varilink / data-app

The Derby Arts and Theatre Association web application, which is based on the Perl, CGI::Aplication framework.

David Williamson @ Varilink Computing Ltd

In this GitHub account I share my work as a Linux and blockchain system administrator and full-stack web developer, where it can be made public. The repositories herein are predominantly not packaged applications but I share them in case they are useful to anybody who is tackling similar challenges and might benefit from seeing how somebody else did it. I also blog on selected topics related to this work for the same reason.

Latest Blog Posts

• Network and DNS in “Services – Docker”
• Testing Ansible Roles Using Containers
• Managing Nodeos P2P Peer Connections
• Nodeos Blocks Log Stride
• Nodeos Replays

Projects

I organise these repositories by project but I don't use the project board functionality within GitHub. There follows a list of projects and the repositories associated with each of them.

Derby Arts and Theatre Association

The Derby Arts and Theatre Association (DATA) is the representative body for amateur performing arts in the Derby area.

Repositories:

• data-asnsible
• data-admin
• data-app
• data-docker
• data-web

The Friends of Bennerley Viaduct

The Friends of Bennerley Viaduct (FoBV) are "Dedicated to restoring, conserving and celebrating Bennerley Viaduct".

Repositories:

• fobv-ansible
• fobv-docker
• fobv-wordpress

New Opera Company

New Opera Company, Derby, founded in 1959 was the only local company singing grand opera.

Repositories:

• newopera-ansible
• newopera-docker
• newopera-wordpress

Telos UK

Telos UK is "a Founding Block Producer on the Telos Public Network".

Repositories:

• telos-eosio
• telos-phplib
• telos-toolbox

Varilink Computing Ltd (Libraries)

Libraries that are used across multiple of my projects.

Repositories:

• libraries-ansible_playbooks
• libraries-ansible_roles
• libraries-php
• libraries-wordpress_plugins

Varilink Computing Ltd (Services)

Services such as backup, mail, etc. deployed across the Varilink Computing Ltd server estate.

Repositories:

• services-ansible
• services-docker

Varilink Computing Ltd (Tools)

Client side tools that I use across multiple projects.

Repositories:

• tools-exim
• tools-gimp
• tools-npm
• tools-proxy
• tools-wordpress

Varilink Computing Ltd (Website)

The website for Varilink Computing Ltd.

Repositories:

• website-ansible
• website-docker
• website-wordpress

data-app's Issues

There is no mechanism to remove redundant CGI session files

These are located in /tmp and all prefixed cgisess_

They build up indefinitely with no clear-down mechanism in place.

Add a third member society membership status

We currently have membership statuses of ACTIVE and INACTIVE. An INACTIVE member society is not listed and user accounts associated with such a member society can no longer list events.

In discussion at committee, we think it would be a good idea to have a status whereby the member society is still listed but events can't be added for it. We would place member societies whose subscription renewal was overdue in this category.

Javascript for YouTube modals loaded on every page

The Javascript for YouTube modals is located towards the end of partials/layouts/footer.tt, which means that it is included on every page. We should use the mechanism whereby selected scripts are included only on the pages that they are required for by pulling in separate script blocks. This is also implemented in partials/layouts/footer.tt.

Add moderation capability

As more member users are adding their own events to the DATA Diary we could do with an admin moderation capability. Possibly set it so that a member user can only add an event in placeholder status and then submit a request to the admins to publish it.

No role set in session cookie

One user had an issue whereby after they had apparently successfully signed in the Admin menu did not become available to them. When I checked their session file on the back-end, for some reason their "role" was set to blank.

Automated Tweets and Facebook posts of event without an image use the "Click me to upload an image" image

Poor name for "base URL" environment variable

The application uses an environment variable named root to hold the base URL for the website. This is a poor variable name for that purpose. It should be changed to something more descriptive, e.g. base_url.

Cookie set when not needed

We are setting a session cookie for every site visitor but this is only needed to support the authentication/authorization when a site visitor needs to access secure content. This is unnecessary and wasteful.

Hyperlinks recorded in database without associated text to display

Hyperlinks can be recorded in the database for organisations. However, there's no capability currently to also record the text to display for a hyperlink.

Format issue when odd number of events in email bulletin

Sort out problems with email bulletin containing an odd number of events. In a desktop email client, the footer then only occupies one of the columns. In the Mailchimp archive, the event on a row by itself is stretched but its image isn't.

Unspecific error message when image upload exceeds size limit

I inadvertently tried to upload an event image in live that exceeds the 2MByte threshold. This failed with a generic "error" message. I'm sure that the message used to be more specific than that. Whether that's the case or not it certainly needs to be. Investigate and fix.

Add ability to send an image or a link to an image to the "Notify Us" form

Validate longest string in event box office or information field

Recently somebody created an event with a very long string in the box office or information field. This cause problems for the printed, programme listing flyer because it wasn't possible to put a line break in that unbroken string. That messed up the format of the flyer. Introduce some validation on data entry to stop this happening.

Preformated A4 and A5 programme listings not working

I'm sure that these worked at go-live but for some reason they're now not working. The A4 format fits all the event details in but has no left and right margins. The A5 format doesn't even get all the text in.

Integrate with Facebook events

The current integration with Facebook merely creates posts to advertise DATA events. Facebook has its own events functionality. Integrate with it.

Extract contacts to a file shared with committee members

Fairly self-explanatory, extract the contacts database to a file that is securely shared with other committee members, say on a daily basis.

Potential invalid HTML in generated Mailchimp templates

I see evidence that the generated Mailchimp templates contain invalid HTML; for example opening and closing tag mismatches.

Identical adjacent Facebook posts or Tweets

When the future events in the DATA Diary are relatively sparse, we can end up with a Facebook post or a Tweet that is identical to the previous one. Upgrade the integration scripts so that they can examine the current feed, determine when this is going to be the case and not post or Tweet when it is.

Replace hard code of copyright year

The copyright year is hard coded as 2018 in partials/layouts/footer.tt. Replace with dynamic generation of the current year or, possibly a range of years, starting with 2018 and going up to the current year - see this article on the the international copyright symbol.

Special characters not uniformly handled across channels

If a special character, for example containing accent marks, is included in an event description then it is not uniformly handled across publicity channels. It may work on the website but not translate correctly to Twitter.

reCAPTCHA integration needs an overhaul

There is a reCAPTCHA secret key hard coded within the DATA::Plugin::Captcha module. Fortunately it's out of date, i.e. not the current secret key in use, which is good news from a security perspective but does imply that this plugin can't be working.

I set reCAPTCHA up for this domain on its own Google account. I should move it to my main Google account where I have all the other domains that I've setup reCAPTCHA integration for.

This domain is still using reCAPTCHA version two and needs to be upgraded to use reCAPTCHA version 3.

Allow a start date to be entered for one-page, programme event listing flyers

When people generate a one-page, programme event listing flyer from the home page, it lists events from the day that the flyer is created. However they may not be planning to use that flyer in programmes until some time later, when some of those events may have already finished. Allow the entry of a start date at the time that the flyer is generated with only events that are current or future at that start date then being included.

Possibility for non member society events to be listed via "Presented By"

"Presented by" defaults to the member society associated with a member user account adding an event. However it can be overridden. This option is given to allow member societies to augment "presented by" with, for example, the names of solo performers they have added for a performance. However, this makes it possible for member users to add an event that is actually not in any way associated with their member society by overriding "presented by".

Include reCAPTCHA script on every public page

Currently in tt/partials/layouts/head.tt we include a script tag for Google's reCAPTCHA API for every page on which there is a form submission. However, since we're using the invisible reCAPTCHA I think perhaps that it should be included on every public page, so that it can observe the behaviour of the user even before an attempt to submit a form. I think that the invisible reCAPTCHA makes its decisions on whether to challenge or not based on this observation.

Completely automate obtaining access tokens

I'm currently manually obtaining a new page access token every time the current token in use is about to expire - see this repository's README. This is because the process relies on a manual authentication.

I believe that it's possible to completely automate the process following the steps shown in the Facebook::OpenGraph synopsis.

Investigate further and if possible implement this automation.

If hyperlinks are stored in the database as true hyperlinks they won't work with the application

Various hyperlink attributes defined in the Django model. This causes the Django admin interface to insist on a full URL or if not it will convert the entered image to a full URL. However the application then interpolates a URL around the information given, which leads to a broken link. So, I'm having to store partial URLs (domain and path) only in the database, which then doesn't work with the Django admin interface.