chore: refactor and comment

Signed-off-by: blam <ben@blam.sh>

.changeset/nine-mails-crash.md

@@ -6,4 +6,4 @@ The `ZipArchiveResponse` now correctly handles corrupt ZIP archives.
 
 Before this change, certain corrupt ZIP archives either cause the inflater to throw (as expected), or will hang the parser indefinitely.
 
-By switching out the `zip` parsing library, we now write to a temporary directory, and load from disk which should ensure that the parsing of the `.zip` files are done correctly because `streaming` of `zip` paths is technically impossible without being able to parse the headers at the end of he file.
+By switching out the `zip` parsing library, we now write to a temporary directory, and load from disk which should ensure that the parsing of the `.zip` files are done correctly because `streaming` of `zip` paths is technically impossible without being able to parse the headers at the end of the file.

packages/backend-common/src/reading/tree/ZipArchiveResponse.ts

@@ -111,7 +111,8 @@ export class ZipArchiveResponse implements ReadTreeResponse {
         }
 
         zipfile.on('entry', async (entry: Entry) => {
-          if (!/\/$/.test(entry.fileName) && this.shouldBeIncluded(entry)) {
+          // Check that the file is not a directory, and that is matches the filter.
+          if (!entry.fileName.endsWith('/') && this.shouldBeIncluded(entry)) {
             zipfile.openReadStream(entry, async (openErr, readStream) => {
               if (openErr || !readStream) {
                 reject(