There are lots of cases where we can inject a User object without
additional efforts. Now $wgUser is only used inside AFComputedVariable,
which is a little bit harder to handle because some instances of that
class are serialized in the DB, and thus we cannot easily change the
constructor until T213006 is resolved.
This partly copies what Ia474f02dfeee8c7d067ee7e555c08cbfef08f6a6 tried
to do, but adopting a different approach for various can*() methods:
they're now static methods in the AbuseFilter class, so future callers
don't need to instantiate an AbuseFilterView class. This also allows to
re-use those methods in an API module for editing filters (T213037).
Bug: T213037
Bug: T159299
Change-Id: I22743557e162fd23b3b4e52951a649d8c21109c8
While this is not as important as throwing for too few parameters, IMHO
it's still important to fail in this case. Mostly because if a function
receives too many parameters, chances are that who wrote the filter
didn't do that intendedly, and thus there may be a hidden bug.
Bonus: fix a few docblocks.
Bug: T230803
Change-Id: Iac2931f17b50ace8c8f4c2faa44b3f54ca134c54
In general, it's not safe to change configuration in the middle of a
test, because services could wind up in an inconsistent state. In
particular, I'm trying to have setMwGlobals() reset services, which will
cause stuff to break if it happens in the middle of a test. So just
specify the settings you want up front, like in setUp().
Change-Id: I00e35ecea6a27468674b2a6e7d9d9eb6518e3bd5
This allows a little bit more of abstraction: we can store other data in the
tree, without having to store it in a specific node (e.g. the variables map,
which is still unused). It also adds a few typehints, and specializes
the return value of eval'ing the AST: previously, it was the one of
evalNode, which wasn't guaranteed to be an AFPData. Now we have this
guarantee. Last but not least, we can now measure runtime metrics for
evalTree, which doesn't recurse.
Bonus: fix a check in the old parser, which used the wrong variable when
reporting outofbounds errors.
Change-Id: Iff806793b1d968e9bb6220f1459f3d0ac587c7da
And fix a couple of minor bugs.
Bug: T156096
Depends-On: I3b85087677607573f4fa68681735dc35348dcd87
Change-Id: Ia4c713a1d45827f6a8bc5566a8d8835c49f8108a
After having removed the export link in
I72f46247f4323fb5bfe7fa74f332076dbd346187, we don't have any tool to
show for new filters. So avoid outputting an empty section.
Change-Id: Ia07bccdbadb7b874397135bc3f7468d6e0b9eb13
Ensure that the variable isn't set before marking it as DUNDEFINED:
that's only for when we cannot use a default, but if the variable is set
we already have one. Most notably, this fixes conditionals handling: right
now, if you have a conditional with an assignment in both
branches, the variable will be undefined. That's obviously wrong, so
it's fixed in this patch.
Plus: catch only AFPExceptions in a test to avoid unintentionally
catching the assert exception; simplify some assignments using wfSetVar.
Depends-On: I446a307e5395ea8cc8ec5ca5d5390b074bea2f24
Change-Id: I8e7f7710b8cb37ada8531b631456a3ce7b27ee45
This patch includes various fixes to how func arguments are handled in
CachingParser:
- Add a comment about a future improvement of checkSyntax, which we
could limit to try building the AST.
- Having enough args for each function is now also checked when
building the AST. This allows implementing the previous point without
stopping to report notenoughargs at syntaxcheck-time (otherwise it'd be
a runtime error). And it also ensure that we check for the params count
inside skipped branches, e.g. inside if/else: these were already only
discovered at runtime in CachingParser. The old parser is not affected
by this change, because when checking syntax it will always execute
all branches, and at runtime it will skip braces altogether.
- Fix arg count for CachingParser, which previously added a bogus param
in case of a function called without parameters. This was fixed for
the other parser in I484fe2994292970276150d2e417801453339e540, and I
just ported the updated fix. Also note that the CachingParser was
already failing for e.g. `count()`, but instead of complaining about
missing arguments, it failed hard when trying to pass NULL to
evalNode.
- Fixed some tests not to use setExpectedException, which caused the
previous point to remain unnoticed: calling that method prevents the
loop from continuing, and thus only the AbuseFilterParser part was
being executed. The new implementation checks the exception ID and is
thus more future-proof if the i18n message changes.
- Fixed some function names in error reporting for the old parser.
- The arg count is now checked outside of the function handlers, thus
it's no more necessary to call checkEnoughArguments at the beginning
of each handler. This also produces clearer error messages in case of
aliases (e.g. set/set_var).
- Check the args count even if some of the args are DUNDEFINED. This is
much easier now that the check is outside of the handler. This will
make syntax check fail for e.g. `contains_any(added_lines)`.
Bug: T156095
Change-Id: I446a307e5395ea8cc8ec5ca5d5390b074bea2f24
This was broken in I34c040dbeb3ab01158fb3db22496def6ccaf72d9. I thought
the members of that object were always arrays, but I was wrong.
Plus typehint a few array parameters and make a couple of methods
private since they're only used in this class.
Bug: T230639
Change-Id: I0c51359769c4b3054f95755a96e7e0a2d8e5bf15
Now it's always wider, and so is the "notes" field. Moreover, the
fallback textarea has the exact same size. Plus removed a parameter
which only made it hard to write a CSS rule for the textarea. Since the
textarea is generated by the same code, and we're always using it for
the same thing (filter syntax, regardless of the final goal), make it
always use the same name.
Bug: T230591
Change-Id: Ibb308e80d954c0e81aa09249c38c39572f157948
Problems fixed:
- Don't hardcode duration in the message
- Move duration to a constant
- Fix wrong parameter order for AbuseFilter::blockAutopromote
- Log a warning if we cannot block autopromotion
- Remove the $performer parameter, as it should only and always be the
filter user.
Bug: T230296
Change-Id: Ice9e4b21033c430cf1fd34182c63ca64ad2f5d3e
If $parser->parse returns a falsey value (=null), that's because the
filter doesn't have any statement. But that's not a valid reason not to
cache the filter. Hence, return whatever parse() is returning inside the
callback, so that the result is always cached.
Change-Id: Ib6b0e72d882dc484456a3be6bbc74da36ef48bf7
In order to bundle AbuseFilter, we have to provide an optimal
configuration for sysadmins. This patch aims to provide such
configuration, and of course it's open to changes.
Changed values:
*Grant abusefilter-view-private to admins. This is mostly for
redundancy-safety (they already have -modify)
*Grant modify-restricted to admins (see T200032)
*Disable the 'degroup' action. This is really dangerous, not ideal to be
enabled by default.
*Grant the abusefilter-revert right to sysops. This is similar to
assigning the modify-restricted right in its rationale.
*Add abusefilter-log-private to viewrestrictedlogs grant. This right
should go together with abusefilter-view-private. (T226987)
*Remove abusefilter-revert from the rollback grant, as that right is
for reverting AF actions and has nothing to do with rollbacks.
This mostly copies the config of WMF production: degrouping is already
disabled there, and modify-restricted should
be assigned to sysops on WMF wikis as well. The assignment of the
abusefilter-view-private and abusefiliter-revert rights is handled in
I09cb462312c5bbc71ee3a8f0fbd91357908130d4 (parent patch) to keep the
status quo.
NOTE: this patch should be *deployed* together with the config patch to
avoid discontinuities. Thus, please ONLY provide a virtual +2 using a
+1, and I'll coordinate the +2 with a deployer.
Bug: T191740
Bug: T200032
Bug: T226987
Depends-On: I09cb462312c5bbc71ee3a8f0fbd91357908130d4
Change-Id: I58f215b8aa3444602532098a3c9dfd25409f29eb
