aboutsummaryrefslogtreecommitdiffstats
path: root/lint/libs/lint_api/src/com
Commit message (Collapse)AuthorAgeFilesLines
* Refactor lint folder structure.Xavier Ducrohet2012-12-0630-11556/+0
| | | | | | | | | | Move folders around to match default gradle/maven folder structure. The custom structure confused some IDEs when importing the project as a Gradle project. Change-Id: I5c059cc6fdf0ac2444b03d435f83f853363cfa32
* Findbugs cleanupTor Norbye2012-12-061-2/+2
| | | | | | | Fix a couple of issues found by findbugs, and some test stability fixes. Change-Id: I97390ea606ea25d6a68e5b7f8245e5e689117995
* Add override detectorTor Norbye2012-11-302-1/+31
| | | | | | | | | | This lint detector finds method declarations which are likely to clash and accidentally override framework classes defined on newer platforms Also, fix method position computation in class files. Change-Id: If0d03feb55e44816f75938dae3164c53ed5b3183
* Add tools:targetApi to control local minSdk (like @TargetApi)Tor Norbye2012-11-301-0/+36
| | | | | | | Also add XML editor quickfix to set the attribute on an API violation. Change-Id: Ife95d73659656e98a6fb1a322354f5fcfcef1888
* Make inefficient weight detector identify more candidatesTor Norbye2012-11-292-3/+51
| | | | | | | | | | | | | | | | | The inefficient weight detector looks for horizontal layouts with weights where all the children are layouts. It didn't have a very good heuristic for finding out whether all children are layouts; it just looked for whether the tag name contains "Layout". This changeset generalizes this a bit, adding other layouts, as well as <fragment> tags to the lint check. It also updates the Master/Detail template to set the baseline alignment attribute to false (to avoid triggering the updated lint check), as well as android:textIsSelectable to clear another recent new lint rule. Change-Id: I887962b5e29a9ad0c5b5c01970b0c66f884125d8
* Add lint check for cut & paste errors in findViewById callsTor Norbye2012-11-271-0/+19
| | | | Change-Id: Ie8ff6629a3e787587516f6fa88c70dfdb7a92d0d
* Lint fix for AOSP unbundledTor Norbye2012-11-201-0/+4
| | | | Change-Id: I306b753319d235b015f44c0927382578805898de
* 39153: lint shows project folder instead of project nameTor Norbye2012-11-053-4/+40
| | | | Change-Id: I415c42e19573bf733de672625a1200e62204ec77
* Fix handling of @SuppressLint on fieldsTor Norbye2012-10-182-10/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | When you have a field initialization like this: private int foo = new ForbiddenClass(); the actual code to perform the initialization lives in a method named <init> (or for a static field, <clinit>). If you tried to suppress lint errors here by adding an annotation on the field, it would not be found by lint, since lint looks at method level annotations and it's sitting on the field node instead. To fix this, the suppress check needs to identify the field scenario, and in that case (when checking whether an error is suppressed) look up the field in the class and check annotations there. There was a second bug: The lint check which looks for invalid suppress annotation sites (since they're not allowed on local variables for class based lint checks) incorrectly concluded that these types of initializations were local variables rather than field initializations. This fixes issue 38626: ADT: Misleading lint rule: SimpleDateFormat Change-Id: I254f3fb5a6132d6cbe39bd425ffe6d67ed7b84ed
* 32849: Fix CRLF line ending handling in the partial formatterTor Norbye2012-10-171-0/+7
| | | | | | | | | | | | | | | | | | Fix the handling of \r characters in the code which applies formatting deltas into existing documents. It could end up inserting the formatted portion in the middle of a \r\n pair, which made Eclipse extremely confused (https://bugs.eclipse.org/bugs/show_bug.cgi?id=375421) This fixes 32849: Eclipse android adt xml editing artifacting (unsynced) lines of text when changes are made in the graphical interface It also adds a lint check to identify *existing* files that already have these mangled line endings, along with a quickfix to make the correction. Change-Id: I1e7024f2786e4cb0233c2c6b98c3d3f942703ea0
* Warn about potentially accidentally overriding methodsTor Norbye2012-10-162-1/+18
| | | | Change-Id: Ic8d20a002f6225cb15f609b3b7582716904e9278
* Add lint support for checking frameworks/base/coreTor Norbye2012-10-164-2/+96
| | | | | | | | | | This changeset adds basic support for running lint against the AOSP frameworks/base/core codebase. It also makes the resource folder provided by the lint client. Change-Id: I7e38b0925cb032f776c54f975b924b91d6ab7a24
* Check notification and action bar iconsTor Norbye2012-10-052-35/+46
| | | | | | | These icons should not use color (and in the case of notification icons, be white). Change-Id: I32b9422735830a01bb069b90a5ad5a76d7aeb5de
* Fix misc lint bugsTor Norbye2012-10-041-0/+17
| | | | | | | I ran lint over some large projects and manually verified the report; this uncovered a couple of false positives. Change-Id: I006d64a03119c6badc40e33df9ab4079e07dc7c5
* Cache applicable issue lists per scope setTor Norbye2012-09-271-6/+51
| | | | | | | | | | Keep sets of applicable issues for each scope set. This should make incremental lint slightly more efficient. Also add a check that icon file extensions match the actual file format. Change-Id: I4a7f5e19b91aa3613480aa831ec42065528f02d3
* 37733: TargetAPI doesn't fails to report unsupported APITor Norbye2012-09-252-3/+79
| | | | | | | Make incremental lint in source files use the Eclipse context to find super classes. Change-Id: I31af7a146c259875857026bc7721294205482a91
* Add labelFor lint checkTor Norbye2012-09-171-3/+4
| | | | Change-Id: If2b905197d5d81c6f98315a0d48ff37091e8282b
* Code cleanupTor Norbye2012-09-172-8/+8
| | | | | | Replace "this.mFoo = foo" with just "mFoo = foo". Change-Id: I44d537ebb62cd7e5c8493a0b50dcb527cd20f91c
* Constants refactoring.Tor Norbye2012-09-139-478/+113
| | | | | | | | | | | | | | | | This changeset moves most constants into the SdkConstants class, and gets rid of AndroidConstants and LintConstants. It also migrates all non-ADT specific constants from AdtConstants into SdkConstants. It furthermore moves various other constants (such as those in XmlUtils and ValuesDescriptors) into the constants class. It also fixes the modifier order to be the canonical modifier order (JLS 8.x). Finally, it removes redundancy and combines various constant aliases such that we don't have both NAME_ATTR and ATTR_NAME pointing to "name", etc. Change-Id: Ifd1755016f62ce2dd80e5c76130d6de4b0e32161
* Add unit test for lint CLITor Norbye2012-09-111-0/+12
| | | | Change-Id: I8434515179b8fd2ee3872506fc237c467a98db44
* Don't report unused resources in projects with errorsTor Norbye2012-09-101-0/+25
| | | | | | | | | | | | | | | | | | | If a project contains errors in the source code, lint won't be able to parse the source code, and in that case it might draw the wrong conclusions about unused resources. This can also happen if there's an actual bug in the Java parser, which is sometimes the case; see for example http://code.google.com/p/projectlombok/issues/detail?id=415 http://code.google.com/p/projectlombok/issues/detail?id=311 In both cases, when we encounter a failure to parse a Java file, we record the fact that not all Java files were properly processed, and rules, such as the UnusedResource detector, can (and now does) use this to for example skip reporting unused resources in this case since it is operating with incomplete data. Change-Id: I00991c10d05965ce151fb0dd322f32229dcd12cd
* 36829: Unused resource check does not handle merged manifestsTor Norbye2012-09-062-1/+24
| | | | Change-Id: I88cfbb3c497cc65e3d16c3ba1fe9ceeaf2613103
* 36951: Lint API Level Warnings for Enums are ConfusingTor Norbye2012-09-041-1/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | When using switch statements on enums, the compiler will generate extra code to make the switch invocation efficient. For example, it generates a method which looks up all the values in the enum, calls ordinal() on each, and then creates an array which has all the enum values indexed by ordinal. The problem is that this helper code will end up referencing fields which may have a higher API requirement than is currently being used, and Lint would flag these as illegal API references. That's not correct in this case; as long as the corresponding jump table does not actually use these references, there's no problem. This changeset makes lint aware of this pattern, and when it encounters an invalid field reference inside one of these synthetic switch constant lookup methods, it will search for the corresponding table lookup, check whether the actual key is used, and only in that case complain. It also makes error messages associated with enums in general a bit better, when the whole enum class itself requires a higher API. (This would generate a message related to invocation of ordinal() which might not be obvious to everybody.) Finally, it cleans up class name formatting such that we never reference inner classes using "$" in error messages. Change-Id: I28a14084c160866e69f63f62b5bc8f21ee50215c
* Add lint check for potential string locale errors (issue 36822)Tor Norbye2012-08-291-0/+4
| | | | Change-Id: I320af69286e7f0bcee395419acfeea7ed8f8d384
* Fix property sheet value completionTor Norbye2012-08-271-0/+7
| | | | | | | | | | | | | | | This changeset fixes the value completion such that you can add custom values into properties that also have enum fields. Rather than have separate completion routines for properties based on whether they contain an enum, a flag, a reference, etc., have a single completer which considers all the various formats and combines the results. In addition to combining results, this now also offers completion on dimensions, and offers theme attribute values for references as well. Change-Id: Idbc1799a34b3a3f14ea567654953925bf12afb8f
* Issue 34322: Fix handling of import statementsTor Norbye2012-08-271-0/+45
| | | | | | | | Fix the commit prefs lint check such that it correctly identifies scenarios where the innerclass "Editor" is referenced without its qualifying top level class, SharedPreferences. Change-Id: I3a22738508b66ce0b3e836feff91f816b2c368e8
* Use simple markup for bold and monospace in lint issuesTor Norbye2012-08-231-0/+154
| | | | | | | | | The explanation text for issues can now contain some simple markup such as * around sentences for bold and ` for monospace. In the HTML report this is converted to HTML markup using <b> and <code> spans, as well as <a> for http:// links. Change-Id: I875a55a0ab9a61c2e8bb493ad94d5e4d51dd1072
* Warn about non-recent targetSdkVersionsTor Norbye2012-08-233-0/+120
| | | | | | | | | | | | | | This changeset changes the default targetSdkVersion written into new projects from being hardcoded to "15" to being the same level as the build target's API level. It also adds a new lint check which looks at the targetSdkVersion and complains if it's not the same as the highest known version (currently 16). The issue explanation points to the javadocs for the android.os.Build.VERSION_CODES class for details on how to upgrade the app. Change-Id: I00c2bd7cd8fa239b9fd1de9b2d35ff9faf87d25f
* Fix class to source mapping for the super() keywordTor Norbye2012-08-212-6/+67
| | | | | | | | | | | | | The API checker needs to look up the corresponding source code source range for a given bytecode method dispatch, and the line number information found in the .class file is not accurate enough. Lint has some heuristics to search for the corresponding source range, but this did not work properly for cases where the method call was a parent constructor, via the "super()" call. It also fixes a bug in handling the default package. Change-Id: I15bb07b6709ab9a53376398fe92715f6e6eba968
* Refactor common.jarXavier Ducrohet2012-08-071-1/+1
| | | | | | | | | | | | | | | Move resources and com.android.util.Pair into layoutlib_api where they belong since layoutlib depends on them and we need to control the API. Made a copy of Pair to stay in common.jar but moved it to com.android.utils.Pair (the one in com.android.util.Pair is marked as deprecated to prevent usage where applicable). Also moved XmlUtil and PositionXmlParser to com.android.utils to match Pair. Change-Id: I21d7057d3f2ce604f86a3bb1fa3c130948c93b89
* 35875: Lint too strict about translating strings by default.Tor Norbye2012-08-061-0/+37
| | | | | | | | | | | | | | | | | | This changeset updates the issue explanation for the translation detector to explain the translatable=false and donottranslate.xml mechanisms to handle non-translatable strings. It adds a quickfix for missing translation items to set the translatable attribute to false. It also makes lint warn where you're using translatable=false on strings in a locale-folder, as well as when you're translating a string defined as translatable=false in the base folder. And finally it also bumps up the severity of the extra translations issue as justified in issue 35875. Change-Id: I7539464b234b0a4b444bf9f188ce5b819f962430
* Merge "Allow lint cli --sources to specify a path, and use from ant lint"Tor Norbye2012-08-061-0/+37
|\
| * Allow lint cli --sources to specify a path, and use from ant lintTor Norbye2012-08-061-0/+37
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | First, this changeset allows the arguments passed to --sources and --classpath (renamed from --classes) to specify not just a directory, but to specify a path as well. This might make it easier to invoke lint from scripts if you have a path variable, so you don't have to split it into multiple arguments. Second, it makes the lint task in ant use these, such that any build.xml customizations to the source paths or class paths are automatically used rather than relying on lint's default structure check. Change-Id: Id8e4caf0010d7fd7245844b3099b5dc0607f0aba
* | Merge "Improve Maven defaults in lint"Tor Norbye2012-08-061-12/+38
|\ \ | |/
| * Improve Maven defaults in lintTor Norbye2012-08-061-12/+38
| | | | | | | | Change-Id: I3c7c13437403fb1392343af079afab9434695566
* | 35804: When importantForAccessibility="no" is set, don't warnTor Norbye2012-08-031-1/+4
|/ | | | | | | | | | | | This changeset fixes 35804: ContentDescription should not flag if importantForAccessibility="no" When the above attribute is set on a View, the accessibility detector will not flag the corresponding view for missing a contentDescription attribute. Change-Id: I953235a04f1b1ee0f05a94604a2f364c09f851a1
* Allow custom source and class foldersTor Norbye2012-08-021-0/+7
| | | | | | | | | | | | Maven places the source files and class files in different locations than our default Eclipse and ant build systems. This changeset adds --sources and --classes arguments to the lint CLI to allow a build script to point to these custom locations. It also also looks in Maven's default location if it doesn't find the normal bin/classes/ folder such that out of the box it might just work even in Maven projects. Change-Id: Ia91b05f0fc9d2c7e504c962d2e794c768085754c
* Fix lint-on-save for .class file detectors in Java filesTor Norbye2012-08-014-8/+86
| | | | | | | | | | | | | | | | | | | | | | | | This changeset fixes the lint-on-save behavior in Java files such that the classfile based checks are run after the .class files are up to date. It also makes lint-on-save work when Project > Build Automatically is turned off, by adding a new resource listener, and it modifies the IFileListener interface to make resource listening more efficient; in particular, it passes the flag mask such that listeners can ignore events such as markers getting added or removed from a file without the content changing. It also makes some improvements to the lint infrastructure. First, it adds an indirection in the LintClient such that reading bytes from files can be customized by the client (to for example add caching or to read contents from memory not yet flushed to disk). It also allows inner classes to share the contents of the source file between each context (while debugging the above I noticed that each inner class node had its own class context and therefore would re-read the source file repeatedly.) Change-Id: Ib9572cebe1269fe05c3af1369610525ea3b44061
* 35588: @SuppressLint does not work on constructors in Java filesTor Norbye2012-07-271-0/+8
| | | | | | | | | | | | | The AST-based check for whether an issue is suppressed did not take constructors into account. This was because the AST node for constructors (ConstructorDeclaration) does not extend the AST node for method declarations (MethodDeclaration). This mistake was made in a couple of other detectors as well. (Also renamed test class which did not have the correct name (detector class + "Test") so jump to test did not work.) Change-Id: I0ecf99ab7d0357a03e300b7197ae84079f0ddbd1
* Improve handling of library projects in lintTor Norbye2012-07-242-1/+72
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This changeset improves the way lint handles library projects. Until now, running lint on project "master" would also look up any library projects referenced by the master project, and analyze the library projects as well. This is necessary in order to correctly compute unused resources for example, since it's possible for a resource to be defined in one project and referenced in another. However, while this behavior is desirable for users who partition their code up into library projects, it has some serious problems for users who are using a third party library project: - Their lint output can be swamped with errors from the library which they have no control over. - If for example the library provides translations into 60 languages, lint will use these 60 languages as the set of languages targeted by the application, and complain about all strings in the master project which are not translated into all the languages. This changeset makes a key change to how library projects are handled: (1) If you run lint on all projects (including the library projects), then there is no change from before. (2) If you run lint and specify just a project, then lint will continue to analyze the project as well as all its libraries, but will only report problems on the user-specified project. The way this is done is by a new "report errors" attribute stored with each project. All projects that are explicitly referenced on the command line (or selected in the Eclipse UI), and all projects that are found recursively if you specify a top level directory, all these projects have their "report errors" flag set. Any remaining projects, which would be those found through library project references, these have their report errors flag cleared. And whenever lint is processing errors, it will filter out errors for projects that are not reporting errors. This addresses issue http://code.google.com/p/android/issues/detail?id=33847 as well as a number of other requests (in StackOverflow and elsewhere) around the ability to filter errors in library projects. Change-Id: I9d084f598c678ecf79cfe70d8ea7a84844333acc
* Merge "Tweaks to the lint CLI offsets"Tor Norbye2012-07-231-1/+5
|\
| * Tweaks to the lint CLI offsetsTor Norbye2012-07-231-1/+5
| | | | | | | | Change-Id: I20592372d41235b36227a0d36014d77c23c7fc58
* | 35049: @SuppressLint("NewApi") doesn't work on local variablesTor Norbye2012-07-231-0/+10
|/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The @SuppressLint annotation can deliberately be placed on not only classes and methods but on parameters and local variables too. For AST-based (Java source-based) lint checks, this works fine. However, some lint checks, such as the API Check, is based on analyzing the bytecode. Annotations placed on local variables and parameters do not make it into the .class file, so these annotations do not work to suppress errors when placed on local variables (or parameters). The @TargetApi annotation applies only to the bytecode based API check, so its metadata only allows the annotation to be placed on methods and classes and constructors. However, the @SuppressLint annotation needs to continue to be available for the AST-based checks. This CL adds a new lint check, a "meta" check, which actually looks for invalid @SuppressLint annotations, and warns about these. With the new lint-on-save behavior, this means you instantly get feedback if you attempt to suppress an annotation in the wrong place. (Note that the quickfix for adding annotations has always enforced this and placed annotations out at the method level, but as shown in issue 35049, developers place them there deliberately themselves.) This CL also fixes an unrelated problem (shown in issue 34198) that the add suppress annotation code could sometimes add multiple versions of the same id into the annotation. Change-Id: I5bc61c6315edfcfc20103d1e580e389dd8e6a09b
* Add bytecode method call dispatch for lint rulesTor Norbye2012-07-237-25/+459
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This changeset adds shared dispatch for bytecode based rules. We already have this for XML and Java source trees: detectors can register interest in a particular element or attribute name, or a particular method call name, and then the source is processed in a single pass, and a quick map lookup dispatches to any detectors interested in that particular item. This has the important property that adding new detectors does not linearly slow down lint: it does not process each XML document an extra time for each newly added rule: the new rule just adds one more entry into a dispatch table, and only for the specifically interesting tag name (for example) is there an extra dispatch to the new detector. This changeset adds a similar mechanism for the ASM based bytecode detectors. Until now, each detector would be handed the class node of the outer class. This changeset adds a couple of new dispatch methods: First, a way to register interests in method calls of a particular type, and second, a way to registere interest in types of ASM nodes. In addition to the new visitor, this changeset rewrites a couple of detectors to use the new dispatch approach (there are new detectors in the pipeline which will also take advantage of this), and cleans up handling of positions a bit. Change-Id: Ib115bd3418b6c63bdcd49fec046e91a73daf38a9
* 35054: lint reports resources as unused when used in pluralsTor Norbye2012-07-171-0/+1
| | | | Change-Id: Id437ef30e1488c96a46ccea45459694fd23db582
* Add code completion and API checks for theme references (?attr)Tor Norbye2012-07-161-0/+1
| | | | | | | | | | | | | This checkin improves the handling of ?attr and ?android:attr theme references, such that (1) The API check considers theme references, and for example flags usage of ?android:attr/dividerHorizontal if your minSdkVersion is lower than 11 (2) Adds theme reference completion to the XML code completion. Change-Id: I2049b828e413802ac81579294f515fb0a4faccdf
* Improvements to bytecode-to-source offset computationTor Norbye2012-07-162-25/+253
| | | | | | | | This changeset improves the way lint searches for the correct source offsets (and line numbers) for symbols found in .class files. Among other improvements this also fixes issue 34686. Change-Id: I7a2f40f0d930a4786dc69a7f0e0bbf5367a30ad4
* Prevent exceptions for invalid project definitionsTor Norbye2012-07-031-0/+3
| | | | Change-Id: Ie2b5929a87144c77f1ce0d43b2520cc498df5081
* Merge "33985: Persistent errors in Java files"Tor Norbye2012-07-021-0/+5
|\
| * 33985: Persistent errors in Java filesTor Norbye2012-07-021-0/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This CL fixes issue 33985. There reason these error markers were sticky, is that the IssueRegistry.PARSER_ERROR error type is special, and was not handled properly by the Issue.isAdequate(Scope) method which is supposed to return whether the given issue can be fully analyzed in the given context. In addition, I also disabled the Java parser generating PARSER_ERROR issues when running in Eclipse, since Eclipse itself will already provide these errors, so all this achieves is creating "multiple annotations on this line" conflicts when you hover over the icon. Change-Id: Ie91604660a69eb007b851e5a821e629211c29690
generated by cgit v1.1 at 2025-08-21 21:58:15 +0000