| Age | Commit message (Collapse) | Author |
|---|
|
Used e.g. for searches `inurl:xxx`
Co-authored-by: Marien Fressinaud <dev@marienfressinaud.fr>
|
|
Be homogeneous in field used, to prevent weird behavior in the future.
|
|
'timestampUsec' field in greader API. (#2773)
- Add the ability to customize 'dateAdded()', so an extension can change it if needed.
|
|
This fix makes the authors part of the "title item", so it's not
considered as another block. This is not perfect since authors will
disappear on small screens, but we can discuss of putting titles on
multi-lines (see https://github.com/FreshRSS/FreshRSS/issues/2344)
|
|
Co-authored-by: Frans de Jonge <fransdejonge@gmail.com>
Co-authored-by: Marien Fressinaud <dev@marienfressinaud.fr>
|
|
The `FreshRSS_Auth::hasAccess` method is called during auth
initialization (`app/FreshRSS.php:78`), only for `user#create` action.
However, at this step, the `user` configuration namespace hasn't be
initialized yet, and so users weren't able to register because of the
exception... quite critical!
|
|
* Pass the entries to 'entry_before_display' extension hook. - #2762
* Fix code-style.
* - Fix 'Whitespace found at end of line'.
Co-authored-by: Frans de Jonge <fransdejonge@gmail.com>
|
|
While testing <https://github.com/FreshRSS/FreshRSS/pull/2766> I noticed that clicking the label activated the wrong dropdown.
|
|
Now FRSS supports more than one admin. Admins have the same rights as
the default user. Admins can promote or demote other users. The default
user is considered as an admin even if it does not have the admin flag
enabled.
See #2096
|
|
The main reason for this checklist is to not forget about documentation,
and to encourage us to write more tests. I hope it'll help us to improve
both!
|
|
Before, the default loader was used. It rendered poorly since it has a white
border.
Now, the theme loader is used.
|
|
* Fix wrong getHeader refactoring
Fix regression introduced by
https://github.com/FreshRSS/FreshRSS/pull/2373
The refactoring required a call to init() even for static functions,
which is most of the time not done.
Removed premature abstraction of `$_SERVER`, which was the root cause of
the bug.
https://github.com/FreshRSS/FreshRSS/issues/2748#issuecomment-569898931
* Refactoring: Move serverIsPublic to Minz_Request
* Add mitigations for wrong configurations
Due to the regression, we have some existing configurations with a bad
base_url
* Forgot one instance
|
|
* Fix logging environment constants
* COPY_LOG_TO_SYSLOG was only working when Syslog was used
* FRESHRSS_ENV was not properly used to set logging level
* Simplify code
Always call `openlog()` even when it is not used
* Always specify syslog properties
https://github.com/FreshRSS/FreshRSS/pull/2745#discussion_r362207445
* $username is also needed further down
* No new line in syslog
|
|
* Fix PostgreSQL install
Fix https://github.com/FreshRSS/FreshRSS/issues/2732
When username is different from database name
* Add more comments
https://github.com/FreshRSS/FreshRSS/pull/2735/files/2b7807d1b3fdd0dc34538faad465934fb737fa3a#r361846714
|
|
I published an article via WebSub, but I never received it. So I checked
my logs to find why and I found the following stacktrace:
```
29/Dec/2019:15:58:32 +0000 "POST /api/pshb.php" 500
NOTICE: PHP message: PHP Fatal error: Uncaught Error: Call to undefined function _t() in /path/app/Models/Category.php:83
Stack trace:
/path/app/Models/CategoryDAO.php(417): FreshRSS_Category->_id(1)
/path/app/Models/CategoryDAO.php(192): FreshRSS_CategoryDAO::daoToCategory(Array)
/path/app/Models/Feed.php(466): FreshRSS_CategoryDAO->searchById('1')
/path/app/Controllers/feedController.php(416): FreshRSS_Feed->cleanOldEntries()
/path/p/api/pshb.php(141): FreshRSS_feed_Controller::actualizeFeed(0, 'https://flus.io...', false, Object(SimplePie))
{main}
thrown in /path/app/Models/Category.php on line 83
```
The `_t` function should be loaded with the Minz_Translate class, but
the latter isn't initialized on WebSub endpoint.
In my opinion, we should not have to care about this kind of detail of
implementation and it reveals a deeper architectural misconception, but
for now the fix should be enough. It’s quite difficult to reproduce
locally though.
|
|
* tec: Add a test target to Makefile
Minor edit put php7-phar on third line
I try to keep on the third line the Alpine-specific PHP extensions (i.e.
that are not by default in Ubuntu), and in alphabetic order
Co-Authored-By: Alexandre Alapetite <alexandre@alapetite.fr>
* Allow to run make test without Docker
|
|
|
|
|
|
For a moment, PHP wasn't able to connect to my database. I tried to
understand what happened, unfortunately, the logs were not very helpful
(it basically showed me information that I had in my configuration
file).
I changed the dsn string by the message from the raised exception, I
think it will be more useful.
Other call of Minz_PDOConnectionException are passing error messages
instead of string connection, so I took the opportunity to rename the
constructor argument.
|
|
https://alpinelinux.org/posts/Alpine-3.11.0-released.html
Apache/2.4.41 (unchanged), PHP 7.3.13 (updated from 7.3.11 when Alpine
3.10 was released, but unchanged since the latest build of 3.10)
|
|
* Rework the travis file to enable PHPUnit
I realized that unit tests weren't executed on Travis. While working on
this file to enable these tests, I started to think we could simplify
it.
I separated jobs so:
- PHP linter and tests are only performed on PHP 7.3
- Translations are tested separatly so they can fail (it was already the
case but it was hard to understand the way it was done)
- PHP 5.6 only checks syntax issues
- the last job checks css, js, etc. (it didn't change)
PHPUnit is not executed on 5.6 because only the version 5 is available
while the latest version is the 8 (https://phpunit.de/supported-versions.html).
I think it's easier to read (each job is more explicit) but I'm not a
Travis expert so maybe there's some room for improvements.
* Fix failing tests
The category `_name` regression was introduced in commit b323ed084.
I wasn't able to understand when and why Search tests was wrong.
The rest is about upgrade of PHPUnit.
|
|
The constant was badly named. The bug was introduced in 3c49986ec,
probably because of a copy/paste.
There is no other occurrences of COPY_LOG_TO_STDERR in the code (checked
with `git grep`).
|
|
For an extension, I needed to call a script from an external domain.
Unfortunately, the CSP headers didn't allow this domain and I had to
patch manually the FreshRSS FrontController for my extension. It's
obviously not a long-term solution since it has nothing to do in the
core of FRSS, and I don't want to apply this patch manually at each
update.
With this patch, I allow changing the CSP header from inside the
controller actions. It allows extensions to modify headers. It's also an
opportunity to remove a bit of code from the FrontController. I wasn't
happy with the previous implementation anyhow.
Reference: https://github.com/flusio/xExtension-Flus/commit/ed12d56#diff-ff12e33ed31b23bda327499fa6e84eccR143
|
|
This is the end of the `dev` branch. Good bye old friend!
|
|
* Change default TAG in Makefile
We are going to drop the `dev` branch. The Docker tag `dev-*` are based
on this branch and so there will be no longer be generated. We must use
images based on the `master` branch (i.e. `latest`, `alpine` and `arm`).
* Remove references to dev branch in documentation
* Synchronize French documentation
|
|
I'm currently playing with a small extension to improve the email sharing
system. It allows to create a message and send an email directly through
FreshRSS. For that, I need to retrieve the shared article. I could use
the link or the title but it would be less robust and less efficient.
Reference: https://github.com/FreshRSS/FreshRSS/issues/300
|
|
It adds two new hooks to have more options for influencing with extensions.
|
|
|
|
When managing users, I've added the language they selected to be able
to communicate with them with the selected language (if I know it).
It could be useful when managing a community or a paid service.
|
|
|
|
When using feed navigation, the previous behavior was to cycle through all
available feeds regardless of their content. To match the behaviour of the
first feed and last feed navigation shortcuts, the navigation now skips
empty feeds. Now it's consistent through out the application.
When using feed navigation with only empty feeds, the new behavior is to
cycle through all available feeds.
See #2385
|
|
* Update i18n cli tools
Few things were bugging me when using the cli tool for i18n. So I've modified
the tools to be easier to use.
First, the tool automatically adds missing keys to all languages. This way, we
always have all keys in all languages.
Second, the tool detects all untranslated keys and adds automatically the todo
comment after the value.
Third, when adding a new key, the key is pushed to all languages at once. There
is no need to duplicate it manually. Thus making the duplication process obsolete.
Fourth, translation and ignore keys are manipulated at the same time. Thus we
don't have obsolete ignored strings anymore.
* Add i18n rules
I find that having the common rules in the Makefile is easier to use,
as long as you know they are here. As it is self documented, people
will see the new rules when using make.
* Use long parameters in Makefile
I find that using long parameters in scripts makes it easier to understand
what's going on. So I've switched all short parameters to long one.
* Format all i18n files
I've used the updated version of the cli tools to have some output
that can be consistently formated. This commit is a huge formating
commit. Nothing was added but some comments were removed in the
process.
|
|
I split the code into more manageable chunks to make it more readable.
It's now easier to get a grasp on it. In the process, I tried to use
the yoda style notation to avoid unintentional variable assignement.
I also tried to exit early to avoid many "if" levels.
I have also added a better way to handle headers and a way to extract
languages from the request. It will be easier to choose which language
to use for i18n.
|
|
* Extract hashPassword method from userController
* Extract and refactor fever key-related methods
* Move update of API password to dedicated action
* Simplify the controller by refactoring feverUtil
* Add locales
|
|
Many broken links.
Related to https://github.com/FreshRSS/FreshRSS/pull/2614
|
|
* [i18n] Add docs po4a script
* Add proof of concept
* Add a few more translations
* Hush ShellCheck and shfmt
* Make that list po4a-friendly
* drat, this document could've probably been auto-generated
* Definitive proof that it's translated from French ;-)
* Add some brand spanking new French translation
* More translation
* Mostly finish that config page
* Fix up FAQ
* More contributing
* Dev first steps
* Let's ignore that admin stuff at the very least for now
* Translate release new version, make French the source first and copy all translations
Then replace French with English in the source.
Much quicker than any alternative route.
* And add the English translation
* Minor stylistic leftover from French
* Most of first steps
* Forgot the extensions
* Use po4a 0.56 to get rid of way too many newlines
* Fix up those newlines
* No point linking to Firefox integration anymore from the new user guide
* Start on main view
* A bunch of main view stuff
* More main view
* And some subscriptions before going to bed
* First steps for devs
* More dev first steps
* Incomplete French → English dev/GH translation
Because I need to ask about that mailing list thing
* Fix typo in docs/en/developers/02_Github.md
* Translate & complete devs/github to English
* Fix up most of extensions
* Is that supposed to be a non-breaking space? Let's see
* Match up some users/mobile access
* More users/mobile access
* Add fresh French translation to Fever API
* Fix typo
* Match frontend todo thingies
* Fix a typo
* Some extensions strings
* Remove Fx subscription service from the docs
Cf. https://github.com/FreshRSS/FreshRSS/pull/2606
* Add translation for https://github.com/FreshRSS/FreshRSS/pull/2643
* fix typo as per https://github.com/FreshRSS/FreshRSS/pull/2643#discussion_r345433009
* Add some more French translations
* Update French translation as per @aledeg comment https://github.com/FreshRSS/FreshRSS/pull/2590#discussion_r345465909
* Translate some of the meaningless stuff
* Translate the rest of contributing.md to French
* Fix conflicts
* Translate Docker first steps to French
* Update with change from #2665
* Add @aledeg corrections
* Overlooked a couple @aledeg corrections thanks to GitHub autohide
* Latest @aledeg suggestions
|
|
Before, the use of the user management page was a little bit tedious
when there was many users. One must select a user to view some
metrics, to update it, or to delete it.
Now, the view is clearer because it shows all users at once with
their metrics. I introduced a detail page that repeats the metrics
but also allow to purge the user's feeds, to update or delete the
user.
This is the first step to make that page more useful and user-friendly.
I have in mind to add a pager for when there is a lot of users, a metric
to know when was the last time the user was using the application, and
a flag to know if the user has admin rights.
See #2096 and #2504 for ideas and inspiration
|
|
While I was looking at the number of articles of my users, I discovered
some of them had none, while having a bunch of feeds though. I took a
look at the logs generated by `app/actualize_script.php` and discovered
that the script stopped strangely (in this example, "OK" for denise is
expected, and more users too):
```
FreshRSS[1681]: FreshRSS Start feeds actualization...
Starting feed actualization at 2019-11-29T16:37:19+00:00
Actualize alice...
Actualize denise...
Results:
alice OK
denise
```
After digging a bit, I quickly realized the script stopped always on
users who didn't validate their emails. And indeed, we trigger a
`Minz_Request::forward(..., true)` for these users, in the `FreshRSS`
class. This function calls the `exit` function, which stops the script.
This patch only allows the feed#actualize action to be executed for
unverified users in order to avoid an early-`exit`. This is a quick-win
solution, but I don't think it's a good one on the long term. I'll
propose an alternative in another patch, later.
|
|
* Remove magicQuotesOff (#2697)
Magic quotes have been deprecated since PHP 5.4 and as of PHP 7.4 `get_magic_quotes_gpc()` displays a warning.
* Remove unused Minz_Helper::stripslashes_r
`Minz_Helper::stripslashes_r` is no longer used.
It's last user was `Minz_Request::magicQuotesOff`
|
|
See https://github.com/FreshRSS/FreshRSS/pull/2662#issuecomment-553575705
|
|
started translating open todos
|
|
* FIX email token check if not in user-config
* fix missing semicolon
|
|
|
|
FreshRSS 1.15.3
|
|
Co-Authored-By: Frans de Jonge <fransdejonge@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
Alexandre Alapetite
Julien-Pierre Avérous
Marien Fressinaud
Offerel
Frans de Jonge
Alexis Degrugillier
romibi
石渠清心
Clemens Neubauer
Alexis Degrugillier
Craig Andrews
Sebastian K
Purexo
