Fix XML parsing #4

pavulon commented 2026年06月28日 20:35:37 +02:00

I've found that the simple XML parser util doesn't handle self-closing tags properly, so for a response like this:

Example

<?xml version='1.0' encoding='UTF-8'?>
<ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
 <Name>bucket</Name>
 <Prefix/>
 <KeyMarker/>
 <VersionIdMarker/>
 <MaxKeys>1000</MaxKeys>
 <IsTruncated>false</IsTruncated>
 <Version>
 <Key>file1.jpg</Key>
 <VersionId>1781648073.939709</VersionId>
 <IsLatest>true</IsLatest>
 <LastModified>2026年06月16日T22:14:34.000Z</LastModified>
 <Size>26702</Size>
 <Owner>
 <ID>user</ID>
 <DisplayName>user</DisplayName>
 </Owner>
 <StorageClass>STANDARD</StorageClass>
 </Version>
 <Version>
 <Key>file2.mp4</Key>
 <VersionId>1781648486.233691</VersionId>
 <IsLatest>true</IsLatest>
 <LastModified>2026年06月16日T22:21:26.000Z</LastModified>
 <Size>3217865</Size>
 <Owner>
 <ID>user</ID>
 <DisplayName>user</DisplayName>
 </Owner>
 <StorageClass>STANDARD</StorageClass>
 </Version>
</ListVersionsResult>

it would produce completely borked json

Before

After

{ "ListVersionsResult": { "Name": "bucket", "Key": { "MaxKeys": "1000", "IsTruncated": "false" }, "VersionId": "1781648073.939709", "IsLatest": "true", "LastModified": "2026年06月16日T22:14:34.000Z", "Size": "26702", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD", "Version": { "Key": "file2.mp4", "VersionId": "1781648486.233691", "IsLatest": "true", "LastModified": "2026年06月16日T22:21:26.000Z", "Size": "3217865", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" } } }

{ "ListVersionsResult": { "Name": "bucket", "Prefix": "", "KeyMarker": "", "VersionIdMarker": "", "MaxKeys": "1000", "IsTruncated": "false", "Version": [ { "Key": "file1.jpg", "VersionId": "1781648073.939709", "IsLatest": "true", "LastModified": "2026年06月16日T22:14:34.000Z", "Size": "26702", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" }, { "Key": "file2.mp4", "VersionId": "1781648486.233691", "IsLatest": "true", "LastModified": "2026年06月16日T22:21:26.000Z", "Size": "3217865", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" } ] } }

I've also added some tests.

test:unit was using the default jest.config.js, so I've created a separate config so that

• you don't need to set provider env variables to run it (seems like jest.config.js is meant mostly for test:providers?)
• you can import source files in tests directly

I've found that the simple XML parser util doesn't handle self-closing tags properly, so for a response like this: <details> <summary>Example</summary> ```xml <?xml version='1.0' encoding='UTF-8'?> <ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"> <Name>bucket</Name> <Prefix/> <KeyMarker/> <VersionIdMarker/> <MaxKeys>1000</MaxKeys> <IsTruncated>false</IsTruncated> <Version> <Key>file1.jpg</Key> <VersionId>1781648073.939709</VersionId> <IsLatest>true</IsLatest> <LastModified>2026年06月16日T22:14:34.000Z</LastModified> <Size>26702</Size> <Owner> <ID>user</ID> <DisplayName>user</DisplayName> </Owner> <StorageClass>STANDARD</StorageClass> </Version> <Version> <Key>file2.mp4</Key> <VersionId>1781648486.233691</VersionId> <IsLatest>true</IsLatest> <LastModified>2026年06月16日T22:21:26.000Z</LastModified> <Size>3217865</Size> <Owner> <ID>user</ID> <DisplayName>user</DisplayName> </Owner> <StorageClass>STANDARD</StorageClass> </Version> </ListVersionsResult> ``` </details> it would produce completely borked json <table> <tr> <th>Before</th> <th>After</th> </tr> <tr> <td> ```json { "ListVersionsResult": { "Name": "bucket", "Key": { "MaxKeys": "1000", "IsTruncated": "false" }, "VersionId": "1781648073.939709", "IsLatest": "true", "LastModified": "2026年06月16日T22:14:34.000Z", "Size": "26702", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD", "Version": { "Key": "file2.mp4", "VersionId": "1781648486.233691", "IsLatest": "true", "LastModified": "2026年06月16日T22:21:26.000Z", "Size": "3217865", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" } } } ``` </td> <td> ```json { "ListVersionsResult": { "Name": "bucket", "Prefix": "", "KeyMarker": "", "VersionIdMarker": "", "MaxKeys": "1000", "IsTruncated": "false", "Version": [ { "Key": "file1.jpg", "VersionId": "1781648073.939709", "IsLatest": "true", "LastModified": "2026年06月16日T22:14:34.000Z", "Size": "26702", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" }, { "Key": "file2.mp4", "VersionId": "1781648486.233691", "IsLatest": "true", "LastModified": "2026年06月16日T22:21:26.000Z", "Size": "3217865", "Owner": { "ID": "user", "DisplayName": "user" }, "StorageClass": "STANDARD" } ] } } ``` </td> </tr> </table> I've also added some tests. `test:unit` was using the default `jest.config.js`, so I've created a separate config so that - you don't need to set provider env variables to run it (seems like `jest.config.js` is meant mostly for `test:providers`?) - you can import source files in tests directly

Jolly_Good commented 2026年06月30日 15:52:57 +02:00

@pavulon thanks a lot for catching this and your contribtion! ❤️

I made a tiny adjustment to that REGEX ... test seems to be passing and SonarQube is also happy. Could you please checkout the DEV and doublecheck if its all fine.

@pavulon thanks a lot for catching this and your contribtion! ❤️ I made a tiny adjustment to that REGEX ... test seems to be passing and SonarQube is also happy. Could you please checkout the DEV and doublecheck if its all fine.

pavulon commented 2026年07月02日 22:47:13 +02:00

Looks good, although figuring out why it had non-linear backtracking goes way over my head 😅

By the way, for @babel/preset-typescript I used version 7.x because jest is still using it.
With 8.x on dev branch I get a bunch of warnings when I run npm install

Looks good, although figuring out why it had non-linear backtracking goes way over my head 😅 By the way, for `@babel/preset-typescript` I used version 7.x because `jest` is still using it. \ With 8.x on `dev` branch I get a bunch of warnings when I run `npm install`

Jolly_Good commented 2026年07月03日 09:37:45 +02:00

I despise every extra dependency ... how about switching to ts-jest - do you have any experience with that?

I despise every extra dependency ... how about switching to `ts-jest` - do you have any experience with that?