In previous versions of the authenticate package, there was something for Facebook authentication. Why does it not exist in the newer versions? Does it have to do with the fact that the Facebook API changed?
The code was superseded by the more comprehensive fb package. It was a mistake on my part to not mention this in the package description itself (you're not the first to get caught by this change). I've updated the cabal file and uploaded a new version to Hackage. Hopefully that will avoid further confusion.
It was removed in this commit. As for why, you'd have to ask Michael Snoyman himself; there's no issue corresponding to commit that I could find. I suspect the code had simply become outdated because of Facebook updates.
However, since Facebook support OAuth (and indeed this seems to be what the old code was based on), you may be able to use authenticate-oath, which contains the OAuth code that was split out from authenticate proper. (If this does work, then that may be why the Facebook-specific support was removed: it would be redundant to the generic OAuth code.)
Related
My code show this error. I didn't use "flutter_secure_storage" dependency but it shows this error. But the app is working well, What is the reason for this and how to fix this also if didn't fix this, the effect of this error?
flutter_secure_storage is a plugin which uses platform specific APIs to interact with the device.
This warning suggests that it is using a deprecated Java API on the Android device. You can't really fix it unless you know Java and are willing to contribute to the plugin. At the same time, you can safely-ish ignore it for now: even deprecated APIs usually last a long time, and it won't affect any functioning until the underlying API is actually removed.
The best thing to do is regularly check for updates to the package, which might fix this warning.
As the error suggests, the package flutter_secure_storage use the file specified in the error, which in turn use or implements a deprecated API.
You might not use that plugin, but you are using some other plugin that use it.
You are not affected by it now because it is just a warning/information.
You could get affected by it when the depreciation goes to be removed instead. But by that time the package will have been updated. So all you have to do is update the package at a later time when said deprecation has been solved.
Because of a context I need to use the older versions of the Facebook API and use the v2.3. When i make a new facebook AppId the version is read only to v2.6 and i get to see this hint
I really hope i can still make apps with older versions, but not sure how yet. Does anyone had this situation before and can lead me to something that helps? Thank you.
If you create a new App, you can only use the current API version. There is no way to downgrade below that version.
It would be pointless anyway, don´t build something new on an API that is deprecated, because it would stop working sooner or later.
Edit: Well, that answer in the other thread is from me. Of course it´s still valid, and it will always be ;)
I downloaded some module from CPAN and added to it extra functionality. Could I post the newest module on the CPAN? If yes, how should I do this? Is it possible at all? Should I contact the person who wrote the original module? Could someone help about the procedure to update this specific module?
You won't be able to upload and index a module as the same name unless you are a co-maintainer. Uploading it as a different name isn't very productive for the community as there are now two (or more) slightly different versions.
Contact the author
Submit your patches to the RT queue for the module (or the issue tracker it specifies in its docs)
If it's on Github, fork the project, make the change, and submit a pull request
If the author is unresponsive (which means no response, not a negative response), we have ways to pass on maintainership of abandoned modules. We take this process very slowly because we want to give the original author or current maintainer every chance to respond. Some people might be swamped at work, on holiday, and so on. However, most situations turn out very well.
You can upload anything that you like to PAUSE, but if you are not some sort of maintainer, PAUSE won't index it. It will still show up in your CPAN account and people can still download it through the CPAN website, but the CPAN clients won't see it (since they work according to the index). Some of the search sites will show it as an "Unauthorized release".
It would be unhelpful to blindly upload a module that you've added functionality to under the same name as the module that it's based on (it happens sometimes and often creates a mess). But you do have options that would be helpful:
First, you could contact the module's author or maintainer, emailing a diff or patch showing what you would like to see added. Be sure to supply relevant documentation as well, and explain your rationale.
It could be that the module author will accept the patch and apply it to an update of the module himself. Or it could be that by way of prevention of creeping featurism, the author/maintainer rejects the upgraded functionality, and that leads you to a second option.
The second option is to get your own PAUSE account, and either subclass or otherwise extend the module. But be sure to give it your own new module name, full documentation, and probably an explanation within the documentation that this is an extension of xyz module with the following additional functionality... If you're extending Math::BigInt, you might call it Math::BigInt::Frobcinate (just an example).
When you do attempt to make contact with the module author, please be patient. Sometimes it takes awhile to let things run their proper course.
I recommend that you attempt to contact the author.
Additionally, you can upload your module changes as a patch via rt.cpan.org; there should be a link from the module's website on CPAN. Refer to the CPAN FAQ: How do I report/fix a bug in a module/script? This tracking system is used for requests such as yours, not just for reporting bugs.
I am using confluence 3.0.1 and have the problem that some images are not shown after logout. I remember there being a fix but cannot find anything despite extensive google searches.
Can someone please point me in the right direction?
Thanks.
Christian
I do not have an installation of Confluence 3.0.1 available, we use a more current version 3.4.9, but perhaps the things are similar.
I think the reason for your missing image is simple. Your profile image is stored in the people directory, and you don't have access to the people directory without login. So even if anonymous access is allowed in your personal wiki, the space logo is taken from the people directory, and there the anonymous user has no access.
There is no workaround known to that problem. Atlassian states that it is a feature, not a bug. See the issue https://jira.atlassian.com/browse/CONF-16626 for a similar case. In newer versions of confluence, a question mark is shown to indicate that a profile picture should exist, but is not shown. So it does not look anymore like a technical bug.
I have been looking to switch to OAuth for my Twitter integration code and now that there is a deadline in less than 7 weeks (see countdown link) it is even more important to make the jump to OAuth. I have been doing Basic Authentication which is extremely easy. Unfortunately OAuth does not appear to be something that I would whip together in a couple of hours.
http://www.countdowntooauth.com/
So I am looking to use a library. I have put together the following list.
MPOAuth
MGTwitterEngine
OAuthConsumer
I see that MPOAuth has some great features with a good deal of testing code in place but there is one big problem. It does not work. The sample iPhone project that is supposed to authenticate with Twitter causes an error which others have identified and logged as a bug.
http://code.google.com/p/mpoauthconnection/issues/detail?id=29
The last code change was March 11 and this bug was filed on March 30. It has been over a month and this critical bug has not been fixed yet. So I have moved on to MGTwitterEngine.
I pulled down the source code and loaded it up in Xcode. Immediately I find that there are a few dependencies and the README file does not have a clear list of steps to fetch those dependencies and integrate them with the project so that it builds successfully. I see this as a sign that the project is not mature enough for prime time. I see also that the project references 2 libraries for JSON when one should be enough. One is TouchJSON which has worked well for me so I am again discouraged from relying on this project for my applications.
I did find that MGTwitterEngine makes use of OAuthConsumer which is one of many OAuth projects hosted by an OAuth project on Google Code.
http://code.google.com/p/oauth/
http://code.google.com/p/oauthconsumer/wiki/UsingOAuthConsumer
It looks like OAuthConsumer is a good choice at first glance. It is hosted with other OAuth libraries and has some nice documentation with it. I pulled down the code and it builds without errors but it does have many warnings. And when I run the new Build and Analyze feature in Xcode 3.2 I see 50 analyzer results. Many are marked as potential memory leaks which would likely lead to instability in any app which uses this library.
It seems there is no clear winner and I have to go with something before the big Twitter OAuth deadline.
Any suggestions?
I've used bengottlieb's Twitter-OAuth without issues.
Build+Analyse finds only one issue (in SA_OAuthTwitterController.m) and that's just a variable that has a value stored during initialization that is never read. I've not seen any leaks in my use of it. It was simple to implement and so far (two apps in the store using it, a third in development) no problems at all.
A new single-file solution is TDOAuth: http://github.com/tweetdeck/TDOAuth.
It's well-tested in that it is used in the TweetDeck iOS clients.
I recently had to implement "post to Twitter".
What I found was that none of the various projects would work out-of-the-box. They're all close, but not quite there.
Eventually I settled on OAuthConsumer, Stig Brautaset's SBJSon, with some ideas more than code borrowed from Ben Gottlieb's Twitter-OAuth-iPhone for pin processing.
I probably would have looked at MGTwitterEngine, but I didn't need to hit that much of the API - just status updates.
I seem to recall having to hack OAuthConsumer to set a few parameters correctly - that was the biggest stumbling block.
Try this Framework out for size: https://github.com/materik/meauth-ios, works well for BitBucket's API and am in progress of testing it with more sites. Please give feedback and contribute.