Merge pull request #28 from seapagan/exclude-only

.github/workflows/test.yml

@@ -0,0 +1,21 @@
+name: Rust
+
+on:
+  push:
+    branches: ["main"]
+  pull_request:
+    branches: ["main"]
+
+env:
+  CARGO_TERM_COLOR: always
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+      - name: Build
+        run: cargo build --verbose
+      - name: Run tests
+        run: cargo test --verbose

Makefile.toml

@@ -1,3 +1,8 @@
-[tasks.cov]
+[tasks.test]
+# this needs the crates 'nextest' and 'cargo-llvm-cov' installed locally
 command = "cargo"
-args = ["tarpaulin", "--out", "html"]
+args = ["llvm-cov", "nextest"]
+
+[tasks.test-html]
+command = "cargo"
+args = ["llvm-cov", "nextest", "--html"]

README-cratesio.md

@@ -306,6 +306,7 @@ Options:
   -l, --lnumbers            Add line numbers to each code file in the output.
   -t, --token <TOKEN>       GitHub personal access token (required for private repos and to pass rate limits)
   -e, --extend-exclude <PATTERN>  Additional file pattern to exclude (can be specified multiple times)
+  -x, --exclude <PATTERN>   File pattern to exclude, replacing the default ignore list (can be specified multiple times)
   -V, --version             Print version information and exit
   -h, --help                Print help
 ```
@@ -351,31 +352,41 @@ Available configuration options:
 - `line_numbers`: Whether to add line numbers by default (default: false)
 - `token`: Your GitHub personal access token (default: none)
 - `extend_exclude`: Additional file patterns to exclude (default: none)
+- `exclude`: File patterns to exclude, replacing the default ignore list
+  (default: none)
 
-The `extend_exclude` option can be specified either by using multiple `-e` flags
-on the command line:
+The `extend_exclude` and `exclude` options can be specified either by using
+multiple `-e` or `-x` flags on the command line:
 
 ```bash
 bundlerepo user/repo -e "*.md" -e "*.txt" -e "docs/*"
+bundlerepo user/repo -x "*.exe" -x "*.dll" -x "node_modules/*"
 ```
 
-Or as an array in the TOML configuration file:
+Or as arrays in the TOML configuration file:
 
 ```toml
 extend_exclude = ["*.md", "*.txt", "docs/*"]
+exclude = ["*.exe", "*.dll", "node_modules/*"]
 ```
 
-These patterns will be **added** to the default ignore list.
-
-> The `extend_exclude` option is useful for excluding additional files that
-> aren't in the default ignore list but that you don't want to include in your
-> XML output. This can help reduce token usage and remove irrelevant files from
-> the LLM context.
->
-> Storing your GitHub token in the configuration file can be more convenient
-> than passing it via command line, especially if you frequently work with
-> private repositories. Just be sure to keep your configuration file secure -
-> use the Global TOML file rather than the Local (in repo) TOML file.
+The `extend_exclude` patterns will be **added** to the default ignore list,
+while the `exclude` patterns will **replace** the default ignore list entirely.
+
+**Important**: When the `exclude` option is used (either via command line or
+config file), both the default ignore list and any `extend_exclude` patterns are
+completely ignored. The `exclude` patterns become the only ignore rules in
+effect.
+
+**Note**: The `extend_exclude` option is useful for excluding additional files
+that aren't in the default ignore list but that you don't want to include in
+your XML output. The `exclude` option gives you complete control over what files
+are ignored, replacing the built-in ignore list. Both options can help reduce
+token usage and remove irrelevant files from the LLM context.
+
+Storing your GitHub token in the configuration file can be more convenient than
+passing it via command line, especially if you frequently work with private
+repositories. Just be sure to keep your configuration file secure.
 
 ## Ignored Files
 

README.md

@@ -319,6 +319,7 @@ Options:
   -l, --lnumbers            Add line numbers to each code file in the output.
   -t, --token <TOKEN>       GitHub personal access token (required for private repos and to pass rate limits)
   -e, --extend-exclude <PATTERN>  Additional file pattern to exclude (can be specified multiple times)
+  -x, --exclude <PATTERN>   File pattern to exclude, replacing the default ignore list (can be specified multiple times)
   -V, --version            Print version information and exit
   -h, --help               Print help
 ```
@@ -345,6 +346,7 @@ clipboard = false
 line_numbers = true
 token = "your-github-token"
 extend_exclude = ["*.md", "*.txt", "docs/*"]  # Additional patterns to exclude
+exclude = ["*.exe", "*.dll", "node_modules/*"]  # File patterns to exclude
 ```
 
 All settings are optional. Settings are applied in the following order of
@@ -364,28 +366,40 @@ Available configuration options:
 - `line_numbers`: Whether to add line numbers by default (default: false)
 - `token`: Your GitHub personal access token (default: none)
 - `extend_exclude`: Additional file patterns to exclude (default: none)
+- `exclude`: File patterns to exclude, replacing the default ignore list
+  (default: none)
 
-The `extend_exclude` option can be specified either by using multiple `-e` flags
-on the command line:
+The `extend_exclude` and `exclude` options can be specified either by using
+multiple `-e` or `-x` flags on the command line:
 
 ```bash
 bundlerepo user/repo -e "*.md" -e "*.txt" -e "docs/*"
+bundlerepo user/repo -x "*.exe" -x "*.dll" -x "node_modules/*"
 ```
 
-Or as an array in the TOML configuration file:
+Or as arrays in the TOML configuration file:
 
 ```toml
 extend_exclude = ["*.md", "*.txt", "docs/*"]
+exclude = ["*.exe", "*.dll", "node_modules/*"]
 ```
 
-These patterns will be **added** to the default ignore list.
+The `extend_exclude` patterns will be **added** to the default ignore list,
+while the `exclude` patterns will **replace** the default ignore list entirely.
+
+> [!IMPORTANT]
+>
+> When the `exclude` option is used (either via command line or config file),
+> both the default ignore list and any `extend_exclude` patterns are completely
+> ignored. The `exclude` patterns become the only ignore rules in effect.
 
 > [!TIP]
 >
 > The `extend_exclude` option is useful for excluding additional files that
 > aren't in the default ignore list but that you don't want to include in your
-> XML output. This can help reduce token usage and remove irrelevant files from
-> the LLM context.
+> XML output. The `exclude` option gives you complete control over what files
+> are ignored, replacing the built-in ignore list. Both options can help reduce
+> token usage and remove irrelevant files from the LLM context.
 >
 > Storing your GitHub token in the configuration file can be more convenient
 > than passing it via command line, especially if you frequently work with

src/cli.rs

@@ -94,6 +94,15 @@ pub struct Flags {
         action = ArgAction::Append
     )]
     pub extend_exclude: Option<Vec<String>>,
+
+    #[arg(
+        long = "exclude",
+        short = 'x',
+        value_name = "PATTERN",
+        help = "Replace the existing exclude patterns with the specified pattern(s). Can be specified multiple times.",
+        action = ArgAction::Append
+    )]
+    pub exclude: Option<Vec<String>>,
 }
 
 pub fn version_info() -> String {

src/filelist.rs

@@ -17,32 +17,45 @@ pub struct FileTree {
 pub fn list_files_in_repo(
     repo_path: &PathBuf,
     extend_exclude: Option<&[String]>,
+    exclude: Option<&[String]>,
 ) -> Vec<String> {
     let mut file_list = Vec::new();
 
-    // Base ignore patterns as String
-    let mut ignore_patterns: Vec<String> = vec![
-        r"(?i)\.gitignore",
-        r"(?i)renovate\.json",
-        r"(?i)requirement.*\.txt",
-        r"(?i)\.lock$",
-        r"(?i)license(\..*)?",
-        // r"(?i)todo\..*",
-        r"(?i)\.github",
-        r"(?i)\.git",
-        r"(?i)\.vscode",
-    ]
-    .into_iter()
-    .map(String::from)
-    .collect();
-
-    // Add additional patterns if provided
-    if let Some(patterns) = extend_exclude {
-        ignore_patterns.extend(patterns.iter().map(|p| {
-            let escaped = regex::escape(p);
-            format!(r"(?i){}", escaped)
-        }));
-    }
+    // Initialize ignore patterns based on whether exclude is set
+    let ignore_patterns: Vec<String> = if let Some(patterns) = exclude {
+        // If exclude is set, use only those patterns
+        patterns
+            .iter()
+            .map(|p| {
+                let escaped = regex::escape(p);
+                format!(r"(?i){}", escaped)
+            })
+            .collect()
+    } else {
+        // Otherwise use default patterns
+        let mut patterns: Vec<String> = vec![
+            r"(?i)\.gitignore",
+            r"(?i)renovate\.json",
+            r"(?i)requirement.*\.txt",
+            r"(?i)\.lock$",
+            r"(?i)license(\..*)?",
+            r"(?i)\.github",
+            r"(?i)\.git",
+            r"(?i)\.vscode",
+        ]
+        .into_iter()
+        .map(String::from)
+        .collect();
+
+        // Add additional patterns if provided
+        if let Some(extend_patterns) = extend_exclude {
+            patterns.extend(extend_patterns.iter().map(|p| {
+                let escaped = regex::escape(p);
+                format!(r"(?i){}", escaped)
+            }));
+        }
+        patterns
+    };
 
     let regex_list: Vec<Regex> = ignore_patterns
         .iter()