feat(bundle-jvm): Automatically collect JVM source files with respecting excludes and gitignore

CHANGELOG.md

@@ -2,6 +2,11 @@
 
 ## Unreleased
 
+### Features
+
+- (bundle-jvm) Allow running directly on a project root (including multi-module repos) by automatically collecting only JVM source files (`.java`, `.kt`, `.scala`, `.groovy`), respecting `.gitignore`, and excluding common build output directories ([#3260](https://github.com/getsentry/sentry-cli/pull/3260))
+- (bundle-jvm) Add `--exclude` option for custom glob patterns to exclude files/directories from source collection ([#3260](https://github.com/getsentry/sentry-cli/pull/3260))
+
 ### Fixes
 
 - Replace `eprintln!` with `log::info!` for progress bar completion messages when the progress bar is disabled (e.g. in CI). This avoids spurious stderr output that some CI systems treat as errors ([#3223](https://github.com/getsentry/sentry-cli/pull/3223)).

src/commands/debug_files/bundle_jvm.rs

@@ -7,15 +7,70 @@ use crate::utils::file_upload::SourceFile;
 use crate::utils::fs::path_as_url;
 use crate::utils::source_bundle::{self, BundleContext};
 use anyhow::{bail, Context as _, Result};
-use clap::{Arg, ArgMatches, Command};
+use clap::{Arg, ArgAction, ArgMatches, Command};
+use log::debug;
 use sentry::types::DebugId;
+use std::borrow::Cow;
 use std::collections::BTreeMap;
+use std::ffi::OsStr;
 use std::fs;
-use std::path::PathBuf;
+use std::path::{Path, PathBuf};
 use std::str::FromStr as _;
 use std::sync::Arc;
 use symbolic::debuginfo::sourcebundle::SourceFileType;
 
+/// File extensions for JVM-based languages.
+const JVM_EXTENSIONS: &[&str] = &[
+    "java", "kt", "scala", "sc", "groovy", "gvy", "gy", "gsh", "clj", "cljc",
+];
+
+/// Directory patterns that are always safe to exclude globally (can never be
+/// valid JVM package names due to leading dots or conventions).
+const SAFE_EXCLUDES: &[&str] = &[
+    "!.cxx",
+    "!.eclipse",
+    "!.fleet",
+    "!.gradle",
+    "!.idea",
+    "!.kotlin",
+    "!.mvn",
+    "!.settings",
+    "!.vscode",
+    "!node_modules",
+];
+
+/// Directory names that are common build output dirs but could also be valid
+/// JVM package names (e.g. `com.example.build`). These are only excluded when
+/// they appear outside of `src/` directories to avoid filtering out legitimate
+/// source packages.
+const AMBIGUOUS_EXCLUDES: &[&str] = &["bin", "build", "out", "target"];
+
+/// Returns true if the file should be excluded because it sits inside an
+/// ambiguous build-output directory that is NOT under a `src/` ancestor.
+///
+/// We check *all* ambiguous directories in the path and exclude if any of them
+/// is not under a `src/` ancestor. This handles nested cases like
+/// `build/src/main/java/com/example/target/Foo.java` where the inner `target`
+/// is under `src`, but the outer `build` is not — the file should be excluded.
+fn is_in_ambiguous_build_dir(relative_path: &Path) -> bool {
+    for ancestor in relative_path.ancestors() {
+        let Some(name) = ancestor.file_name().and_then(|n| n.to_str()) else {
+            continue;
+        };
+        if AMBIGUOUS_EXCLUDES.contains(&name) {
+            // Check if any ancestor *above* this directory is named "src".
+            let has_src_above = ancestor
+                .ancestors()
+                .skip(1) // skip the ambiguous dir itself
+                .any(|a| a.file_name() == Some(OsStr::new("src")));
+            if !has_src_above {
+                return true;
+            }
+        }
+    }
+    false
+}
+
 pub fn make_command(command: Command) -> Command {
     command
         .hide(true) // experimental for now
@@ -47,6 +102,17 @@ pub fn make_command(command: Command) -> Command {
                 .value_parser(DebugId::from_str)
                 .help("Debug ID (UUID) to use for the source bundle."),
         )
+        .arg(
+            Arg::new("exclude")
+                .long("exclude")
+                .value_name("PATTERN")
+                .action(ArgAction::Append)
+                .help(
+                    "Glob pattern to exclude files/directories. Can be repeated. \
+                     By default, common build output and IDE directories are excluded \
+                     (build, .gradle, target, .idea, .vscode, out, bin, etc.).",
+                ),
+        )
 }
 
 pub fn execute(matches: &ArgMatches) -> Result<()> {
@@ -75,7 +141,36 @@ pub fn execute(matches: &ArgMatches) -> Result<()> {
         ))?;
     }
 
-    let sources = ReleaseFileSearch::new(path.clone()).collect_files()?;
+    let user_excludes = matches
+        .get_many::<String>("exclude")
+        .into_iter()
+        .flatten()
+        .map(|v| format!("!{v}"));
+
+    let all_excludes = SAFE_EXCLUDES
+        .iter()
+        .copied()
+        .map(Cow::Borrowed)
+        .chain(user_excludes.map(Cow::Owned));
+
+    let sources = ReleaseFileSearch::new(path.clone())
+        .extensions(JVM_EXTENSIONS.iter().copied())
+        .ignores(all_excludes)
+        .respect_ignores(true)
+        .collect_files()?;
+
+    let sources: Vec<_> = sources
+        .into_iter()
+        .filter(|source| {
+            let relative = source.path.strip_prefix(&source.base_path).unwrap();
+            if is_in_ambiguous_build_dir(relative) {
+                debug!("excluding (build output): {}", source.path.display());
+                return false;
+            }
+            true
+        })
+        .collect();
+
     let files = sources.iter().map(|source| {
         let local_path = source.path.strip_prefix(&source.base_path).unwrap();
         let local_path_jvm_ext = local_path.with_extension("jvm");
@@ -100,3 +195,72 @@ pub fn execute(matches: &ArgMatches) -> Result<()> {
 
     Ok(())
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::path::Path;
+
+    #[test]
+    fn test_excludes_build_output_at_module_root() {
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "app/build/generated/Foo.java"
+        )));
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "build/generated/Foo.java"
+        )));
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "module/target/classes/Foo.java"
+        )));
+        assert!(is_in_ambiguous_build_dir(Path::new("bin/Foo.class")));
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "out/production/Foo.java"
+        )));
+    }
+
+    #[test]
+    fn test_keeps_source_packages_under_src() {
+        assert!(!is_in_ambiguous_build_dir(Path::new(
+            "src/main/java/com/example/build/Builder.java"
+        )));
+        assert!(!is_in_ambiguous_build_dir(Path::new(
+            "app/src/main/java/com/example/target/Target.java"
+        )));
+        assert!(!is_in_ambiguous_build_dir(Path::new(
+            "src/main/kotlin/com/example/out/Output.kt"
+        )));
+    }
+
+    #[test]
+    fn test_excludes_build_dir_containing_src() {
+        // build/src/... should still be excluded — src is *inside* build, not above it
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "build/src/main/java/Foo.java"
+        )));
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "app/build/src/generated/Foo.java"
+        )));
+    }
+
+    #[test]
+    fn test_excludes_nested_ambiguous_dirs_under_build() {
+        // build/src/.../target/ — inner `target` is under src, but outer `build` is not
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "build/src/main/java/com/example/target/Foo.java"
+        )));
+        assert!(is_in_ambiguous_build_dir(Path::new(
+            "target/src/main/java/com/example/out/Foo.java"
+        )));
+    }
+
+    #[test]
+    fn test_keeps_files_without_ambiguous_dirs() {
+        assert!(!is_in_ambiguous_build_dir(Path::new(
+            "src/main/java/com/example/Foo.java"
+        )));
+        assert!(!is_in_ambiguous_build_dir(Path::new("Foo.java")));
+        assert!(!is_in_ambiguous_build_dir(Path::new(
+            "app/src/main/java/Foo.java"
+        )));
+    }
+}

src/utils/file_search.rs

@@ -20,6 +20,7 @@ pub struct ReleaseFileSearch {
     ignores: BTreeSet<String>,
     ignore_file: Option<String>,
     decompress: bool,
+    respect_ignores: bool,
 }
 
 #[derive(Eq, PartialEq, Hash)]
@@ -37,6 +38,7 @@ impl ReleaseFileSearch {
             ignore_file: None,
             ignores: BTreeSet::new(),
             decompress: false,
+            respect_ignores: false,
         }
     }
 
@@ -78,6 +80,11 @@ impl ReleaseFileSearch {
         self
     }
 
+    pub fn respect_ignores(&mut self, respect: bool) -> &mut Self {
+        self.respect_ignores = respect;
+        self
+    }
+
     pub fn collect_file(path: PathBuf) -> Result<ReleaseFileMatch> {
         // NOTE: `collect_file` currently do not handle gzip decompression,
         // as its mostly used for 3rd tools like xcode or gradle.
@@ -105,11 +112,11 @@ impl ReleaseFileSearch {
         let mut collected = Vec::new();
 
         let mut builder = WalkBuilder::new(&self.path);
-        builder
-            .follow_links(true)
-            .git_exclude(false)
-            .git_ignore(false)
-            .ignore(false);
+        builder.follow_links(true);
+
+        if !self.respect_ignores {
+            builder.git_exclude(false).git_ignore(false).ignore(false);
+        }
 
         if !&self.extensions.is_empty() {
             let mut types_builder = TypesBuilder::new();

tests/integration/_cases/debug_files/debug_files-bundle-jvm-help.trycmd

@@ -18,6 +18,9 @@ Options:
       --debug-id <UUID>          Debug ID (UUID) to use for the source bundle.
       --log-level <LOG_LEVEL>    Set the log output verbosity. [possible values: trace, debug, info,
                                  warn, error]
+      --exclude <PATTERN>        Glob pattern to exclude files/directories. Can be repeated. By
+                                 default, common build output and IDE directories are excluded
+                                 (build, .gradle, target, .idea, .vscode, out, bin, etc.).
       --quiet                    Do not print any output while preserving correct exit code. This
                                  flag is currently implemented only for selected subcommands.
                                  [aliases: --silent]