youngj / envaya Goto Github PK

should either be added to UploadTest, or convert ReportingTest to use *chrome

On the http://envaya.org/org/feed page, allow visitors to load previous updates (currently you can only see the latest 20)

Google Translate sometimes is unable to translate certain content (often, very long regions of text). Currently Envaya will show the original language text to the user, but we should probably display some sort of error message.

Over time reporting forms will change, and defining the forms in code (as in engine/reporthandler/fcs/narrative.php) will become difficult to manage. Form definitions should be stored in the database.

Some reports require periodic responses using the same template, but the current system only lets an organization submit one response using a particular template (although the grantmaker can manually create a duplicate form).

currently the order of pages in menu is predefined and any new custom pages appear at the end of the list.

the alignment of different size photos on http://envaya.org/envaya/team looks weird

When people upload many images at once, they currently show up as individual news updates. Instead they should go under a photos tab, with a UI specific for browsing photos.

There should be a selenium test that creates an email template and sends it to envaya users.

sometimes people accidentally register the same organization twice... there should be code to merge their content into one organization

it should tell people to use a standard version on a regular browser. too much javascript and typing for a mobile phone.

if somebody adds a document, it should say document, etc.

see http://www.scribd.com/developers/api?method_name=Javascript+API

Many Envaya users have their images within Word documents and are confused when they can't just upload the Word document to attach the image. Instructing users to copy the image into something like MS Paint and save as a .jpg is too complicated. We should find a way to extract images from .doc files and either use the first image from the document or let them choose from all the images.

view filenames have inconsistent naming conventions (some camel case, some with underscores). also the engine/views/default/org directory isn't organized very well.

see if Selenium's attachFile method in *chrome mode works to test uploading photos

we should be able to do ip geolocation without relying on external service using data from http://www.countryipblocks.net/e_country_data/Africa_cidr.txt

Feedback email pointed out that it is not obvious how to sign up (since the only link is on the home page):

"Nilipokuwa Dodoma Novemba mwaka huu nikakutana na Erasmo Tulo, Meneja wa
Shirika la Dira Theatre linalofanya kazi kama za Shirika langu. Huyu alipa
kadi yake ya biashara na kwa maajabu nikaona ana tovuti (Envaya)! siku hiyo
hiyo niliifungua tovuti yake na kuijionea kazi zake. Nikamuuliza akanipa
habari kiasi za envaya. Sikutilia maanani sana hasa kwa kuwa kila
nilipotaka kuingia nilitakiwa neno la utambulisho na namba ya siri."

translation in English:
"When I was in Dodoma this November, I met with Erasmo Tulo, manager of
the Dira Theatre organization which does work like mine. He gave me
his business card and with wonder I saw he had a website (Envaya)!
That day I opened his website and saw his work for myself. I asked him
and he gave me some information about Envaya. I didn't understand why
each time when I tried to log in I was asked for a username and
password. "

there are currently no tests for this

the back/forward buttons on firefox for mac have a very small target area

several organizations have created multiple accounts for their organization... we should detect probable duplicates and encourage them to sign in using their organization's existing account

if you click to the next image before it finishes preloading, it never shows up

there should be a selenium test for this. either need to create a bunch of initial feed items in scripts/install_test_data.php so that the button appears, or create the feed items during the test

the TCI iframe should be invisible, but it isn't

Currently the entire grid is encoded in json inside a csv field, which is not very friendly. Probably each row/column of the grid should have its own column in the csv.

and allow users to restore old versions

the setup scripts use ubuntu, but nobody has run envaya's selenium test scripts on ubuntu yet

should display facebook/twitter/donate buttons prominently.

photo region should be closer to 4x3 so it is easier to find other photos that work there. shouldn't be bigger than large size for image uploads so we can use user uploaded photos there.

UploadTest currently only tests a custom logo, but we should also test uploading a large header image.

Would be necessary for selenium to use different config settings for the http server. Also would simplify things for everything to be started by runserver.php

Many Envaya users have a vast amount of content in Microsoft Word files, and often want to publish that on their website.

People can already copy and paste content into the Envaya tinymce editor, but this doesn't work for documents with images or with complex layouts.

However there are some problems with allowing CSOs to upload Word documents in the normal way: 1) readers would need Word (or Word Viewer, OpenOffice, etc.) on their computers in order to access the information; 2) the barrier to downloading files is much larger than the barrier of opening a webpage; 3) computer viruses are especially common in Tanzania and we don't want to facilitate them spreading.

Current plan is to integrate Envaya with Scribd's API, allowing documents to be displayed inside a normal webpage without any required software or any risk of spreading viruses.

When the browser width is very small, the header should wrap to multiple lines, but the background image doesn't line up correctly

There should be selenium tests for creating/editing/activating featured organizations.

there should be a selenium test for the google map on http://envaya.org/org/browse

Sometimes users apparently try to upload images which should be stored on S3, but a broken image icon appears instead.

http://envaya.org/dhahabu/post/6590 has an empty img tag:

That seems like a bug with the tinymce editor, but there could also be other cases where the S3 upload fails.

Currently users are limited in the UI to creating a fixed set of pages (home, news, history, projects, team, reports, contact). They should be able to create other basic pages too (similar to the history or projects pages).

should be selenium test for this (although won't be able to test recaptcha)

the language links at the bottom of the page look funky

as a global variable it needs to be imported whenever it is used, which is tedious. it should probably be a static class instead.

also, the codebase intermixes 'config' and 'settings' terms inconsistently.

there are currently no tests for this.

In particular:

• editing functionality

Also disable JavaScript during the selenium tests if possible.

There should be helpful error messages that explain why uploading failed

http://wiki.nginx.org/NginxHttpProxyModule#proxy_cache looks like it would work

currently they can choose the address (widget_name) when the page is first created, but changing the address requires deleting the page and creating a new one.

should be a selenium test for this

each photo should have a link to the organization it comes from... also, create UI for envaya admins to select photos that go in the slideshow

There are currently no selenium tests for the reporting system.

on the mobile version of e.g. http://envaya.org/org/feed , the drop down filters dominate the screen on small devices, so should probably move them to a separate page and make users click a link to change the filters.