Vienna 2.5 BETA released

[Michael Ströck]

Click to download: https://sourceforge.net/projects/vienna-rss/files/TestVersions/2.5.0.2500/Vienna2.5.0.2500-BETA.zip/download

Requirements
Vienna 2.5 BETA requires Mac OS X 10.5 Leopard or higher to run.

Pre-release software
The Vienna 2.5 BETA version is pre-release software, and it may contain bugs. If you discover a bug, please create a new topic here in the Vienna forum and report the bug to us. Some tips for writing good bug reports can be found at http://forums.cocoaforge.com/viewtopic.php?f=18&t=9210. For safety, we recommend that you back up your files before running the beta version. Vienna's files are located at "~/Library/Preferences/uk.co.opencommunity.vienna2.plist" and "~/Library/Application Support/Vienna". When 2.5 goes out of beta, Vienna will automatically update to the release version.

Changes
- Add support for plugins. For details, see: http://www.vienna-rss.org/?page_id=120
- Add a facility for automatic URL shortening via bit.ly that can be used by all plugins.
- Add "Share with Twitter" plugin and toolbar button.
- Add "Share with Facebook" plugin and toolbar button.
- Add "Share with Evernote" plugin and toolbar button.
- Web plugins (especially audio- and video-players) now continue to operate while in the background.
- Fix bug that caused Vienna to become unresponsive after deleting a feed in Unified view.
- Update to the current version of Sparkle, which prevents auto-updating to a version of Vienna which will not run on the user's system.
- Fix bug where deleting a feed in Unified view mode would cause Vienna to stop working correctly.

[cortig]

Hi All,

I noticed a couple of things with the new beta:
- New plugins can be installed in ~/Library/Application Support/Vienna
- the link plugins parse a raw URL. There is no way (that I found) to parse the URL with encoded text (for crazy links using a lot of &, ?, etc). I wanted to create a plugin to share with Papers (mekentosj.com). I can use a papers://url/ link and get the link plugin to add the URL, but unless the URL has encoded text, many parsed URL are truncated in Papers.
I therefore created a Script plugin based on Mike Dimicco‘s NNW/Vienna to Papers script and this works like a charm :-)
I sent an e-mail to Mike Dimicco to see if he was OK with me sharing the plugin.

A Link plugin would have been the easiest though. Couldn’t Vienna automatically parse encoded text for the URL? Wouldn’t it work in all cases then?

Corentin

[cortig]

I sent an e-mail to Mike DiMicco to see if he was OK with me sharing the plugin.

He is and here is the plugin if anyone needs it (I didn't find another way to share it on the main site):
http://www.cortig.net/files/ShareWithPapers.zip
The server seems to be having some difficulties today, so please be patient if the file appears to be unreachable at first. My hosting provider is looking into it).

Corentin

[Michael Ströck]

A Link plugin would have been the easiest though. Couldn’t Vienna automatically parse encoded text for the URL? Wouldn’t it work in all cases then?

Hi Corentin,

1) Wow, that was quick - thanks for your contribution! Is it OK if we put this up on our website for distribution?

2) I'm not sure what exactly you mean. Could you give me a URL that fails with Papers? (To be completely honest, I we weren't really thinking about local apps when we wrote the link-plugin stuff. It was meant more for web service, all of which to their own parsing on top of what you hand to them - but it would be nice to get this working as well!)

Michael

[cortig]

Hi Michael,

1) Sure you can post it there! It was quick to develop the plugin because you guys made it easy :-)
I also contacted Mike (since I am using his script as the core of the plugin) and he didn’t have any problem with his script being made available.
Let me know if my site is still unreachable. I guess I can send it by e-mail somehow…

2) I use Vienna to keep track of scientific publications. Once I find an article I like, I send it to Papers, for consultation, archiving, etc.
A lot of URLs are built on searches and contain special characters like “?”.
This URL fails for instance:
http://www.ncbi.nlm.nih.gov/pubmed/19827607?dopt=Abstract
I systematically end-up on the main PubMed page.

Corentin

[Michael Ströck]

No problem, the download worked alright! We'd just like to have a repository of these things on the website where people can download it.

I'll look at the problem you reported. Remember, this is a beta :-)

[cortig]

I'll look at the problem you reported. Remember, this is a beta :-)

I know, that’s why I reported the problems on the forum :-)
If no one reports them, you can never fix them ;-)

Corentin

[Michael Ströck]

I do not have Papers installed. Just to be sure, before I download Papers and start working on this: Where does it break, exactly? You mentioned the "?" character - if it breaks there, that's Paper's bug. That's a perfectly legal character in URIs. Actually, it would be wrong to encode it. The same goes for "&".

[cortig]

I do not have Papers installed. Just to be sure, before I download Papers and start working on this: Where does it break, exactly? You mentioned the "?" character - if it breaks there, that's Paper's bug. That's a perfectly legal character in URIs. Actually, it would be wrong to encode it. The same goes for "&".

Well I don’t know exactly since I can’t see the URL of the page in Papers. All I can tell you is that I’m systematically redirected to the root of the site if I make a Link plugin, but it works just fine through the Script plugin I made.

Corentin

[Michael Ströck]

Well I don’t know exactly since I can’t see the URL of the page in Papers. All I can tell you is that I’m systematically redirected to the root of the site if I make a Link plugin, but it works just fine through the Script plugin I made.

OK, I see. I still think this is Paper's bug - "http://www.ncbi.nlm.nih.gov/pubmed/19827607?dopt=Abstract" is a legal URL - nothing to encode there. I just put our URL encoding through its paces. For example:

www.özéarñ.com/there be spaces.cgi?text=Hello Günter¡ (an IDN with other crazy stuff)

correctly (UTF8-) encodes to

http://www.xn--zar-bma6bza.com/there%20be%20spaces.cgi?text=Hello%20G%C3%BCnter%C2%A1

Sorry, nothing I can do here.

[cortig]

OK, I see. I still think this is Paper's bug - "http://www.ncbi.nlm.nih.gov/pubmed/19827607?dopt=Abstract" is a legal URL - nothing to encode there. I just put our URL encoding through its paces. For example:

www.özéarñ.com/there be spaces.cgi?text=Hello Günter¡ (an IDN with other crazy stuff)

correctly encodes to

http://www.xn--zar-bma6bza.com/there%20be%20spaces.cgi?text=Hello%20G%C3%BCnter%C2%A1

Sorry, nothing I can do here.

I wasn’t talking about UTF-8 or other encoding. I was referring to translating some special characters such as , / ? : @ & = + $ # into HTML entities.
The Papers folks have a Web Browser JavaScript bookmarklet for Papers and I can see that it uses the “encodeURIComponent” function.
The script I am using also performs a similar operation to translate these characters.
I agree that there is nothing wrong in what Vienna currently does. It’d be great though if Vienna could automatically translate the special characters into HTML entities, but if it can’t, well I guess there is always the AppleScript ;-)

Corentin

[Michael Ströck]

I wasn’t talking about UTF-8 or other encoding. I was referring to translating some special characters such as , / ? : @ & = + $ #

Yeah, me too :-) Again, the URL http://www.ncbi.nlm.nih.gov/pubmed/19827607?dopt=Abstract already is encoded. Try putting it through JavaScript's encodeURI yourself: http://www.w3schools.com/jsref/tryit.as ... _encodeuri

If Papers can't use that URL as is, there's something wrong with its URL handling, and nothing further Vienna could possibly do.

Could you post a zip of the link plugin version that you wrote?

[cortig]

I think I posted a bad link…
PubMed is a tricky service. The full links I get through RSS actually look something like this:
http://www.ncbi.nlm.nih.gov/entrez/quer ... t=Abstract
The link eventually redirects to this:
http://www.ncbi.nlm.nih.gov/pubmed/2007 ... t=Abstract

The redirected link is fine and works through my "link" version of the plugin.
The original one fails. I suspect the & are the problem there. If I truncate the link at the &, I get to the default page I end up with in Papers.
Though… I realize it also is a valid URL :-[, but the & is often problematic in various environments.

I posted the link version there (again, not sure whether the file made it to the server already since 1and1 is have serious issue today).
http://www.cortig.net/files/ShareWithPapers_Link.zip


Corentin

[cortig]

Well in any case, I can also simply paste the code for the plist here :->

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
   <key>Name</key>
   <string>ShareWithPapers</string>
   <key>Default</key>
   <true/>
   <key>Type</key>
   <string>Link</string>
   <key>URL</key>
   <string>papers://url/$ArticleLink$</string>
   <key>FriendlyName</key>
   <string>Papers</string>
   <key>Tooltip</key>
   <string>Share the current article with Papers</string>
   <key>ButtonImage</key>
   <string>PapersButton</string>
   <key>MenuPath</key>
   <string>Share With Papers</string>
</dict>
</plist>


Corentin

[cortig]

I made some tests using the AppleScript itself to see what it did.
Through the script:
http://www.ncbi.nlm.nih.gov/entrez/quer ... t=Abstract
becomes:
papers://url/http:%2F%2Fwww.ncbi.nlm.ni ... 3DAbstract

Everything gets translated — in hex — including the / & and ?
I realize this falls outside of the range of applications Vienna should have to perform and it looks like we'll have to stick with the AppleScript.
It's not that bad anyway since it seems to be fairly snappy, at least on my Mac.

Corentin

[Michael Ströck]

Actually, I now understand what's happening - I missed the part about the custom url scheme papers://. The URL in this case is papers://url/ and the rest is just data that is handed to that, so encoding it in this way is right. Do many other applications to that? In that case, we might have to implement it and parse out the URL scheme if it is not http:// or https://.

[cortig]

Do many other applications to that? In that case, we might have to implement it and parse out the URL scheme if it is not http:// or https://.

I suspect all applications that support custom links need something like that.

Corentin

[randian]

I'm confused because of the numbering change. The 2.5 version number used to be "2.5.0.2500.1507". The new beta dropped the "15xx". What version is the beta, 1509? 1510?

[AlanR]

I'm confused because of the numbering change. The 2.5 version number used to be "2.5.0.2500.1507". The new beta dropped the "15xx". What version is the beta, 1509? 1510?

The versions with the four digit extension (.1507 above) are not versions released by the Vienna team.

For a couple of years now I've been building Vienna from source and posting the apps to vienna.mpaka.com. I originally starting posting the builds I created from Subversion source when I saw a poster here asking when a fix would be available and I knew that I'd been running my own build with that fix for over a month.

Generally the builds at vienna.mpaka.com will be more “up to date” than anything you'll find officially available. The Vienna team is busy enough working on the app that they don't need to waste their time publishing new builds every time a change or two is made. In any case it would make their job more confusing, developers have to stick to certain logical break points when they decide to make a release. I'm not as constrained as they are and basically all I do is let Xcode create the build and I publish it, a simple job for me.

The new beta (2.5.0.2500) is at the moment one revision behind what you can find at vienna.mpaka.com (SVN version 1510). I started adding the SVN revision number extension so that people who do use use my builds can easily tell that it's one of mine and what revision it is. I don't co-ordinate closely with the dev team but I'm pretty sure they'll never use the X.X.X.XXXX.YYYY numbering I'm using, that's why I decided to take that route.

I hope all of this is clear, if you've any questions just ask.

[czanderna]

I quickly made a plugin for myself for posting the article to delicious. And I thought it would be nice if other users would get the chance to use it.
I hope it's okay to post this plugin here. If it's not so, please tell me where to post it.

Here is my plugin: ShareWithDelicious