update-tags-encoding: escape '/' and ',' in tags

This is more conflict resistant than the old encoding where tags was a
file with a line per tag.

src/comment.rs

@@ -212,16 +212,17 @@ mod tests {
 
     #[test]
     fn read_comment_0() {
-        let comment_dir =
+        let comment_dir = std::path::Path::new(
-            std::path::Path::new("test/0001/dd79c8cfb8beeacd0460429944b4ecbe95a31561/comments/9055dac36045fe36545bed7ae7b49347");
+            "test/0001/dd79c8cfb8beeacd0460429944b4ecbe/comments/9055dac36045fe36545bed7ae7b49347",
+        );
         let comment = Comment::new_from_dir(comment_dir).unwrap();
         let expected = Comment {
             uuid: String::from("9055dac36045fe36545bed7ae7b49347"),
             author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-07T15:26:26-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T10:08:38-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
-            description: String::from("This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe95a31561\n\nIt has multiple lines\n"),
+            description: String::from("This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe\n\nIt has multiple lines\n"),
             dir: std::path::PathBuf::from(comment_dir),
         };
         assert_eq!(comment, expected);

src/issue.rs

@@ -48,6 +48,10 @@ pub enum IssueError {
     ChronoParseError(#[from] chrono::format::ParseError),
     #[error("Failed to parse issue")]
     IssueParseError,
+    #[error("invalid escape character {escape:?} in tag file {filename:?}")]
+    TagInvalidEscape { escape: String, filename: String },
+    #[error("invalid trailing escape character ',' in tag file {filename:?}")]
+    TagTrailingEscape { filename: String },
     #[error("Failed to parse state")]
     StateParseError,
     #[error("Failed to run git")]
@@ -137,13 +141,7 @@ impl Issue {
                 } else if file_name == "dependencies" && direntry.metadata()?.is_dir() {
                     dependencies = Self::read_dependencies(&direntry.path())?;
                 } else if file_name == "tags" {
-                    let contents = std::fs::read_to_string(direntry.path())?;
+                    tags = Self::read_tags(&direntry)?;
-                    tags = contents
-                        .lines()
-                        .filter(|s| s.len() > 0)
-                        .map(|tag| String::from(tag.trim()))
-                        .collect();
-                    tags.sort();
                 } else if file_name == "comments" && direntry.metadata()?.is_dir() {
                     Self::read_comments(&mut comments, &direntry.path())?;
                 } else {
@@ -525,12 +523,82 @@ impl Issue {
         Ok(())
     }
 
+    fn read_tags(tags_direntry: &std::fs::DirEntry) -> Result<Vec<String>, IssueError> {
+        if !tags_direntry.metadata()?.is_dir() {
+            eprintln!("issue has old-style tags file");
+            return Err(IssueError::IssueParseError);
+        }
+        let mut tags = Vec::<String>::new();
+        for direntry in tags_direntry.path().read_dir()? {
+            if let Ok(direntry) = direntry {
+                let tag = Issue::tag_from_filename(&direntry.file_name().to_string_lossy())?;
+                tags.push(tag);
+            }
+        }
+        tags.sort();
+        Ok(tags)
+    }
+
+    /// Perform un-escape on a filename to make it into a tag:
+    /// ",0" => ","
+    /// ",1" => "/"
+    fn tag_from_filename(filename: &str) -> Result<String, IssueError> {
+        let mut tag = String::new();
+        let mut token_iter = filename.split(',');
+        let Some(start) = token_iter.next() else {
+            return Err(IssueError::StdIoError(std::io::Error::from(
+                std::io::ErrorKind::NotFound,
+            )));
+        };
+        tag.push_str(start);
+        for token in token_iter {
+            match token.chars().nth(0) {
+                Some('0') => {
+                    tag.push(',');
+                    tag.push_str(&token[1..]);
+                }
+                Some('1') => {
+                    tag.push('/');
+                    tag.push_str(&token[1..]);
+                }
+                Some(bogus) => {
+                    return Err(IssueError::TagInvalidEscape {
+                        escape: String::from(bogus),
+                        filename: String::from(filename),
+                    });
+                }
+                None => {
+                    return Err(IssueError::TagTrailingEscape {
+                        filename: String::from(filename),
+                    });
+                }
+            }
+        }
+        Ok(tag)
+    }
+
+    // Perform escape on a tag to make it into a filename:
+    // "," => ",0"
+    // "/" => ",1"
+    fn tag_to_filename(tag: &str) -> String {
+        let mut filename = tag.replace(",", ",0");
+        filename = filename.replace("/", ",1");
+        return filename;
+    }
+
     fn commit_tags(&self, commit_message: &str) -> Result<(), IssueError> {
-        let mut tags_filename = self.dir.clone();
+        let mut tags_dir_name = self.dir.clone();
-        tags_filename.push("tags");
+        tags_dir_name.push("tags");
-        let mut tags_file = std::fs::File::create(&tags_filename)?;
+        match std::fs::remove_dir_all(&tags_dir_name) {
+            Err(e) if e.kind() == std::io::ErrorKind::NotFound => (),
+            Err(e) => return Err(e.into()),
+            Ok(_) => (),
+        }
+        std::fs::create_dir(&tags_dir_name)?;
         for tag in &self.tags {
-            writeln!(tags_file, "{}", tag)?;
+            let mut tag_filename = tags_dir_name.clone();
+            tag_filename.push(Issue::tag_to_filename(tag));
+            std::fs::File::create(&tag_filename)?;
         }
         self.commit(commit_message)?;
         Ok(())
@@ -551,19 +619,118 @@ mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
+    #[test]
+    fn parse_tag_0() {
+        assert_eq!(
+            Issue::tag_from_filename("hello").unwrap(),
+            String::from("hello")
+        );
+    }
+
+    #[test]
+    fn parse_tag_1() {
+        assert_eq!(
+            Issue::tag_from_filename("hello,0world").unwrap(),
+            String::from("hello,world")
+        );
+    }
+
+    #[test]
+    fn parse_tag_2() {
+        assert_eq!(
+            Issue::tag_from_filename("hello,1world").unwrap(),
+            String::from("hello/world")
+        );
+    }
+
+    #[test]
+    fn parse_tag_3() {
+        assert_eq!(
+            Issue::tag_from_filename(",0hello,1world,0").unwrap(),
+            String::from(",hello/world,")
+        );
+    }
+
+    #[test]
+    fn parse_tag_4() {
+        // std::io::Error does not impl PartialEq :-(
+        let filename = "hello,";
+        match Issue::tag_from_filename(filename) {
+            Ok(tag) => panic!(
+                "tag_from_filename() accepted invalid input {:?} and returned {:?}",
+                filename, tag
+            ),
+            Err(_e) => (),
+        }
+    }
+
+    #[test]
+    fn parse_tag_5() {
+        // std::io::Error does not impl PartialEq :-(
+        let filename = "hello,world";
+        match Issue::tag_from_filename(filename) {
+            Ok(tag) => panic!(
+                "tag_from_filename() accepted invalid input {:?} and returned {:?}",
+                filename, tag
+            ),
+            Err(_e) => (),
+        }
+    }
+
+    #[test]
+    fn tag_to_filename_0() {
+        let tag = "hello";
+        assert_eq!(Issue::tag_to_filename(tag), "hello");
+    }
+
+    #[test]
+    fn tag_to_filename_1() {
+        let tag = "hello,";
+        assert_eq!(Issue::tag_to_filename(tag), "hello,0");
+    }
+
+    #[test]
+    fn tag_to_filename_2() {
+        let tag = "/hello";
+        assert_eq!(Issue::tag_to_filename(tag), ",1hello");
+    }
+
+    #[test]
+    fn tag_to_filename_3() {
+        let tag = "hello/bye,boo";
+        assert_eq!(Issue::tag_to_filename(tag), "hello,1bye,0boo");
+    }
+
+    #[test]
+    fn tag_to_filename_4() {
+        let tag = ",,,///,,,";
+        assert_eq!(Issue::tag_to_filename(tag), ",0,0,0,1,1,1,0,0,0");
+    }
+
+    #[test]
+    fn tag_to_filename_5() {
+        let tag = ",0,0,1,1";
+        assert_eq!(Issue::tag_to_filename(tag), ",00,00,01,01");
+    }
+
     #[test]
     fn read_issue_0() {
-        let issue_dir = std::path::Path::new("test/0000/3943fc5c173fdf41c0a22251593cd476d96e6c9f/");
+        let issue_dir = std::path::Path::new("test/0000/3943fc5c173fdf41c0a22251593cd476/");
         let issue = Issue::new_from_dir(issue_dir).unwrap();
         let expected = Issue {
-            id: String::from("3943fc5c173fdf41c0a22251593cd476d96e6c9f"),
+            id: String::from("3943fc5c173fdf41c0a22251593cd476"),
             author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T12:14:26-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:36:25-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
             done_time: None,
             tags: Vec::<String>::from([
                 String::from("TAG2"),
+                String::from("bird/wing"),
+                String::from("bird/wing/feather"),
+                String::from("deer,antler"),
+                String::from("deer,antler,tassle"),
+                String::from("hop,scotch/shoe"),
                 String::from("i-am-also-a-tag"),
                 String::from("tag1"),
             ]),
@@ -581,12 +748,12 @@ mod tests {
 
     #[test]
     fn read_issue_1() {
-        let issue_dir = std::path::Path::new("test/0000/7792b063eef6d33e7da5dc1856750c149ba678c6/");
+        let issue_dir = std::path::Path::new("test/0000/7792b063eef6d33e7da5dc1856750c14/");
         let issue = Issue::new_from_dir(issue_dir).unwrap();
         let expected = Issue {
-            id: String::from("7792b063eef6d33e7da5dc1856750c149ba678c6"),
+            id: String::from("7792b063eef6d33e7da5dc1856750c14"),
             author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T12:14:26-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:37:07-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
             done_time: None,

src/issues.rs

@@ -96,13 +96,13 @@ mod tests {
 
         let mut expected = Issues::new();
 
-        let uuid = String::from("7792b063eef6d33e7da5dc1856750c149ba678c6");
+        let uuid = String::from("7792b063eef6d33e7da5dc1856750c14");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(crate::issue::Issue {
             id: uuid,
             author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T12:14:26-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:37:07-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
             done_time: None,
@@ -115,19 +115,24 @@ mod tests {
             dir,
         });
 
-        let uuid = String::from("3943fc5c173fdf41c0a22251593cd476d96e6c9f");
+        let uuid = String::from("3943fc5c173fdf41c0a22251593cd476");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(
             crate::issue::Issue {
                 id: uuid,
                 author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T12:14:26-06:00")
+                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:36:25-06:00")
                     .unwrap()
                     .with_timezone(&chrono::Local),
                 done_time: None,
                 tags: Vec::<String>::from([
                     String::from("TAG2"),
+                    String::from("bird/wing"),
+                    String::from("bird/wing/feather"),
+                    String::from("deer,antler"),
+                    String::from("deer,antler,tassle"),
+                    String::from("hop,scotch/shoe"),
                     String::from("i-am-also-a-tag"),
                     String::from("tag1"),
                 ]),
@@ -149,13 +154,13 @@ mod tests {
 
         let mut expected = Issues::new();
 
-        let uuid = String::from("3fa5bfd93317ad25772680071d5ac3259cd2384f");
+        let uuid = String::from("3fa5bfd93317ad25772680071d5ac325");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(crate::issue::Issue {
             id: uuid,
             author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T11:59:44-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:37:46-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
             done_time: Some(
@@ -172,7 +177,7 @@ mod tests {
             dir,
         });
 
-        let uuid = String::from("dd79c8cfb8beeacd0460429944b4ecbe95a31561");
+        let uuid = String::from("dd79c8cfb8beeacd0460429944b4ecbe");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         let mut comment_dir = dir.clone();
@@ -184,8 +189,8 @@ mod tests {
             crate::comment::Comment {
                 uuid: comment_uuid,
                 author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-07T15:26:26-06:00").unwrap().with_timezone(&chrono::Local),
+                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T10:08:38-06:00").unwrap().with_timezone(&chrono::Local),
-                description: String::from("This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe95a31561\n\nIt has multiple lines\n"),
+                description: String::from("This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe\n\nIt has multiple lines\n"),
                 dir: std::path::PathBuf::from(comment_dir),
             }
         );
@@ -193,7 +198,7 @@ mod tests {
             crate::issue::Issue {
                 id: uuid,
                 author: String::from("Sebastian Kuzminsky <seb@highlab.com>"),
-                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-03T11:59:44-06:00")
+                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T10:08:24-06:00")
                     .unwrap()
                     .with_timezone(&chrono::Local),
                 done_time: None,
@@ -216,13 +221,13 @@ mod tests {
 
         let mut expected = Issues::new();
 
-        let uuid = String::from("3fa5bfd93317ad25772680071d5ac3259cd2384f");
+        let uuid = String::from("3fa5bfd93317ad25772680071d5ac325");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(crate::issue::Issue {
             id: uuid,
             author: String::from("sigil-03 <sigil@glyphs.tech>"),
-            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-05T13:55:49-06:00")
+            creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:38:40-06:00")
                 .unwrap()
                 .with_timezone(&chrono::Local),
             done_time: None,
@@ -235,14 +240,14 @@ mod tests {
             dir,
         });
 
-        let uuid = String::from("dd79c8cfb8beeacd0460429944b4ecbe95a31561");
+        let uuid = String::from("dd79c8cfb8beeacd0460429944b4ecbe");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(
             crate::issue::Issue {
                 id: uuid,
                 author: String::from("sigil-03 <sigil@glyphs.tech>"),
-                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-05T13:55:49-06:00")
+                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:39:20-06:00")
                     .unwrap()
                     .with_timezone(&chrono::Local),
                 done_time: None,
@@ -256,22 +261,22 @@ mod tests {
             },
         );
 
-        let uuid = String::from("a85f81fc5f14cb5d4851dd445dc9744c7f16ccc7");
+        let uuid = String::from("a85f81fc5f14cb5d4851dd445dc9744c");
         let mut dir = std::path::PathBuf::from(issues_dir);
         dir.push(&uuid);
         expected.add_issue(
             crate::issue::Issue {
                 id: uuid,
                 author: String::from("sigil-03 <sigil@glyphs.tech>"),
-                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-05T13:55:49-06:00")
+                creation_time: chrono::DateTime::parse_from_rfc3339("2025-07-24T08:39:02-06:00")
                     .unwrap()
                     .with_timezone(&chrono::Local),
                 done_time: None,
                 tags: Vec::<String>::new(),
                 state: crate::issue::State::WontDo,
                 dependencies: Some(vec![
-                    crate::issue::IssueHandle::from("3fa5bfd93317ad25772680071d5ac3259cd2384f"),
+                    crate::issue::IssueHandle::from("3fa5bfd93317ad25772680071d5ac325"),
-                    crate::issue::IssueHandle::from("dd79c8cfb8beeacd0460429944b4ecbe95a31561"),
+                    crate::issue::IssueHandle::from("dd79c8cfb8beeacd0460429944b4ecbe"),
                 ]),
                 assignee: None,
                 description: String::from("issue with dependencies\n\na test has begun\nfor dependencies we seek\nintertwining life"),

test/0000/3943fc5c173fdf41c0a22251593cd476d96e6c9f/tags

@@ -1,3 +0,0 @@
-tag1
-TAG2
-i-am-also-a-tag

test/0001/dd79c8cfb8beeacd0460429944b4ecbe/comments/9055dac36045fe36545bed7ae7b49347/description

@@ -0,0 +1,3 @@
+This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe
+
+It has multiple lines

test/0001/dd79c8cfb8beeacd0460429944b4ecbe95a31561/comments/9055dac36045fe36545bed7ae7b49347/description

@@ -1,3 +0,0 @@
-This is a comment on issue dd79c8cfb8beeacd0460429944b4ecbe95a31561
-
-It has multiple lines

tools/update-tags-encoding

@@ -0,0 +1,79 @@
+#!/bin/bash
+#
+# Check out the `entomologist-data` branch in a temporary worktree.
+# For each issue with a `tags` file, replace the old-style tags file with a new-style tags dir.
+# git commit
+
+set -e
+#set -x
+
+function escape_tag() {
+    TAG="$1"
+    TAG=$(echo "${TAG}" | sed -re 's/,/,0/g')
+    TAG=$(echo "${TAG}" | sed -re 's/\//,1/g')
+    echo "${TAG}"
+}
+
+BRANCH=""
+
+if [[ -n "$1" ]] && [[ -d "$1" ]]; then
+    echo "updating ent db in directory '$1'"
+    pushd "$1"
+else
+    if [[ -n "$1" ]]; then
+        # better be a branch
+        BRANCH="$1"
+    else
+        BRANCH="entomologist-data"
+    fi
+    echo "updating ent db in branch '${BRANCH}'"
+    WORKTREE_DIR=$(mktemp --directory)
+    git worktree add "${WORKTREE_DIR}" "${BRANCH}"
+    pushd "${WORKTREE_DIR}" > /dev/null
+fi
+
+# Now our current working directory is the ent db that we're supposed
+# to update.
+#
+# If $BRANCH is empty, we're in a directory not tracked by git and we
+# just change the files.
+#
+# If $BRANCH is not empty, we're in a git worktree of the branch we're
+# supposed to change, so we commit as we go.
+
+for ISSUE_ID in $(find . -maxdepth 1 -type d -regextype posix-extended -regex '\./[0-9a-f]{32}'); do
+    ISSUE_ID=$(basename "${ISSUE_ID}")
+    if ! [[ -f "${ISSUE_ID}/tags" ]]; then
+        continue
+    fi
+
+    pushd "${ISSUE_ID}" > /dev/null
+
+    echo "${ISSUE_ID} has tags:"
+    TAGS=$(cat tags)
+    echo "${TAGS}"
+    rm tags
+
+    if [[ -n "${BRANCH}" ]]; then
+        git rm -f tags
+    fi
+
+    mkdir tags
+    for TAG in ${TAGS}; do
+        TAG=$(escape_tag "${TAG}")
+        touch "tags/${TAG}"
+    done
+
+    if [[ -n "${BRANCH}" ]]; then
+        git add tags
+        git commit -m "issue ${ISSUE_ID}: update tags to new format"
+    fi
+
+    popd > /dev/null
+done
+
+popd > /dev/null
+
+if [[ -n "${BRANCH}" ]]; then
+    git worktree remove "${WORKTREE_DIR}"
+fi