Downloads Page on the Jena website

[Rob Vesse <rv...@dotnetrdf.org>]

Some of you may have noticed from the flood of emails to the commits list
that I've been making some changes to how downloads are managed on the Jena
website.

I noticed some chatter on other project lists and JIRA about Infrastructure
cracking down on projects that were incorrectly linking directly to
dist.apache.org for their releases and while we haven't been brought to task
yet I realised that we are one of the guilty projects.

Therefore I've reworked the downloads page to use the Apache mirroring
script so /download/index.mdtext now becomes the template for the download
page and the new /download/index.cgi invokes an ASF Infra provided script
which processes that page and produces the actual download page directing
people to download via the mirrors rather than directly from dist.apache.org

One side effect of this was that I had to change any links to the download
page to explicitly reference /download/index.cgi rather than
/download/index.html or just /download/ (as otherwise the ASF servers pick
index.html) and index.html now contains broken links because you are
displaying the template which doesn't have the mirror information inserted.
If you need to link to the download page in future please ensure you
reference index.cgi rather than index.html

Rob

Re: Downloads Page on the Jena website

[Rob Vesse <rv...@dotnetrdf.org>]

Yes I think I like the inline MD5 and PGP links better

I fixed the breadcrumbs.js script to ignore any query string portion in
rendering the breadcrumbs

Rob

On 17/12/2013 11:51, "Andy Seaborne" <an...@apache.org> wrote:

>On 16/12/13 13:18, Andy Seaborne wrote:
>> On 16/12/13 11:13, Rob Vesse wrote:
>>> Some of you may have noticed from the flood of emails to the commits
>>>list
>>> that I've been making some changes to how downloads are managed on the
>>> Jena
>>> website.
>>>
>>> I noticed some chatter on other project lists and JIRA about
>>> Infrastructure
>>> cracking down on projects that were incorrectly linking directly to
>>> dist.apache.org for their releases and while we haven't been brought
>>> to task
>>> yet I realised that we are one of the guilty projects.
>>>
>>> Therefore I've reworked the downloads page to use the Apache mirroring
>>> script so /download/index.mdtext now becomes the template for the
>>> download
>>> page and the new /download/index.cgi invokes an ASF Infra provided
>>>script
>>> which processes that page and produces the actual download page
>>>directing
>>> people to download via the mirrors rather than directly from
>>> dist.apache.org
>>>
>>> One side effect of this was that I had to change any links to the
>>> download
>>> page to explicitly reference /download/index.cgi rather than
>>> /download/index.html or just /download/ (as otherwise the ASF servers
>>> pick
>>> index.html) and index.html now contains broken links because you are
>>> displaying the template which doesn't have the mirror information
>>> inserted.
>>> If you need to link to the download page in future please ensure you
>>> reference index.cgi rather than index.html
>>>
>>> Rob
>>
>> Rob,
>>
>> Thanks for doing this.  Much better.
>>
>> I fixed some links for the source sigs and duplicate Fuseki.
>>
>> Shall I turn the PGP and MD5 into inline items so it looks like:
>>
>> * apache-jena-2.11.0.tar.gz (PGP, MD5)
>>
>> (Just saving on vertical space.)
>>
>> The indent on the maven fragment being one, not two as it clearly is in
>> the mdtext (and other pages) is currently defeating me.
>>
>>      Andy
>>
>
>I've put in the links as inline (PGP, MD5), rather than a bullet point
>list. Tell me if you think it's better or worse.
>
>One problem I'm noticing is that the breadcrumbs go wild when you change
>the mirror. e.g.
>
>DOWNLOAD / INDEX.CGI?PREFERRED=HTTP%3A%2F%2FAPACHE.MIRROR.ANLX.NET%2F
>	
>	Andy
>

Re: Downloads Page on the Jena website

[Andy Seaborne <an...@apache.org>]

On 16/12/13 13:18, Andy Seaborne wrote:
> On 16/12/13 11:13, Rob Vesse wrote:
>> Some of you may have noticed from the flood of emails to the commits list
>> that I've been making some changes to how downloads are managed on the
>> Jena
>> website.
>>
>> I noticed some chatter on other project lists and JIRA about
>> Infrastructure
>> cracking down on projects that were incorrectly linking directly to
>> dist.apache.org for their releases and while we haven't been brought
>> to task
>> yet I realised that we are one of the guilty projects.
>>
>> Therefore I've reworked the downloads page to use the Apache mirroring
>> script so /download/index.mdtext now becomes the template for the
>> download
>> page and the new /download/index.cgi invokes an ASF Infra provided script
>> which processes that page and produces the actual download page directing
>> people to download via the mirrors rather than directly from
>> dist.apache.org
>>
>> One side effect of this was that I had to change any links to the
>> download
>> page to explicitly reference /download/index.cgi rather than
>> /download/index.html or just /download/ (as otherwise the ASF servers
>> pick
>> index.html) and index.html now contains broken links because you are
>> displaying the template which doesn't have the mirror information
>> inserted.
>> If you need to link to the download page in future please ensure you
>> reference index.cgi rather than index.html
>>
>> Rob
>
> Rob,
>
> Thanks for doing this.  Much better.
>
> I fixed some links for the source sigs and duplicate Fuseki.
>
> Shall I turn the PGP and MD5 into inline items so it looks like:
>
> * apache-jena-2.11.0.tar.gz (PGP, MD5)
>
> (Just saving on vertical space.)
>
> The indent on the maven fragment being one, not two as it clearly is in
> the mdtext (and other pages) is currently defeating me.
>
>      Andy
>

I've put in the links as inline (PGP, MD5), rather than a bullet point 
list. Tell me if you think it's better or worse.

One problem I'm noticing is that the breadcrumbs go wild when you change 
the mirror. e.g.

DOWNLOAD / INDEX.CGI?PREFERRED=HTTP%3A%2F%2FAPACHE.MIRROR.ANLX.NET%2F
	
	Andy

Re: Downloads Page on the Jena website

[Andy Seaborne <an...@apache.org>]

On 16/12/13 11:13, Rob Vesse wrote:
> Some of you may have noticed from the flood of emails to the commits list
> that I've been making some changes to how downloads are managed on the Jena
> website.
>
> I noticed some chatter on other project lists and JIRA about Infrastructure
> cracking down on projects that were incorrectly linking directly to
> dist.apache.org for their releases and while we haven't been brought to task
> yet I realised that we are one of the guilty projects.
>
> Therefore I've reworked the downloads page to use the Apache mirroring
> script so /download/index.mdtext now becomes the template for the download
> page and the new /download/index.cgi invokes an ASF Infra provided script
> which processes that page and produces the actual download page directing
> people to download via the mirrors rather than directly from dist.apache.org
>
> One side effect of this was that I had to change any links to the download
> page to explicitly reference /download/index.cgi rather than
> /download/index.html or just /download/ (as otherwise the ASF servers pick
> index.html) and index.html now contains broken links because you are
> displaying the template which doesn't have the mirror information inserted.
> If you need to link to the download page in future please ensure you
> reference index.cgi rather than index.html
>
> Rob

Rob,

Thanks for doing this.  Much better.

I fixed some links for the source sigs and duplicate Fuseki.

Shall I turn the PGP and MD5 into inline items so it looks like:

* apache-jena-2.11.0.tar.gz (PGP, MD5)

(Just saving on vertical space.)

The indent on the maven fragment being one, not two as it clearly is in 
the mdtext (and other pages) is currently defeating me.

	Andy