- For consistency, functions should always return null on non-existing data.
- WordPressPluginUpdateBridge appears to have used its own cache instance in the past. Obviously not used anymore.
- Since $key can be anything, the cache implementation must ensure to assign the related data reliably; most commonly by serializing and hashing the key in an appropriate way.
- Even though the default path for storage is perfectly fine, some people may want to use a different location. This is an example how a cache implementation is responsible for its requirements.
* [BingSearch] Make the bridge compatible with PHP 5.6
The use of isset() with an expression is not possible in PHP 5.6. I
fixed it by replacing isset() with "null !== ".
This seems like a weird bug somewhere.
Either the HTML parser should return the valid page, or the CMS should
not convert the URL first, or the URL validation regex is buggy.
* [AutoJMBridge] Fix bridge after website change
The website was totally reworked, so the bridge had to be reworked too.
The bridge parameters changed, therefore old RSS feed will not work
anymore, but it was impossible to do it in another way.
In case of a unparsable date, the text to DateTime object failed, and
this resulted to a Fatal error while using this DateTime object . To
prvent this fatal error, if the date parsing failse, then a DateTime
object is created with the actual date.
* rework to parse new layout
* skip incomplete rows
The last row could have fewer columns if there are less rows than the items limit. This usually should not happen, though.
* use constant for skipping
Bandcamp tags pages have a new layout and now use a POST API endpoint to view each page of releases.
Output of this bridge should be almost the same as before, with a few small improvements:
- Small album image in 'content', larger album image in 'enclosures'
- RSS item titles/authors are appended with the releaser in parentheses if the artist name and the releaser are different (i.e. Record Label's Bandcamp releases an album called Bar by the band named Foo, it would get the title 'Foo - Bar (Record Label)' and the author 'Foo (Record Label)')
* Fix issue with CachetAPI Pagination
Fixing issue that only the oldest 20 entries were shown.
_Background:_
_Cachet has a, lets call it odd, system of pagination. On the first page you see the incidents first created, so they are not what you want to see. But on the last page you can have 1 or 20 of the newest incidents. So you have to take the incidents from the last page (call it Pmax) and combine them with the incidents from Pmax - 1._
This commit adds support for a new parameter which specifies the type
of cache to use for caching. It is specified in config.ini.php:
[cache]
type = "..."
Currently only one type of cache is supported (see /caches). All uses
of 'FileCache' were replaced by this configuration option.
Note: Caching currently depends on files and folders (due to FileCache).
Experience may vary depending on the selected cache type. For now always
check if FileCache is working before testing alternative types.
References #1000
In item content, the image is now a clickable link to the post;
In item title a ▶ is prepended if the post contains a video; it's impossible to tell from the content image.
Images in Twitter feeds are currently being scaled by adding ':orig'
(original image) and ':thumb' (thumbnail) to image URIs in the feed.
This can cause issues with feed readers that don't handle colons in
URIs correctly.
Image scaling can now be disabled by adding '&noimgscaling=on' to the
query. This parameter is optional to stay compatible to existing feeds.
References #957
1. Author Name now doesn't include Episode Title
2. It now fetches Episode Creation Timestamp, to allow correct sorting in podcatchers
3. Description is now the actual show notes, and not an <audio> tag
* core: Add bridge parameter auto detection
This adds a new 'detect' action which accepts a URL from which an
appropriate bridge is selected and relevant parameters are extracted.
The user is then automatically redirected to the selected bridge.
For example to get a feed from: https://twitter.com/search?q=%23rss-bridge
we could send a request to:
'/?action=detect&format=Atom&url=twitter.com/search%3Fq%3D%2523rss-bridge'
which would redirect to:
'/?action=display&q=%23rss-bridge&bridge=Twitter&format=Atom'.
This auto detection happens on a per-bridge basis, so a new function
'detectParameters' is added to BridgeInterface which bridges may implement.
It takes a URL for an argument and returns a list of parameters that were
extracted, or null if the URL isn't relevant for the bridge.
* [TwitterBridge] Add parameter auto detection
* [BridgeAbstract] Add generic parameter detection
This adds generic "paramater detection" for bridges that don't have any
parameters defined. If the queried URL matches the URI defined in the
bridge (ignoring https://, www. and trailing /) an emtpy list of parameters is
returned.
* [ZoneTelechargementBridge] Switch to the new Website
The website zone-telechargement1.org decided that he will be using a new
domain at the end of november :
https://www.annuaire-telechargement.com/
The bridge uses the new domain but still uses the same filename and
class name to keep the existing feed working.
Replaces 'debugMessage' by specialized debug function 'Debug::log'.
This function takes the same arguments as the previous 'debugMessage'.
A separate Debug class allows for further optimization and separation
of concern.
This commit adds filters to remove embedded videos and view counts from
all posts. This doesn't remove the preview image for videos, which are
embedded separately.
Hidden elements are used for error conditions and generally made
visible using JavaScript. Since RSS-Bridge doesn't support JS, these
error messages are shown in the final feed. For example:
"It looks like you may be having problems playing this video. If so,
please try restarting your browser."
This commit removes all hidden elements to prevent error messages being
added to the feed.
- "It looks like you may be having problems playing this video. If so,
please try restarting your browser."
FB includes origin information (i.e. "YOUTUBE.COM") as well as
descriptions with embedded media (images and video).
These details are currently being removed by the bridge.
This commit changes implementation to only remove origin information
and keep the media description in place. The media description consists
of two elements - title and description. The title provided by FB is
included in an anchor, which gets replaced by a paragraph with the
same contents to improve readability.
References #912
Move CACHE_DIR from index.php to /lib/RssBridge.php and change name
to PATH_CACHE.
PATH_CACHE is one of the core paths of RSS-Bridge and should therefore
be defined in the core file RssBridge.php.
`defaultLinkTo` makes anchors point to the correct path which broke
parsing because it expected href to start with `#gistcomment`.
This commit changes the implementation to make `defaultLinkTo` point
to the correct page (using `getURI` instead of `self::URI`) and search
with `*=` instead of `^=`.
Instagram returns "HTTP/1.1 301 Moved Permanently" on each request
to "https://instagram.com/" because the correct location is
"https://www.instagram.com/".
Instagram will respond with "HTTP/1.1 301 Moved Permanently" if the
URI for the requested user doesn't end with a slash.
Notice: This is only a minor enhancement to prevent error 301 from
happening. The previous version worked fine as is.
Using the classical www.zone-telechargement1.org as base URL, the bridge will
always be redirected to the actual wwX.zone-telechargement1.org final URL. This
makes the bridge more robust to URL changes.
This commit collects the original contents from a different
tag to prevent this issue. The root cause is unknown but closely
related to the regex.
References #877
The function 'defaultLinkTo' applied to the source HTML does break
regex matches later in the bridge. We need to apply the function
right before adding the contents to the item for the bridge to work
properly.
References #856
- Base URL updated
- Show name has different styles on the Website, use another way to get the show name
- Entry URIs are now unique to make sure RSS readers don't treat episodes as duplicates
- No more new lines in the feed or item title
- Entry URIs are unique to make sure RSS readers don't treat episodes as duplicates
- Titles are unique to make sure RSS readers don't treat streams and downloads as duplicates
The bridge has been expanded to better cover the whole GQ magazine.
It should support all countries (provided they all use the same absurdly shitty publication system).
It is guaranteed to be only tested with sexactu articles (that I now obtain by loading Maïa Mazaurette author page).
This commit adds a new optional parameter 'limit' which can be used
to limit the number of items returned by this bridge (i.e. '&limit=10')
As requested in #669
The URI "https://facebook.com/username?_fb_noscript=1" returns two
posts per user. Some profiles, however, are very active, causing the
bridge to miss items if more than two posts are send within the cache
duration (5 minutes).
The alternative suggested in #669 is to use a different URI:
"https://facebook.com/pg/username/posts?_fb_noscript=1"
While the contents of this URI essentially look the same when viewed
in a browser, it actually returns more than 10 posts depending on the
profile.
References #669
* Fix typo with status field
* Comply with other bridges
Change the uri element of an item to point, not on the magnet link, but on the page, as similar bridges do.
* Improved to return name & uri matching with query
This change makes it possible for the feed reader to discover a title and url consistent with the user's search.
* [ZoneTelechargementBridge] Added Bridge for ww2.zone-telechargement1.org
Goal for this bridge is to follow the episode publication of a TV show
season while it's broadcasted on the TV.
* [AutoJMBridge] New Bridge
This bridge will show all the car offers AutoJM has for the model you
choosed and using your filter. Very useful to wait for a cheap price for
a new car !
This commit fixes#782 by updating the parameter value of 'Maison &
Jardin', but this means the user has to update his RSS Feed URL (.because
of the bridge structure, it would be a nightmare to fix it in another
way)
This commits add all the categories available on Dealabs Website.
* [ExtremeDownloadBridge] Added Bridge for ww1.extreme-d0wn.com
Goal for this bridge is to follow the episode publication of a TV show season
while it's broadcasted on the TV.
* Adding a bridge for theyetee.com
The bridge fetches daily shirts at theyetee.com.
The Yetee offers two new shirts each day, but you can buy them only for a few hours !
Unfortunately, the site don't provide RSS feed, so the only way to keep up to date on new shirt is their daily mailing ... until now !
* Debug mode improvements
- Improve debug warning message
- Restore error reporting in debug mode
- Fix 'notice' messages for unset fields
* Add parsing utility functions
html.php
- extractFromDelimiters
- stripWithDelimiters
- stripRecursiveHTMLSection
- markdownToHtml (partial)
bridges
- remove now-duplicate functions
- call functions from html.php instead
* [Anidex] New bridge
Anime torrent tracker
* [Anime-Ultime] Restore thumbnail
* [CNET] Recreate bridge
Full rewrite as the previous one was broken
* [Dilbert] Minor URI fix
Use new self::URI property
* [EstCeQuonMetEnProd] Fix content extraction
Bridge was broken
* [Facebook] Fix "SpSonsSoriSsés" label
... which was taking space in item title
* [Futura-Sciences] Use HTTPS, More cleanup
Use HTTPS as FS now offer HTTPS
Clean additional useless HTML elements
* [GBATemp] Multiple fixes
- Fix categories: missing "break" statements
- Restore thumbnail as enclosure
- Fix date extraction
- Fix user blog post extraction
- Use getSimpleHTMLDOMCached
* [JapanExpo] Fix bridge, HTTPS, thumbnails
- Fix getSimpleHTMLDOMCached call
- Upgrade to HTTPS as JE now offers HTTPS
- Restore thumbnails as enclosures
* [LeMondeInformatique] Fix bridge, HTTPS
- Upgrade to HTTPS as LMI now offers HTTPS
- Restore thumbnails using small images
- Fix content extraction
- Fix text encoding issue
* [Nextgov] Fix content extraction
- Restore thumbnail and use small image
- Field extraction fixes
* [NextInpact] Add categories and filtering by type
- Offer all RSS feeds
- Allow filtering by article type
- Implement extraction for brief articles
- Remove article limit, many brief articles are publied all at once
* [NyaaTorrents] New bridge
Anime torrent tracker
* [Releases3DS] Cache content, restore thumbnail
- Use getSimpleHTMLDOMCached
- Restore thumbnail as enclosure
* [TheHackerNews] Fix bridge
- Fix content extraction including article body
- Restore thumbnail as enclosure
* [WeLiveSecurity] HTTPS, Fix content extraction
- Upgrade to HTTPS as WLS now offers HTTPS
- Fix content extraction including article body
* [WordPress] Reduce timeout, more content selectors
- Reduce timeout to use default one (1h)
- Add new content selector (articleBody)
- Find thumbnail and set as enclosure
- Fix <script> cleanup
* [YGGTorrent] Increase limit, use cache
- Increase item limit as uploads are very frequent
- Use getSimpleHTMLDOMCached
* [ZDNet] Rewrite with FeedExpander
- Upgrade to HTTPS as ZD now offers HTTPS
- Use FeedExpander for secondary fields
- Fix content extraction for article body
* [Main] Handle MIME type for enclosures
Many feed readers will ignore enclosures (e.g. thumbnails) with no MIME type. This commit adds automatic MIME type detection based on file extension (which may be inaccurate but is the only way without fetching the content).
One can force enclosure type using #.ext anchor (hacky, needs improving)
* [FeedExpander] Improve field extraction
- Add support for passing enclosures
- Improve author and uri extraction
- Fix 'notice' PHP error messages
* [Pull] Coding style fixes for #802
* [Pull] Implementing changes for #802
- Fix coding style issues with str append
- Remove useless CACHE_TIMEOUT
- Use count() instead of $limit
- Use defaultLinkTo() + handle strings
- Use http_build_query()
- Fix missing </em>
- Remove error_reporting(0)
- warning CSS (@LogMANOriginal)
- Fix typo in FeedExpander comment
* [Main] More documentation for markdownToHtml
See #802 for more details
The bridge would generate empty titles if the content is longer than
50 characters, but doesn't have further spaces in it. With this commit
the title is correctly generated based on the contents, taking missing
spaces into account.
References #786
https://cad-comic.com/ now provides feeds at
- https://cad-comic.com/feed (rss)
- https://cad-comic.com/feed/atom (atom)
Thus multiple alternatives are available to choose from, making this
bridge obsolete:
- FilterBridge (using one of the feeds above)
- WordPressBridge (on the main site)
- One of the two available feeds
References #752
This commit fixes an issue caused by self closing tags not supported
by simplehtmldom (<source>).
Adds a monkey patch to extend simplehtmldom with the ability to detect
that particular tag. Most of the code added is copied directly from
simplehtmldom (see vendor/simplehtmldom) with adjustments to account
for RSS-Bridge formatting.
Related to: https://sourceforge.net/p/simplehtmldom/bugs/83/
Notice: The tag itself is valid according to Mozilla:
The HTML <picture> element serves as a container for zero or more
<source> elements and one <img> element to provide versions of an
image for different display device scenarios. The browser will
consider each of the child <source> elements and select one
corresponding to the best match found; if no matches are found
among the <source> elements, the file specified by the <img>
element's src attribute is selected. The selected image is then
presented in the space occupied by the <img> element.
-- https://developer.mozilla.org/en-US/docs/Web/HTML/Element/picture
References #753
* use defaultLinkTo
* remove duplicate video links
* remove line ending before "Reposted" label
* return newline before reposted string
* remove comments
* use video links that won't require login
* set title if video has no title
Adds a new option '&title_from_content=on' to build the title for feed
items from the feeds content. The title is generated from the first
whitespace after 50 characters of the content or the entire content if
the total size is lower than 50 characters.
References #587
This commit fixes a few things related to static::URI
1) Remove trailing slash from the URI to simplify using 'defaultLinkTo'
2) Use static::URI instead of self::URI for consistency
3) Remove custom implementation of 'defaultLinkTo'
Images are collected for each post and added to enclosures. Images or
animtions from lh3.googleusercontent.com are specifically handled in
order to return the animated version of the gif and the original sized
image (this is normally taken care of by JS in the browser).
Adds a new bridge for https://gist.github.com
The bridge generates feeds for comments on a particular gist based on
the gist ID or full URI. For better readability the general behavior
of code sections is manually restored with the original CSS styles
from GitHub.
New bridge for Skimfeed: https://skimfeed.com
Generates feeds for all features of Skimfeed:
- News (the ones displayed on the front page)
- Hot topics ("What's Hot" section on the front page)
- Tech news (preconfigured feeds in the menu bar)
- Custom feeds (using the configuration system of Skimfeed), see
https://skimfeed.com/custom.php
The number of items returned by the bridge can be limited for all
categories ('&limit=...'). This parameter is optional, all categories
are unlimited by default!
Authors are added with HTML anchors in order to allow quick navigation
to source channels.
The bridge ships with developer tools to auto-generate lists in the
future (especially useful for 'Tech news'!)
References #748
Following changes in the JSON data and selecting images for the
content (320x240 or bigger) and enclosure (largest version). All of
the data is now extracted from the JSON data instead of parsing the
DOM.
References #754
