jodal / comics Goto Github PK

Feeds for sets with many comics can be so slow (>10s response time) that e.g. Netvibes fails to fetch the feed.

Reported 2010-04-07 by Berge

The feeds should be updated to only work given a personal secret key in the feed URL.

Misfile (http://www.misfile.com/) requested by Alex ([email protected]) at 2010-03-05

The latest release of a comic is found based on max id instead of max pub_date, which isn't always correct, though it is fast and generally works.

The offending code is in comics.core.utils.comic_releases.get_latest_releases().

See http://www.pythonware.com/library/pil/handbook/image.htm#Image.verify or http://stackoverflow.com/questions/4780424/detecting-corrupted-images-in-bash-script for details.

Would be nice with a link to the specific comic aggregated and not just the comic site.
Example:
When displaying XKCD strip 1286 then link to http://xkcd.com/1286/ and not just http://xkcd.com/.

Also, is it really necessary to redirect through the aggregator when clicking on the link to the comic site? What is the gain? It makes it (slightly) harder to share the link to the comic with friends who don't have an account for the comic aggregator.

The natural action once a page of comics has been read is to go to the next. If the images where preloaded in the background while you read the page load would be incredibly faster. That would be like way cool.... They don't even need to ever be shown as just loading them in the background will cache them in the browser.

"Havet" from http://nettserier.no/havet requested by @sveinhal at Twitter at 2010-05-14.

"Scenes from a Multiverse" from http://amultiverse.com/ at 2010-07-01.

When it shows up with '20 new releases', it can be a chore finding the correct place to start reading. It could be solved by reading from the top and downwards, but then you could encounter multi-issue stories in the wrong order.

It would be nice for the new page load to start out scrolled to the oldest unread comic, so you could just read from there 'till the top.

Consider adding a management command for finding and removing image files in which is not referenced in the database.

http://satwcomic.com/

Requested by Berge

Look into if switching is worth it.

"Axe Cop" from http://axecop.com/ requested at 2010-06-06.

For XKCD the "alt" (or mouseover) text is displayed alongside the comic.
Certain other comics, such as "Scandinavia and the World" and "Nerf Now" also got some descriptive text for the comic, it would be nice if this text also is displayed next to the comic.

When user sets (#26) have been implemented, set editing should be improved:

• The "Automatically add new comics to the set" and "Hide comics without matching releases from view" settings can be moved to the account editing page.
• Adding to or removing from a comics set should be possible to do with a small add/remove widget in any comic view.
  • The static widget should just be a stupid toggle, which doesn't change when the comic is or is not in the user's set. This is to make the page more cacheable.
  • Using JavaScript, the widget should be improved to vary with the comic being in the user's set or not.

Given the following scenario:

• The latest XKCD comic is from yesterday, 2011-11-21.
• I'm at XKCD's "latest view" at /c/xkcd/ and see the release from yesterday.
• I press "Previous".

I would expect to get to the release before the one I was just viewing, e.g. the latest release from before yesterday, e.g. from 2011-11-18.

In reality, I arrive at /c/xkcd/2011/11/21/ and see the exact same release. If I press "Previous" once more, I get to the release I wanted to get to in the first place: /c/xkcd/2011/11/18/

http://buttersafe.com/

http://timetrabble.com/

Firekanta from http://www.kantenstrek.no/galleri requested by John Roger Taraldsen ([email protected]) at 2010-04-30.

Bug comic (http://www.bugcomic.com/) requested at 2010-07-03 by [email protected].

The sample apache vhost in the documentation contains the line "Allow from all". In Apache 2.4 this will not work and should be replaced by "Require all granted".

See http://stackoverflow.com/questions/18392741/apache2-ah01630-client-denied-by-server-configuration for more information.

I imagine that 10.04 fixes this, but in the documentation for 9.04, installing South from pip doesn't pull in Python's setuptools unless you already have them. You'll get lines that look like this:

Downloading/unpacking South
  Downloading South-0.6.2.tar.gz (51Kb): 51Kb downloaded
  Running setup.py egg_info for package South
    Traceback (most recent call last):
      File "<string>", line 3, in <module>
    ImportError: No module named setuptools.command
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):  

  File "<string>", line 3, in <module>

ImportError: No module named setuptools.command

----------------------------------------
Command python setup.py egg_info failed with error code 1 
Storing complete log in ./pip-log.txt

Solution is to add python-setuptools as a dependency in the 9.04 install line. Python-pip isn't pulling it in if you have python-pkg-resources already installed, though the above package seems to not want "either-or."

Consider allowing an existing user to connect his account with his Google/Twitter/Facebook account, and to authenticate using this external account.

This may not be necessary, given long enough validity of the session cookie.

Intel's Insides from http://www.intelsinsides.com/ requested by Norvald at 2010-06-25.

Add a form to the top of the user's set page which allows the user to import a set of comics from any old named set.

The axecop crawler is no longer working after they have changed the feed. It seems like the full resolution image is no longer in the feed.

A REST API for accessing comics, releases, images and a user's set have been requested by users multiple times.

One should consider using one of:

• http://django-rest-framework.org/
• http://django-tastypie.readthedocs.org/

The API should probably have a service for authenticating using a user's username and password. This service will return the user's secret key. This key should be used to authenticate the client on all further requests, and may be stored by the client for future usage. A user should be able to reset the secret key from his account management page.

Reimplement comics sets to be associated with users instead of names.

• The old front page should be removed.
• The old set views at /s/setname should be the new front page.

http://www.menagea3.net/strips-ma3/

E.g. Darths & Drois has a huge formatted text associated with each comic. Since these texts often are half the fun, comics should support fetching larger pieces of text with formatting, and keep a sane amount of this formatting, e.g. headers and bullet lists.

This feature was removed when moving from sets to "my comics" in v2.0. @adamcik has requested that the feature is reintroduced.

http://www.collectedcurios.com/sequentialart.php

http://www.hejibits.com/

http://www.icanbarelydraw.com/

http://www.redmeat.com/redmeat/current/index.html

Pretty please :)

http://nichtlustig.de/

The all comics list at /account should indicate which comics I have already added.

When running ./manage.py to set up the local sqlite3 database, it's created with the permissions of the user running manage.py. In basically all setups, this won't be the user of the webserver.

This doesn't manifest itself typically at first (the comics application will have read access to the database via default global-read file permissions); it showed up for me when I tried to create a new comics set. This was the first time that it tried to actually write to the database.

I'm using Apache as my webserver: the solution for me was to move the sqlite3 database into its own directory ("database" in my case), giving ownership / write permissions to the webserver group for the database directory as well as the file.

The problem is that this just manifests itself as a 500 in production mode (or a more-helpful traceback in dev mode); a test for good permissions on the database file and a helpful message / debugging steps would be more useful.

Another solution is to add a couple steps in the documentation that fix the permissions ownership problem.

Add the option of turning off open user registration and require new users to have an invite from an existing user.

http://www.optipess.com/

http://thepunchlineismachismo.com/

During crawling, title and/or text with quotes are not handled correctly. This is the reason e.g. the XKCD comment is often missing.

Feature request from Einar Ryeng:

Add a text area for each comic for entering the text of the comic into. The comic texts should be searchable.

Some comics are released as tif, in this case it should probably be converted to a web friendly format. Comics should also probably have a whitelist of accepted web formats.

Consider using Celery instead of cronjobs for crawling new comics.

http://beyondthetree.wordpress.com/

The size of comic images can be reduced without loosing any quality by pushing them through lossless optimization, like e.g. Yahoo's www.smushit.com can do for you. Smushing comics would save a lot of disk space and reduce page load time.

Some work have already been done by @adamcik on a tool for smushing the images in comics' archive. See the file bin/smush in the repo for details.

I miss it :(

"Not invented here" from http://notinventedhe.re/ requested 2010-03-17