Report Files content is assumed to always be text

[miguel-ferreira-cko]

Environment

• Checkout SDK version: 4.0.15
• Platform and version: .NET 7.0.203
• Operating System and version: Windows 22H2

Description

Report File content should not always be treated as text. In some cases it can be a binary format (e.g. PDF) which cannot be encoded as a text in .NET (be it ASCII, UTF8...) and must be treated as a byte array / stream / base64 so that it can be properly stored.

Expected behavior

Report File content accurately contains the original file bytes.

Current behavior

Report File content contains a string encoding of the original file bytes, which when encoded back into binary form will be missing data.

Steps to reproduce

1. Download the PDF file via command line (e.g. PowerShell) Invoke-RestMethod -Method Get -Uri 'https://api.checkout.com/reports/REPORT_ID/files/FILE_ID' -Headers @{"Authorization"="SECRET"} -OutFile original.pdf
2. Download the same PDF file using the Checkout SDK and attempt to save in any of the available text encodings

using Checkout;
using System.Text;

ICheckoutApi api = CheckoutSdk.Builder().StaticKeys()
    .SecretKey("SECRET")
    .Environment(Checkout.Environment.Production)
    .Build();

var reportId = "REPORT ID WHERE .Files[0].Format == PDF";

var report = await api
    .ReportsClient()
    .GetReportDetails(reportId);

var file = await api
    .ReportsClient()
    .GetReportFile(reportId, report.Files[0].Id);

File.WriteAllBytes("report.pdf", Encoding.Unicode.GetBytes(file.Body));
File.WriteAllBytes("report2.pdf", Encoding.UTF8.GetBytes(file.Body));
File.WriteAllBytes("report3.pdf", Encoding.ASCII.GetBytes(file.Body));
File.WriteAllBytes("report4.pdf", Encoding.UTF32.GetBytes(file.Body));
File.WriteAllBytes("report5.pdf", Encoding.Latin1.GetBytes(file.Body));
File.WriteAllText("report6.pdf", file.Body, Encoding.ASCII);

3. Observe that none of the PDF files that had their bytes text encoded previously is keeping the original file bytes:
   

Possible solution

• I may be able to implement this bug fix

Additional information

[armando-rodriguez-cko]

Thank you @miguel-ferreira-cko, we will review the information you provided shortly and get back to you to resolve this issue.

[a-ibarra]

@miguel-ferreira-cko We raised this issue time ago, the thing is that this endpoint is a redirect, and previously this endpoint was always returning a string, in worst case we should return only the string of the redirect and merchant should handle the type of the content based on the response.

[miguel-ferreira-cko]

@armando-ibarra-cko this endpoint has always returned 302 Redirect into a pre-signed location within AWS S3 that downloads a file. The HttpClient in .NET follows redirect links by default.

The response type of GetReportFile on the SDK could simply be the HttpClient Response itself

[armando-rodriguez-cko]

Thank you @miguel-ferreira-cko for the information, we will review and we will try to improve this point. We will let you know when we have something.

[armando-rodriguez-cko]

@miguel-ferreira-cko can you please confirm that this behavior is no longer happening?

[armando-rodriguez-cko]

We have closed this issue due to inactivity, please make a new comment to reopen it, thanks.