-
Notifications
You must be signed in to change notification settings - Fork 374
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix: print declared type in output #1143
Merged
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov ReportPatch and project coverage have no change.
Additional details and impacted files@@ Coverage Diff @@
## master #1143 +/- ##
=======================================
Coverage 46.97% 46.98%
=======================================
Files 365 365
Lines 61156 61159 +3
=======================================
+ Hits 28729 28735 +6
+ Misses 30066 30065 -1
+ Partials 2361 2359 -2
Flags with carried forward coverage won't be shown. Click here to find out more.
☔ View full report in Codecov by Sentry. |
piux2
changed the title
print declared type in output
fix: print declared type in output
Sep 18, 2023
thehowl
approved these changes
Sep 18, 2023
moul
approved these changes
Sep 18, 2023
7 tasks
gfanton
pushed a commit
to gfanton/gno
that referenced
this pull request
Nov 9, 2023
This the first PR required for this feature gnolang#1141 --------- Co-authored-by: piux2 <> Co-authored-by: Manfred Touron <94029+moul@users.noreply.github.com>
Merged
thehowl
added a commit
that referenced
this pull request
May 31, 2024
This is part 2 of 3 of the solution for issue #1141. The part 1 of 3 of the solution can be found in issue #1143. In this part of the solution, we have made several improvements: - Support both named and unnamed type assignments in assignment statements and function return values. - Resolved the issue related to incorrect method selectors that is caused by mixing named and unnamed assignments. - Added 62 file tests to ensure the correctness of the code. - Included 2 realm tests to further validate the cross realm assignment and method selector. - Enhanced the support for GNO file tests in nested directories. This allows us to organize tests in intuitively named folders. To achieve the above improvements in the preprocessing phase, we made the following changes: - Introduced an isNamed() function on the Type Interface and marked named types with isNamed() returning true. This helps distinguish between named and unnamed types. - Followed the specifications to convert the right-hand side type into a constant function type. - As for determining the package associated with a test file, we've maintained the original convention. We keeps relying on the comment directive "//PKGPATH: gno.land/r/xyz" in the test file itself to identify the package it belongs to. We do not using the local folder structure to derive the package for file tests. Therefore the tests in tests/files folder can be stored in any intuitively named sub directories. **NOTE:** The named and unnamed type conversions that involve the decomposition of function calls returning multiple values in the preprocess have not yet been included in this pull request. This functionality will be addressed in part 3 of 3 of the entire solution. <!-- please provide a detailed description of the changes made in this pull request. --> <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [ ] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [ ] Updated the official documentation or not needed - [ ] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs - [ ] Provided any useful hints for running manual tests - [ ] Added new benchmarks to [generated graphs](https://gnoland.github.io/benchmarks), if any. More info [here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md). </details> --------- Co-authored-by: deelawn <dboltz03@gmail.com> Co-authored-by: Morgan Bazalgette <morgan@morganbaz.com>
DIGIX666
pushed a commit
to DIGIX666/gno
that referenced
this pull request
Jun 2, 2024
This is part 2 of 3 of the solution for issue gnolang#1141. The part 1 of 3 of the solution can be found in issue gnolang#1143. In this part of the solution, we have made several improvements: - Support both named and unnamed type assignments in assignment statements and function return values. - Resolved the issue related to incorrect method selectors that is caused by mixing named and unnamed assignments. - Added 62 file tests to ensure the correctness of the code. - Included 2 realm tests to further validate the cross realm assignment and method selector. - Enhanced the support for GNO file tests in nested directories. This allows us to organize tests in intuitively named folders. To achieve the above improvements in the preprocessing phase, we made the following changes: - Introduced an isNamed() function on the Type Interface and marked named types with isNamed() returning true. This helps distinguish between named and unnamed types. - Followed the specifications to convert the right-hand side type into a constant function type. - As for determining the package associated with a test file, we've maintained the original convention. We keeps relying on the comment directive "//PKGPATH: gno.land/r/xyz" in the test file itself to identify the package it belongs to. We do not using the local folder structure to derive the package for file tests. Therefore the tests in tests/files folder can be stored in any intuitively named sub directories. **NOTE:** The named and unnamed type conversions that involve the decomposition of function calls returning multiple values in the preprocess have not yet been included in this pull request. This functionality will be addressed in part 3 of 3 of the entire solution. <!-- please provide a detailed description of the changes made in this pull request. --> <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [ ] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [ ] Updated the official documentation or not needed - [ ] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs - [ ] Provided any useful hints for running manual tests - [ ] Added new benchmarks to [generated graphs](https://gnoland.github.io/benchmarks), if any. More info [here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md). </details> --------- Co-authored-by: deelawn <dboltz03@gmail.com> Co-authored-by: Morgan Bazalgette <morgan@morganbaz.com>
omarsy
pushed a commit
to TERITORI/gno
that referenced
this pull request
Jun 3, 2024
This is part 2 of 3 of the solution for issue gnolang#1141. The part 1 of 3 of the solution can be found in issue gnolang#1143. In this part of the solution, we have made several improvements: - Support both named and unnamed type assignments in assignment statements and function return values. - Resolved the issue related to incorrect method selectors that is caused by mixing named and unnamed assignments. - Added 62 file tests to ensure the correctness of the code. - Included 2 realm tests to further validate the cross realm assignment and method selector. - Enhanced the support for GNO file tests in nested directories. This allows us to organize tests in intuitively named folders. To achieve the above improvements in the preprocessing phase, we made the following changes: - Introduced an isNamed() function on the Type Interface and marked named types with isNamed() returning true. This helps distinguish between named and unnamed types. - Followed the specifications to convert the right-hand side type into a constant function type. - As for determining the package associated with a test file, we've maintained the original convention. We keeps relying on the comment directive "//PKGPATH: gno.land/r/xyz" in the test file itself to identify the package it belongs to. We do not using the local folder structure to derive the package for file tests. Therefore the tests in tests/files folder can be stored in any intuitively named sub directories. **NOTE:** The named and unnamed type conversions that involve the decomposition of function calls returning multiple values in the preprocess have not yet been included in this pull request. This functionality will be addressed in part 3 of 3 of the entire solution. <!-- please provide a detailed description of the changes made in this pull request. --> <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [ ] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [ ] Updated the official documentation or not needed - [ ] No breaking changes were made, or a `BREAKING CHANGE: xxx` message was included in the description - [x] Added references to related issues and PRs - [ ] Provided any useful hints for running manual tests - [ ] Added new benchmarks to [generated graphs](https://gnoland.github.io/benchmarks), if any. More info [here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md). </details> --------- Co-authored-by: deelawn <dboltz03@gmail.com> Co-authored-by: Morgan Bazalgette <morgan@morganbaz.com>
7 tasks
piux2
added a commit
that referenced
this pull request
Jul 19, 2024
This is the last part of the solution for issue #1141. The 1 of 3 of the solution can be found in PR #1143. The 2 of 3 of the solution can be found in PR #1246 It decomposes function calls that return multiple values in the preprocess. ### Here is the problem to solve: ` u1, n2 = x() ` How do we ensure that the returned multiple values from a function call adhere to named and unnamed type assignment specifications? Additionally, we want to solve this problem during preprocessing instead of at runtime to minimize the impact on runtime performance. ### The main ideas: u1, n2 = x() << decompose the statement to the following two lines // .tmp_1, .tmp_2 := x() // u1, n2 = .tmp_1, .tmp_2 then we can apply name and unname type conversion specs to the second line. u1, n2 = _tmp_1, _tmp_2 ### Here are the example code and the explanation ``` // decompose_filetest.gno package main type nat []int func x() (nat, []int) { a := nat{1} b := []int{2} return a, b } func main() { var u1 []int var n2 nat u1, n2 = x() // .tmp_1, .tmp_2 := x() // u1, n2 = .tmp_1, .tmp_2 println(u1) println(n2) } // Output: // slice[(1 int)] // (slice[(2 int)] main.nat) ``` ### Here is the simplified recursive tree of the transformation in the preprocess <img width="1336" alt="image" src="https://github.com/gnolang/gno/assets/90544084/306a4770-457d-4131-a82a-2de5c6b0dadf"> ### Here are the major steps involved in this decomposition during preprocessing: - Create hidden temporary name expressions .tmp1, .tmp2. In Go, a leading dot is not valid in variable names, ensuring that users cannot create names that clash with these hidden variables. - Create two statements in the block: one for defining and one for assigning. ``` .tmp1, .tmp2 := x() u1, n2 = .tmp_1, .tmp_2 ``` - Preprocess each newly created statements - Replace the original statement with the two newly created statements. ### Here are two additional changes to facilitate above. - Update the FuncValue's body in `updates := pn.PrepareNewValues(pv) `since its source Body has been changed during preprocessing. - Replace all ` for index := range Body` with `for i:=0; i < len(Body); i++` in transcribe.go since the body length might change due to the decomposition. <!-- please provide a detailed description of the changes made in this pull request. --> <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [x] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [ ] Updated the official documentation or not needed - [x] No breaking changes were made - [x] Added references to related issues and PRs - [ ] Provided any useful hints for running manual tests - [ ] Added new benchmarks to [generated graphs](https://gnoland.github.io/benchmarks), if any. More info [here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md). </details> --------- Co-authored-by: Miloš Živković <milos.zivkovic@tendermint.com> Co-authored-by: Morgan <git@howl.moe>
gfanton
pushed a commit
to gfanton/gno
that referenced
this pull request
Jul 23, 2024
This is the last part of the solution for issue gnolang#1141. The 1 of 3 of the solution can be found in PR gnolang#1143. The 2 of 3 of the solution can be found in PR gnolang#1246 It decomposes function calls that return multiple values in the preprocess. ### Here is the problem to solve: ` u1, n2 = x() ` How do we ensure that the returned multiple values from a function call adhere to named and unnamed type assignment specifications? Additionally, we want to solve this problem during preprocessing instead of at runtime to minimize the impact on runtime performance. ### The main ideas: u1, n2 = x() << decompose the statement to the following two lines // .tmp_1, .tmp_2 := x() // u1, n2 = .tmp_1, .tmp_2 then we can apply name and unname type conversion specs to the second line. u1, n2 = _tmp_1, _tmp_2 ### Here are the example code and the explanation ``` // decompose_filetest.gno package main type nat []int func x() (nat, []int) { a := nat{1} b := []int{2} return a, b } func main() { var u1 []int var n2 nat u1, n2 = x() // .tmp_1, .tmp_2 := x() // u1, n2 = .tmp_1, .tmp_2 println(u1) println(n2) } // Output: // slice[(1 int)] // (slice[(2 int)] main.nat) ``` ### Here is the simplified recursive tree of the transformation in the preprocess <img width="1336" alt="image" src="https://github.com/gnolang/gno/assets/90544084/306a4770-457d-4131-a82a-2de5c6b0dadf"> ### Here are the major steps involved in this decomposition during preprocessing: - Create hidden temporary name expressions .tmp1, .tmp2. In Go, a leading dot is not valid in variable names, ensuring that users cannot create names that clash with these hidden variables. - Create two statements in the block: one for defining and one for assigning. ``` .tmp1, .tmp2 := x() u1, n2 = .tmp_1, .tmp_2 ``` - Preprocess each newly created statements - Replace the original statement with the two newly created statements. ### Here are two additional changes to facilitate above. - Update the FuncValue's body in `updates := pn.PrepareNewValues(pv) `since its source Body has been changed during preprocessing. - Replace all ` for index := range Body` with `for i:=0; i < len(Body); i++` in transcribe.go since the body length might change due to the decomposition. <!-- please provide a detailed description of the changes made in this pull request. --> <details><summary>Contributors' checklist...</summary> - [x] Added new tests, or not needed, or not feasible - [x] Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory - [ ] Updated the official documentation or not needed - [x] No breaking changes were made - [x] Added references to related issues and PRs - [ ] Provided any useful hints for running manual tests - [ ] Added new benchmarks to [generated graphs](https://gnoland.github.io/benchmarks), if any. More info [here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md). </details> --------- Co-authored-by: Miloš Živković <milos.zivkovic@tendermint.com> Co-authored-by: Morgan <git@howl.moe>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This the first PR required for this feature #1141