Merge pull request #14 from isbm/isbm-cleanup-filters

Update extensions and replace hardcoded values
tinythings · Nov 2, 2023 · f78d245 · f78d245
2 parents 8e8c0f7 + 9455e78
commit f78d245
Show file tree

Hide file tree

Showing 2 changed files with 32 additions and 10 deletions.
diff --git a/src/filters/defs.rs b/src/filters/defs.rs
@@ -7,17 +7,31 @@ pub const DOC_STUB_FILES: &[&str] =
     &["AUTHORS", "COPYING", "LICENSE", "DEBUG", "DISTRIB", "DOC", "HISTORY", "README", "TERMS", "TODO"];
 
 /// Docfiles
-pub const DOC_F_EXT: &[&str] = &[".txt", ".doc", ".rtf", ".md", ".rtx", ".tex", ".xml"];
+pub const DOC_F_EXT: &[&str] = &[".txt", ".doc", ".rtf", ".md", ".rtx", ".tex", ".xml", ".htm", ".html", ".log"];
 
 /// Docfiles portable
 pub const DOC_FP_EXT: &[&str] = &[".eps", ".pdf", ".ps"];
 
+/// Typically, docs
+pub const DOC_LOCATIONS: &[&str] = &["/usr/share/doc"];
+
 /// Headers
 pub const H_SRC_F_EXT: &[&str] = &[".h", ".hpp"];
 
 /// Archives
 pub const ARC_F_EXT: &[&str] = &[".gz", ".bz2", ".xz", ".zip", ".tar"];
 
 /// Graphic files
-pub const IMG_F_EXT: &[&str] =
-    &[".bmp", ".jpg", ".jpeg", ".png", ".gif", ".xpm", ".xbm", ".tif", ".tiff", ".pbm", ".svg", ".ico"];
+pub const IMG_F_EXT: &[&str] = &[
+    ".ani", ".bmp", ".dib", ".pcx", ".jpg", ".jpeg", ".jpx", ".jxr", ".png", ".gif", ".xpm", ".xbm", ".tif", ".tiff", ".iff",
+    ".lbm", ".pbm", ".pgm", ".pict", ".svg", ".ico", ".ai",
+];
+
+/// Manpages
+pub const D_MANPAGES: &str = "/usr/share/man";
+
+/// Localisation
+pub const D_L10N: &str = "/usr/share/locale";
+
+/// Intetrnaetiomns... i18n
+pub const D_I18N: &str = "/usr/share/i18n";
diff --git a/src/filters/texts.rs b/src/filters/texts.rs
@@ -4,6 +4,8 @@ use std::{
     path::{Path, PathBuf},
 };
 
+use super::defs;
+
 pub struct TextDataFilter {
     data: HashSet<PathBuf>,
     remove_manpages: bool,
@@ -45,7 +47,7 @@ impl TextDataFilter {
         _p.pop();
 
         self.remove_manpages
-            && p.to_str().unwrap().starts_with("/usr/share/man")
+            && p.to_str().unwrap().starts_with(defs::D_MANPAGES)
             && _p.file_name().unwrap().to_str().unwrap().starts_with("man")
     }
 
@@ -55,21 +57,27 @@ impl TextDataFilter {
             return false;
         }
 
+        for c in defs::DOC_STUB_FILES {
+            if p.file_name().unwrap_or_default().to_str().unwrap_or_default().contains(c) {
+                return true;
+            }
+        }
+
         let p = p.to_str().unwrap();
 
-        for c in ["/doc/", "changelog", "readme"] {
-            if p.to_lowercase().contains(c) {
+        for c in ["/doc/"].iter().chain(defs::DOC_STUB_FILES.iter()) {
+            if p.to_lowercase().contains(c.to_lowercase().as_str()) {
                 return true;
             }
         }
 
-        for c in ["/usr/share/doc"] {
+        for c in defs::DOC_LOCATIONS {
             if p.starts_with(c) {
                 return true;
             }
         }
 
-        for c in [".txt", ".doc", ".md", ".rtx", ".htm", ".html"] {
+        for c in defs::DOC_F_EXT {
             if p.ends_with(c) {
                 return true;
             }
@@ -80,12 +88,12 @@ impl TextDataFilter {
 
     /// Is localisation
     fn filter_l10n(&self, p: &Path) -> bool {
-        self.remove_l10n && p.to_str().unwrap().starts_with("/usr/share/locale")
+        self.remove_l10n && p.to_str().unwrap().starts_with(defs::D_L10N)
     }
 
     /// Is internationalisation
     fn filter_i18n(&self, p: &Path) -> bool {
-        self.remove_i18n && p.to_str().unwrap().starts_with("/usr/share/i18n")
+        self.remove_i18n && p.to_str().unwrap().starts_with(defs::D_I18N)
     }
 }