Sebb,
I'm going to revert your change for multiple repos:
Firstly on the project page it links to a list of all the Git repos, not to
individual ones - so it to generates an invalid link:
https://gitbox.apache.org/repos/asf#gora-site (anchor doesn't exist)
https://gitbox.apache.org/repos/asf#gora (lists all the gora repos)
I can see that you wanted them listed on the Infra JIRA template. I'll
change this to link to the list of repos. So instead of:
https://gitbox.apache.org/repos/asf/gora.git
https://gitbox.apache.org/repos/asf/gora-site.git
it will point to
https://gitbox.apache.org/repos/asf#gora
Niall
On Wed, 23 Apr 2025 at 12:15, Niall Pemberton <niall.pember...@gmail.com>
wrote:
> I've added a new "project-data-plugin.rb" to "pre-process" the project
> data - it creates a new list of hashes (easier to process) and adds the
> "project_id" as an attribute and sets values if they are not specifically
> set (name, fullname, domain etc):
>
> https://github.com/niallkp/attic-jekyll-site/blob/main/_plugins/project-data-plugin.rb
>
> For example, the nasty hack to sort the project list is gone and now using
> the standard liquid "natural sort" on project names (rather than
> project_id):
>
> https://github.com/niallkp/attic-jekyll-site/blob/main/_includes/project-list.html
>
> Also, duplicated logic for things like project_name and sorting is gone
> from the tracking page, making it simpler:
>
> https://github.com/niallkp/attic-jekyll-site/blob/main/tracking.md?plain=1
>
> Niall
>
> On Mon, 21 Apr 2025 at 12:45, Niall Pemberton <niall.pember...@gmail.com>
> wrote:
>
>> On Mon, 21 Apr 2025 at 09:24, sebb <seb...@gmail.com> wrote:
>>
>>> On Sun, 20 Apr 2025 at 18:36, Niall Pemberton <niall.pember...@gmail.com>
>>> wrote:
>>> >
>>> > On Sun, 20 Apr 2025 at 17:51, Niall Pemberton <
>>> niall.pember...@gmail.com>
>>> > wrote:
>>> >
>>> > >
>>> > >
>>> > > On Sun, 20 Apr 2025 at 09:44, sebb <seb...@gmail.com> wrote:
>>> > >
>>> > >> On Sun, 20 Apr 2025 at 01:10, Niall Pemberton <
>>> niall.pember...@gmail.com>
>>> > >> wrote:
>>> > >> >
>>> > >> > Hi,
>>> > >> >
>>> > >> > I have created a PoC to move the current Attic site generation
>>> from
>>> > >> XDOC &
>>> > >> > Forrest to Markdown & Jekyll:
>>> > >> > https://github.com/niallkp/attic-jekyll-site
>>> > >> >
>>> > >> > Each project has a YAML file with details about it and its
>>> retirement
>>> > >> and
>>> > >> > that drives everything:
>>> > >> > - generates a project page
>>> > >> > - updates the Process Tracking Page
>>> > >> > - updates the menu/list of projects in the Attic
>>> > >> > - Generates the "flag" files for the Attic Banner on the site &
>>> Wiki
>>> > >> >
>>> > >> > At the moment it is pretty much functionally complete, but I've
>>> put a
>>> > >> TODO
>>> > >> > list in the README.md or the project:
>>> > >> >
>>> https://github.com/niallkp/attic-jekyll-site/blob/main/README.md
>>> > >> >
>>> > >> > Thoughts/feedback welcome
>>> > >>
>>> > >> Looks good, and could be worth pursuing, but there are some missing
>>> > >> features that are essential for a full evaluation.
>>> > >>
>>> > >> There needs to be an action to generate the output so it can be
>>> checked.
>>> > >> Ideally so the site can be viewed directly.
>>> > >>
>>> > >
>>> > > I have manually published the generated site here for now, so you
>>> can have
>>> > > a look:
>>> > > https://niallkp.github.io/
>>> > >
>>>
>>> Thanks, very useful.
>>>
>>> > There were no "In Progress" projects (i.e. project's whose move to
>>> Attic is
>>> > not complete), so I've just added the data files for Gora & Mnemonic.
>>>
>>> I see you have also added ODE, which has non-standard CWiki names;
>>> useful.
>>>
>>
>> I need to double check, but I think I've added all the project data files
>> except the two "revived" projects (Ambari & XMLBeans?).
>>
>>
>>> > The main bits of the site are generated from the project data files.
>>> Just
>>> > to highlight a few of the generated pages:
>>> > - The Tracking page: https://niallkp.github.io/tracking.html
>>> > - Project Pages e.g. Gora
>>> https://niallkp.github.io/projects/gora.html
>>> > - The list of projects on each page
>>> > - The site flag files are generated in "/flagged/site": e.g.
>>> > https://niallkp.github.io/flagged/site/any23.txt
>>>
>>> These don't need any content; also at present Puppet expects directories.
>>>
>>> > - The cwiki flag files are generated in "/flagged/site": e.g.
>>> > https://niallkp.github.io/flagged/cwiki/mnemonic.txt
>>>
>>> The contents are fine, but they cannot be accessed from there by Puppet.
>>>
>>
>> OK, I just threw something in as a PoC, but happy to polish the fine
>> detail. I'll take a look at getting it to replicate whats currently done -
>> but it might not be for a few days.
>>
>>
>>
>>> > Projects that are "in progress" also have a link to a "templates"
>>> page, so
>>> > for example, here is Gora's Template page:
>>> > - https://niallkp.github.io/templates/gora.html
>>> > The "template" page has text for creating the JIRA and sending out the
>>> user
>>> > & announcement emails.
>>>
>>> That's very useful.
>>> If possible, add buttons to copy the content to the clipboard?
>>> i.e. as per process.md
>>>
>>
>> Good idea, I'll add this to the todo list
>>
>>
>>> > Obviously the "Look & Feel" is crap. Its using the default "Minima"
>>> theme
>>> > from Jekyll. We can chose a different theme or try to re-create the
>>> L&F of
>>> > the existing site
>>>
>>> `The tracking.md page seems a bit out of place: it looks more like a
>>> template.
>>>
>>
>> We could move the table generation part into an html include (don't think
>> markdown files can be includes or templates, but maybe I'm wrong) - but
>> there wouldn't be much else in tracking.md then, just the page elements you
>> mention below.
>>
>>
>>> Also missing are the required page elements (under 'Related Apache
>>> links' on the current site) and some Attic links (see 'The Apache
>>> Attic')
>>>
>>
>> +1, todo
>>
>>
>>> And there needs to be a way to create the initial project.yaml files.
>>> The current site has a Python script to do this; it just requires the
>>> project name.
>>>
>>
>> Good idea, I'll add this to the todo list
>>
>>
>>> > Niall
>>> >
>>> >
>>> >
>>> > >
>>> > > Niall
>>> > >
>>> > >
>>> > >
>>> > >>
>>> > >> It looks like the CWKI flags are generated as part of the website.
>>> > >> However, the Confluence webserver host does not have access to the
>>> > >> Attic website (unlike the TLP server, which hosts Attic as well as
>>> the
>>> > >> retired sites).
>>> > >>
>>> > >> I don't think it makes sense to checkout the entire Attic website on
>>> > >> the Confluence server merely to get access to the flag files.
>>> > >> Probably need to use a separate GHA to generate the required files
>>> in
>>> > >> a separate branch.
>>> > >> (AFAICT, Git does not support checkout of a sub-directory, at least
>>> > >> not in a way that can be used by Puppet config)
>>> > >>
>>> > >> There are some Python scripts in the existing codebase that need to
>>> be
>>> > >> converted to use the new format.
>>> > >> Probably only retire.py needs significant work.
>>> > >>
>>> > >> Minor nit: the plugin files say that the flagged files are used by a
>>> > >> Lua filter to decide whether to add the banner.
>>> > >> That is not the case; the httpd config extracts the project name
>>> from
>>> > >> the request, and checks to see if there is a flag file with that
>>> name.
>>> > >> If so, the Lua filter is applied to the output.
>>>
>>
>> need to remember to update this as well
>>
>> Niall
>>
>>
>>> > >>
>>> > >> > Thanks
>>> > >> >
>>> > >> > Niall
>>> > >>
>>> > >
>>>
>>
