0xe1f/grr

Lightweight Replacement for Google Reader

grr's Issues

How to create account

After following all the steps in the README.md file, I find myself unable to login.

Specifically:
I have no idea how to use the "admin_secret" variable. I tried logging in using "admin", "root" and empty usernames, and the secret as password, but nothing worked, it keeps saying "Incorrect username or password".

Then I decided to resort to openid.
Unfortunately I can't use my regular OpenID, but am forced to choose between Google and Yahoo.
I chose Google, chose my google account, gave permission to know my email address, and then I was redirected back at the grr login page.
But I still wasn't logged in.

All these tests have been performed with:
create_unknown_accounts true.
login_throttling_window false
admin_secret invented_password
role_user user
role_admin admin

Am I missing something?

Thanks!

schema_full.mysql missing changes from V9 and v10 files

I'm not sure if this is a bug or not. I thought I add this here just in case others are having problems using grr initially. I'm using commit: 0394a73

When installing a fresh grr and using schema_full.mysql to setup the database two changes are missing.

1. the new columns introduced in schema_upgrade_v9.mysql, which cause reading/staring not to work while going thru feeds. You will see errors like this in error_log:
   PHP Fatal error: Call to a member function bind_param() on a non-object in /path/to/grr/classes/MySqlStorage.php on line 1715
2. the column change in schema_upgrade_v10.mysql which cause the admin page to give 500 error. You will also see errors like this in error_log:
   PHP Fatal error: Call to a member function execute() on a non-object in /path/to/grr/classes/MySqlStorage.php on line 991

To fix both just run mentioned files in your database.

Debian packaging

Hi!

Is there any plan to create a Debian package?

If you do not want to do this do you mind if I try to add this package to the Debian?

Rewrite the background updater in Go

Rewrite the feed loader/updater in Go, with multithreading support

Hiding read feeds

Is there any way (even if I need to modify source code) to only show the feeds that have unread items?

Create an installation wizard

Create an installer wizard which will guide the user through the initial configuration, database setup and administrative user creation.

Wizard will be initiated when the application is first accessed, and must confirm that the user has access to the file system.

Examples: WordPress, Gallery.

Remove unused feeds

When all users have unsubscribed from a feed, crawler will still continue to index and records for articles still exist in database.

At beginning of each background crawl process, remove feeds and their corresponding articles if no longer used by any users.

Handling feeds containing articles with no publish date

Some ancient feeds (RSS1.0/RDF) appear to sometimes have articles with no publication date. See - http://www.gentoo.org/rdf/en/gentoo-news.rdf

I've coded up an enhancement to the RssParser that looks up the guid of an article if there is no publication date. If it finds an existing article that has already been parsed with the same guid, then it sets the publish date timestamp to the found article's publication date within the RSS parser. This stops it being reset everytime (which is the current problem) and meaning the same articles appear again at the top of your reader list, even though they are read articles.

This requires access to the 'db' attribute/variable to make a prepared statement to do the select enquiry on the database within the RssParser class. As it doesn't have access to the private Crawler 'db' attribute, I'm currently hacking it by using $GLOABLS['db'] and having the Crawler class set this after successfully connecting to MySQL

Would be interested to know if you have a better approach to this given your current code structure. As by the time it exits the parser, the articles have a default publish date of the current timestamp, if nothing else sets it.

My 'hack' is working and doing its job now, so the articles don't keep being reassigned a new publication date. But before submitting a pull request, thought I'd run it past you first.

Name

I think the project could use a new name: "grr" is already Günther's Rss Reader, which has been using the name since 2005 and has a most recent release about a year ago. I noticed because I was trying to find your project, and both "grr rss" and "grr reader" searches on Google kept pulling up the wrong one.