Added "References" header field to cover more potential use cases and enhance versatility. This field will allow applications to track series of related messages.
Test for "References" field has also been added for validation.
Also included are string methods for Content-type objects with relevant tests, ensuring accurate string conversion. Unnecessary duplicate method of string conversion for Charset has been removed to streamline the code and improve readability.
We can no parse simple mails (multipart is not working yet). The existing implementation was made more efficient by refactoring the EML file parsing and header extraction mechanism. Added 'strings' and 'bytes' packages to facilitate these changes. Previously, headers and body were parsed separately which was unnecessarily complex and increased the chance of errors. Now, with the new function 'readEML' and the helper function 'parseEMLBodyParts', we are able to parse headers and body together which not only simplifies the code but also increases its reliability. Specifically, 'bytes.Buffer' now helps us capture body while parsing, which removes need for separate handling. Additionally, certain headers like 'charset' and body types are also accounted for in the new implementation, enhancing the completeness of information extracted from EML files.
The diff modifies how the email library handles the extraction of the mime media type from an email header. It uses the mime.ParseMediaType function to parse the content type header. The function gives back the media type as a string and a mapping of different associated parameters. This mapping was previously just printed, but now the charset parameter is also used for setting the charset of the email if it exists.
Added two new functions `EMLToMsg` and `readEML` to the `mail` package. `EMLToMsg` function opens and parses a .eml file and returns a pre-filled Msg pointer. `readEML` opens an EML file and uses net/mail to parse the header and body. These changes are made to provide support for EML file parsing, which is a common requirement in many email-based applications.
This commit introduces a new type 'JSONlog' that satisfies the Logger interface for structured JSON logging. This includes new methods 'Debugf', 'Infof', 'Warnf' and 'Errorf' to log messages at different levels and an associated test 'jsonlog_test.go' to ensure correct functionality. This enhances the logging functionality by providing clarity in logs and eases debugging process.
Moved Level constants and directionPrefix() function from log/stdlog.go to log/log.go. This change improves the organization of the code by grouping log level constants and related functions in the same file. Furthermore, this removes redundancy and ensures consistency of log level across different files.
The deleted .idea XML configuration files (markdown.xml, modules.xml, Project_Default.xml) were not required for the functioning of the project. Removing these files helps in reducing clutter and simplifying the project structure. Also, these files can often contain user-specific settings or paths, and may cause conflicts when working in a team.
This commit updates the Go version from 1.20 to 1.21 on three GitHub workflow files: golangci-lint.yml, codecov.yml, and sonarqube.yml. The change ensures we use the latest Go version which is more efficient and comes with additional features, leading to improved project performance.
The .idea/vcs.xml file was deleted because it is not needed for the project. It was initially created by the IntelliJ IDEA but we're managing version control configurations through another way, making this file redundant.
Deleted the `.idea/go-mail.iml` file which was committed in error and should not be part of the source code repository. Updated the `.gitignore` file to ignore the whole `.idea` directory, since it contains configuration files that are specific to each developers environment and not required for the project build. The changes in the repository will now make it more clutter-free and prevent unintentional commits of unnecessary files in the future.
In the updated version of msgwriter.go, an additional error handling process has been included. If an error is detected when creating a new part in the message writer, this error is stored and prevents executing the writeBody function. This fixes nil pointer dereference in `mw.writeBody` if an error occured previously.
In the updated version of msgwriter.go, an additional error handling process has been included. If an error is detected when creating a new part in the message writer, this error is stored and prevents executing the writeBody function. This fixes nil pointer dereference in `mw.writeBody` if an error occured previously
The logging system in the smtp.go and log package has been refactored. A new custom log type `Log` was introduced, that includes the message direction, format and arguments. The `Logger` interface and the `Stdlog` implementation were modified to accept this new type. This change provides a clearer understanding of message direction inside logs, allowing for easier debugging and reduced complexity. This change does not affect features or disrupt user functionality. Additionally, it allows for custom implementations of the log.Logger interface to without being forced to use the C --> S/C <-- S direction logging.
A `const CallDepth` placeholder has been introduced and has been set to 2. This is used as an argument in Output method calls in various log levels like debug, info, warn, and error in place of previously hardcoded '2'. This change improves the readability of the code. The '2' represented the call depth of log.Logger's Output method, which wasn't immediately clear with the hardcoded value.
- Bump version in doc.go
- Add sonarlint to .gitignore
- Update README to reflect the changes and contributors
- Fixed typos in auth_cram_md5_118.go, msg.go, msg_test.go, reader.go and smtp_ehlo_117.go
Fixes#121. `defer` in for loops can possibly lead to leaks.
The whole handling including the defer are now moved into anonymous functions outside the for loop and are called form the loop instead. This way the defer is handle when the inline function finishes
Since we are Go <1.20 compatible, we can't easily combine errors here. The suggested golangci-lint fix will not work neither, since the %w verb can't be used twice. Therefore, for now we will ignore the linting error
Closes#119
