Quantcast
Channel: feed2exec:115cd87edd3e8f10d6619f2e7992cc60d6fd2b56 commits
Browsing latest articles
Browse All 40 View Live

add breaking test for debian-planet

View Article



return bytes properly in fetch

this fixes double-encoding issues detected in the previous test. our documentation was *saying* we were returning bytes from fetch() but, actually, we were returning a string. it *seems* that...

View Article

expand tests documentation and add plugin design docs

View Article

expand on the use of vcr

View Article

refactor design docs more

View Article


silence user-agent variation in planet test

View Article

prepare 0.8.0 changelog

View Article

fix path to bash completion

View Article


properly install bash completion in debian package

View Article


package should be arch: all, like other python packages, not any

View Article

Merge tag '0.8.0' into betamax

fixed encoding issues, folder support in OPML, bash completion

View Article

refactor feeds to remove global

we now use a class static session that can be used from plugins this slightly simplifies the implementation of plugins as we do not need to load the module at a specific point to get the right...

View Article

update test suite documentation since switch to betamax

View Article


Merge branch 'betamax'

View Article

add requests-file dependency

get rid of our custom wrapper code around requests to handle file:// resources: that is handled by a third-party library and removes about 40 lines of code we didn't need this makes session...

View Article


use getter/setters for class session

this makes the code cleaner and more idiomatic this does require plugins to use the private member to access the session, especially for the test suite, but there are ways of avoiding that for...

View Article

add badges showing debian versions

View Article


add license and version badges

View Article

reorder feed class hierarchy

the feed fetch functionality does not need to be in the "storage" hierarchy...

View Article

move opml import in the FeedStorage abstract class

this makes more sense than having it as a standalone function. also, it is not storage specific, so let's make a API-facing class that will expose more generic functions.

View Article

move OPML export facilities along with import

View Article


cosmetic: move down sqlite class with other sqlite storage

View Article


split --config and --database settings, using xdg module

the xdg module saves us some code and also extends configuration to support global config files like /etc or /var/cache (in theory - in practice, the freedesktop standards only use ~/.cache). using...

View Article

fix some new flake8 nitpicking

View Article

prepare 0.9.0

View Article


add verbose message removed by mistake

View Article

initialize FeedFetcher session singleton properly

without this, the wayback plugin would fail because it FeedFetcher._session is None. doing this in the constructor makes the accessor faster as well, as we don't need to do another check. this should...

View Article

properly call FeedFetcher constructor

View Article

API refactoring: merge classes in a controller

The FeedStorage and FeedFetcher class were fairly similar: they were wrappers around the storage subsystem, which is basically a "model", so they are a "controller" in the classic sense. So merge them...

View Article



API refactoring: create a Feed class

A "Feed" is the basic representation of a RSS/Atom feed. It derives from the feedparser dict but we use our class to regroup functionality that was loose around the module, like normalize() and...

View Article

security: avoid possible config setting override

It may be possible for parsed feed data to override configuration that is passed to plugins and other components. Normally, feedparser doesn't send those settings (e.g. output or args) that could...

View Article

cosmetic: consistent test comparison

We should do "expected == actual" because that is the way the diff is ordered on display. That's a rather opiniated way of doing things, but at least we should be consistent across the board, which we...

View Article

more possible naming conflicts

View Article


add transmission plugin

this plugin is introduced because the exec plugin cannot safely guess a subdirectory to save feed into without resorting to using plugins. now the transmission plugin has nicer defaults and properly...

View Article

fix some feeds descriptions in readme

View Article

move fetch functionality to Feed

this allows to make for simpler tests, because we can easily fetch and parse a single object instead of having to run over a full config furthermore, it is more logical for plugins: they get passed a...

View Article


move catchup down into plugins

output plugins are now actually called when catchup is used. responsability to skip slow and write operations are delegated to plugins, which makes them easier to develop and test

View Article


make test_plugins run standalone

View Article

expand plugin documentation to include catchup

View Article

mention the catchup parameter in docs

View Article
Browsing latest articles
Browse All 40 View Live




Latest Images