API /reader/api/0/stream/items/contents #1774

Alkarex · 2018-01-25T22:19:48Z

For FeedMe https://play.google.com/store/apps/details?id=com.seazon.feedme

For FeedMe

Frenzie · 2018-01-30T13:19:50Z

So they use a slightly incompatible Google Reader API?

Alkarex · 2018-01-30T13:20:34Z

@Frenzie More some features that I did not implement because I had not seen anybody using them :-)

For FeedMe. This token is not used by e.g. The Old Reader API. There is the Authorization header anyway. TODO: Check security consequences

FeedMe uses GET for some parameters typically given by POST

Frenzie

Is there such a thing as API docs?

Frenzie · 2018-02-07T10:09:02Z

p/api/greader.php

@@ -216,9 +216,13 @@ function token($conf) {
 function checkToken($conf, $token) {
 //http://code.google.com/p/google-reader-api/wiki/ActionToken
 	$user = Minz_Session::param('currentUser', '_');
+	if ($user !== '_' && $token == '') {


Strict for the first comparison, loose for the second (match empty string and 0 in particular)

Frenzie · 2018-02-07T10:15:16Z

p/api/greader.php

@@ -476,73 +528,37 @@ function streamContents($path, $include_target, $start_time, $count, $order, $ex
 			break;
 	}

-	if (!empty($continuation)) {
+	if ($continuation != '') {


So we're fine if it's FALSE or 0? Also, is $continuation trimmed anywhere or could you end up with something empty-adjacent like a space?

Also, like above, != or !==?

Yes, loose comparison there. The test should reject empty string, false, 0, null, etc. and only keep substantial strings.
But good point with sanitising those parameters. Will do so.

Ah, so != something falsy (loose) basically means exactly the same thing as empty()? I know that PHP plays a bit fast and loose with things like 1 and 0 vs true and false, but I didn't realize empty strings were included in being falsey. I thought any string was truthy. Apparently the string '0' is also falsey.

Confusingly, on that last note, 0.0 (float) is falsey and '0.0' (string) is truthy.

Yes indeed. empty() is mostly relevant when working with potentially non-existing entries such as in a array, and it is similar to doing !isset($t[1]) || $t[1] == ''. In the case where we have a variable and not an index, there is little point of using empty() and it used to be slower (I am not sure whether there is any difference nowadays).

I find empty() slightly clearer in the sense that I don't necessarily think of an empty string as falsey while I do think of it as empty, but that's just my problem I suppose. :-P

Python is much more consistent in counting empty arrays, empty strings, etc. as falsey than the likes of JavaScript or PHP. But in Ruby, for example, empty strings are truthy, and in a sense they are in Lua too. In POSIX and related shells you also have to be well aware to use checks like -z (unset or empty) and -n (set and non-empty).

Incidentally, can we get this kind of fun out of PHP?

bla = '0'; if (bla && bla == false) console.log('/sigh') /sigh

It's truthy (non-empty string) but at the same time '0' == false. Perhaps even more confusingly, you can get the same behavior out of new Boolean(false) (the existence of any object is inherently truthy in JS, but its value == false).

https://www.destroyallsoftware.com/talks/wat :-)

Fun, but no PHP. Well, at least I got the last one right after the hint. I didn't really dare think not a number at first because that just seemed too logical. xD

All right, I have one for you. Guess the output :-P

<?php echo ctype_digit(1234) ? 'true' : 'false', "\n"; echo ctype_digit(123) ? 'true' : 'false', "\n"; echo ctype_digit('123') ? 'true' : 'false', "\n";

My intuitive guess would be false false true but I'm guessing there's a catch. Now to check… ;-)

Edit: hm, true false true. Well then…

If an integer between -128 and 255 inclusive is provided, it is interpreted as the ASCII value of a single character (negative values have 256 added in order to allow characters in the Extended ASCII range). Any other integer is interpreted as a string containing the decimal digits of the integer.

Close enough: true false true. :-)

If an integer between -128 and 255 inclusive is provided, it is interpreted as the ASCII value of a single character (negative values have 256 added in order to allow characters in the Extended ASCII range). Any other integer is interpreted as a string containing the decimal digits of the integer.

Alkarex · 2018-02-07T21:01:32Z

A new version of FeedMe will be released when we release FreshRSS 1.10.0 with this patch.

Alkarex · 2018-02-08T13:06:03Z

Preview:

Alkarex · 2018-02-08T19:14:56Z

@MaluNoPeleke and others: The /dev branch of FreshRSS (soon 1.10.0) and FeedMe 3.5.3+ (currently available as beta in the Play Store) should work fine together. Tests welcome.

Alkarex · 2018-02-08T19:28:48Z

#328

bramboomen · 2018-02-18T17:06:04Z

Hi, I just set-up a fresh install of FreshRSS dev and FeedMe 3.5.4 and encountered an issue.
The list of articles does not synchronize. As far as I can tell the connection is working because starred items do synchronize and FeedMe has the complete list of my subscriptions but no unstarred, read or unread articles.
I did have some issues with setting up the API, in particular with the PATH_INFO variable in nginx, but currently both api checks pass and the EasyRSS application is working without problem.
I would try some manual API calls to further debug the problem, but I have no idea where to begin in that endeavour...

Alkarex · 2018-02-18T20:23:06Z

@blankoworld Thanks for the test.
Please make a detailed log by enabling the two following lines

FreshRSS/p/api/greader.php

Lines 740 to 743 in c79b8fb

    
           //Minz_Log::debug('----------------------------------------------------------------', API_LOG); 
        
           //Minz_Log::debug(debugInfo(), API_LOG);

and either set 'environment' => 'development', in ./data/config.php or change the lines to Minz_Log::warning().
It will be in ./data/users/_/log_api.txt

bramboomen · 2018-02-18T20:59:57Z

@Alkarex
Happy to, let me know if I can do anything else!
log_api.txt

Alkarex · 2018-02-18T21:29:40Z

@bramboomen Could you please ensure that you have a number of unread articles before trying a first synchronisation?
If that still does not work, would it be possible for you to make me a test account on your FreshRSS instance?

bramboomen · 2018-02-18T21:49:20Z

@Alkarex
There were both unread and read articles but they only show up if I star them.
I will send you an email with the information for the test-account.

Alkarex · 2018-02-18T22:00:18Z

Thanks for the test account. It seems to work fine for me :-)
Can you give a try to the test account you have just made for me?

Alkarex · 2018-02-18T22:07:19Z

Ah, in FeedMe, under Settings / Synchronisation Mode, pick "All". (I am not sure of the exact wording of the English version)

bramboomen · 2018-02-18T22:12:43Z

@Alkarex
Synchronizing 'all' instead of 'per feed' did the trick, thank you!

Alwaysin · 2018-02-28T22:49:11Z

Had the same problem!
I guess a lot of people might conclude "this is not working" and go somewhere else.

Anyway, this reader is a nice addition, I'll use it a couple days to see if it fits my needs :)

Alkarex · 2018-03-01T19:41:41Z

The next version of FeedMe will have the proper synchronisation setting by default.

fcoiffie · 2018-03-03T13:57:39Z

Regarding the FeedMe configuration, it's not very easy to guess that the Host is: http://adresse_serveur/p/api/greader.php

Alkarex · 2018-03-03T14:05:45Z

@fcoiffie Indeed, therefore:

It is shown in FreshRSS, in the API test page (same page than where one set the API password)
It is in the documentation https://freshrss.github.io/FreshRSS/en/users/06_Mobile_access.html#compatible-clients
I have already suggested to the developer of FeedMe to auto-add api/greader.php if not already included in the address provided by the user.
I will try to see if we can add a redirect without affecting performances.

Please note that is is better to expose only the ./p/ directory of FreshRSS (and not its parent), to have a public URL like https://freshrss.example/ instead of https://freshrss.example/p/.

* API /reader/api/0/stream/items/contents For FeedMe * Fix continuation * Continuation in stream/items/ids * Fix multiple continuations * Allow empty POST tokens For FeedMe. This token is not used by e.g. The Old Reader API. There is the Authorization header anyway. TODO: Check security consequences * API compatibility FeedMe: add/remove feed FeedMe uses GET for some parameters typically given by POST * A bit of sanitization * Links to FeedMe * API favicons more robust when base_url is not set * Changelog FeedMe

API /reader/api/0/stream/items/contents

304f6a0

For FeedMe

Alkarex added this to the 1.10.0 milestone Jan 25, 2018

Alkarex added the Work in progress 🚧 Wait before merging label Jan 25, 2018

aledeg added the API 🤝 API for other clients label Jan 28, 2018

Fix continuation

9ad1e3c

Alkarex added 4 commits January 30, 2018 15:04

Continuation in stream/items/ids

8c17351

Fix multiple continuations

5a8415c

Allow empty POST tokens

575322f

For FeedMe. This token is not used by e.g. The Old Reader API. There is the Authorization header anyway. TODO: Check security consequences

API compatibility FeedMe: add/remove feed

a69eae5

FeedMe uses GET for some parameters typically given by POST

Alkarex modified the milestones: 1.11.0, 1.10.0 Feb 5, 2018

Alkarex removed the Work in progress 🚧 Wait before merging label Feb 5, 2018

Frenzie approved these changes Feb 7, 2018

View reviewed changes

A bit of sanitization

24648c5

Links to FeedMe

b93032f

API favicons more robust when base_url is not set

c3adf2c

Alkarex mentioned this pull request Feb 8, 2018

Make API fever compatible #1406

Closed

Changelog FeedMe

fcdd156

Alkarex merged commit 79f8b44 into FreshRSS:dev Feb 8, 2018

Alkarex deleted the API_FeedMe branch February 8, 2018 19:11

This was referenced Feb 8, 2018

Liste des lecteurs RSS mobiles #328

Closed

FreshRSS 1.10.0 #1786

Merged

Frenzie mentioned this pull request Feb 27, 2018

1.10 displays _ for author in EasyRSS #1799

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

API /reader/api/0/stream/items/contents #1774

API /reader/api/0/stream/items/contents #1774

Alkarex commented Jan 25, 2018 •

edited

Frenzie commented Jan 30, 2018

Alkarex commented Jan 30, 2018

Frenzie left a comment

Frenzie Feb 7, 2018

Alkarex Feb 7, 2018

Frenzie Feb 7, 2018

Alkarex Feb 7, 2018

Frenzie Feb 7, 2018 •

edited

Alkarex Feb 7, 2018

Frenzie Feb 7, 2018

Alkarex Feb 7, 2018

Frenzie Feb 7, 2018

Alkarex Feb 7, 2018

Frenzie Feb 7, 2018 •

edited

Alkarex Feb 7, 2018

Alkarex commented Feb 7, 2018

Alkarex commented Feb 8, 2018

Alkarex commented Feb 8, 2018

Alkarex commented Feb 8, 2018

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018 •

edited

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018

Alkarex commented Feb 18, 2018

bramboomen commented Feb 18, 2018

Alwaysin commented Feb 28, 2018

Alkarex commented Mar 1, 2018

fcoiffie commented Mar 3, 2018

Alkarex commented Mar 3, 2018

API /reader/api/0/stream/items/contents #1774

API /reader/api/0/stream/items/contents #1774

Conversation

Alkarex commented Jan 25, 2018 • edited

Frenzie commented Jan 30, 2018

Alkarex commented Jan 30, 2018

Frenzie left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Frenzie Feb 7, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Frenzie Feb 7, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Alkarex commented Feb 7, 2018

Alkarex commented Feb 8, 2018

Alkarex commented Feb 8, 2018

Alkarex commented Feb 8, 2018

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018 • edited

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018

bramboomen commented Feb 18, 2018

Alkarex commented Feb 18, 2018

Alkarex commented Feb 18, 2018

bramboomen commented Feb 18, 2018

Alwaysin commented Feb 28, 2018

Alkarex commented Mar 1, 2018

fcoiffie commented Mar 3, 2018

Alkarex commented Mar 3, 2018

Alkarex commented Jan 25, 2018 •

edited

Frenzie Feb 7, 2018 •

edited

Frenzie Feb 7, 2018 •

edited

Alkarex commented Feb 18, 2018 •

edited