once, and filtered/searched client-side, if we need
> that.
> 
> Top 10 is, obviously, not doable. However, I missed that it's based on
> an access metric; I thought it was simply the 'last 10'.
> 
> Interestingly, we have plenty of pages that are making outgoing
> requests and failing; for instance, the number 1 on top 10:
> http://www.gnustep.org/softwareindex/showdetail.php?app=61

Well, someone has moved the refererred page and not updated the
index.

> 
> Should we rethink the index?

We should fix such bugs. The mechanism is there (send an update request).

> 
>> 
>>> - We could have a tiny nginx host which does nothing but route
>>> requests to serving backends based on the path:
>>> - either proxy the request to gnustep.github.io,
>>> - or proxy it to a host with the software index (or even return a
>>> 301 to something like index.gnustep.org)
>> 
>> Keep it simple...
>> 
>> Why not just redirect the entry point www.gnustep.org/index.html to
>> the new page and leave the remainder like the softwareindex on the
>> existing infrastructure (apache, php, mysql)?
>> 
>> This should suffice
>> 
>> www.gnustep.org/index.html;
>> 
>> <meta http-equiv = "refresh" content = "0 ; url = 
>> https://gnustep.github.io/index.html"/>
> 
> Because this situation relegates the "main site" to a secondary URL,
> permanently. I've done that before and I regret it. And it prevents us
> from even attempting to preserve location of other URLs which would be
> completely fine as static pages. (Search engines will eventually index
> gnustep.github.io as the sole main site. But right now it's not even
> on the first page of results on three major search engines I tested
> on.)

Ok, I see. It is also not a good idea if you want do redirect all "deep links".