All notable changes to the Hello HTTP software will be documented in this file.
The format is based on Keep a Changelog.
- Prettify button in JSON request editor. This includes GraphQL and gRPC.
Changes since 1.5.2
This version introduces the ability of user scripting. It is limited to pre-flight currently, and may be expanded to other areas according to user voices and demands.
Automatic tests are also implemented and added to the development pipeline. Although it is not visible to end users, it helps to discover some bugs or unexpected behaviors under different connection settings. The coverage of tests will be expanded, and Hello HTTP is stepping away from bugs. Lots of stuffs are reimplemented and optimized. New regressions might be introduced due to reimplementation. Please let me know and I will try to fix it within a short period of time.
- Pre-flight user scripting
- Subproject configuration
- Duplicating an Environment -- there is a new Duplicate button in the top-right corner after selecting an environment in the Environment Editor
- Duplicating a Request Example
- Request Examples dropdown list -- useful if you have lots of examples or long example names
- A connection now terminates immediately, even if request is not completely sent, as soon as a response with an error HTTP status (e.g. 4XX, 5XX) is received.
- For request response copying, it now truncates a request body if it is over 2 MB.
- In transport logs, individual payload is truncated if it is over 512 KB. It is configurable per direction in Subproject Configuration.
- In transport logs, the remaining HTTP/2 data payloads are truncated if total HTTP/2 data size is over 2 MB per direction. It is configurable per direction in Subproject Configuration.
- Hello HTTP now only receives HTTP responses up to 22 MB.
- The display formats of events in transport log are slightly updated. Some variables are enclosed with a square bracket.
- Now all artifacts are generated by GitHub-hosted runners.
- Pressing Ctrl-Enter/Cmd-Enter inside a request body textarea did not fire the current request.
- Wrong color was used to highlight empty string literals in JSON.
- In "Body"/"Query"/"Header", when there are lots of inherited values, key-value editors for "This Example" should still appear but it did not.
- There was an unexpected extra line break after a request body in the content of request response copying.
- Out of memory error when a large request body is used.
- An unfrequented error, ConcurrentModificationException, when receiving and persisting multiple payloads.
- Memory leak when more requests are sent.
- Memory leak when a connection is not completed gracefully.
- Documentation bug -- if "Default" is selected for HTTP protocol version, HTTP/1.1 instead of HTTP/2 is used for the "http" protocol.
- Memory usage -- now the memory usage of request body and response body is linear.
- Disk usage -- now large requests and responses are truncated before storing, so it is less likely for a subproject to grow over several hundreds of megabytes.
Changes since 1.6.0-beta.3
- Wrong color was used to highlight empty string literals in JSON
- "Copy All" button was missing for plain HTTP and GraphQL requests
Changes since 1.6.0-beta.2
WARNING: Please make a backup before start using a pre-release version!
- Could not send HTTP request on some systems, primary Windows
Changes since 1.6.0-beta.1
WARNING: Please make a backup before start using a pre-release version!
- Could not send any HTTP request
Changes since 1.5.2
WARNING: Please make a backup before start using a pre-release version!
This version introduces the ability of user scripting. It is limited to pre-flight currently, and may be expanded to other areas according to user voices and demands.
Automatic tests are also implemented and added to the development pipeline. Although it is not visible to end users, it helps to discover some bugs or unexpected behaviors under different connection settings. The coverage of tests will be expanded, and Hello HTTP is stepping away from bugs. Lots of stuffs are reimplemented and optimized. New regressions might be introduced due to reimplementation. Please let me know and I will try to fix it within a short period of time.
- Pre-flight user scripting
- Subproject configuration
- Duplicating an Environment -- there is a new Duplicate button in the top-right corner after selecting an environment in the Environment Editor
- Duplicating a Request Example
- Request Examples dropdown list -- useful if you have lots of examples or long example names
- A connection now terminates immediately, even if request is not completely sent, as soon as a response with an error HTTP status (e.g. 4XX, 5XX) is received.
- For request response copying, it now truncates a request body if it is over 2 MB.
- In transport logs, individual payload is truncated if it is over 512 KB. It is configurable per direction in Subproject Configuration.
- In transport logs, the remaining HTTP/2 data payloads are truncated if total HTTP/2 data size is over 2 MB per direction. It is configurable per direction in Subproject Configuration.
- Hello HTTP now only receives HTTP responses up to 22 MB.
- The display formats of events in transport log are slightly updated. Some variables are enclosed with a square bracket.
- Now all artifacts are generated by GitHub-hosted runners.
- In "Body"/"Query"/"Header", when there are lots of inherited values, key-value editors for "This Example" should still appear but it did not.
- There was an unexpected extra line break after a request body in the content of request response copying.
- Out of memory error when a large request body is used.
- An unfrequented error, ConcurrentModificationException, when receiving and persisting multiple payloads.
- Memory leak when more requests are sent.
- Memory leak when a connection is not completed gracefully.
- Documentation bug -- if "Default" is selected for HTTP protocol version, HTTP/1.1 instead of HTTP/2 is used for the "http" protocol.
- Memory usage -- now the memory usage of request body and response body is linear.
- Disk usage -- now large requests and responses are truncated before storing, so it is less likely for a subproject to grow over several hundreds of megabytes.