Stop after parsing. #9364

chriseth · 2020-07-08T20:09:50Z

Fixes #8739

Todo:

error out if binary was requested in combination with "stop after parsing"
implement this in standard-io

chriseth · 2020-07-08T20:10:04Z

Would be deligted if someone took this over :)

Marenz · 2020-07-21T12:45:19Z

I am taking this over :)

Marenz · 2020-07-22T16:06:40Z

Pushed an update. this is still WIP, but feel free to give feedback on the current state.

chriseth · 2020-07-23T11:25:30Z

libsolidity/ast/Types.cpp

@@ -3494,7 +3497,7 @@ string FunctionType::externalSignature() const
 	}

 	// "inLibrary" is only relevant if this is not an event.
-	bool const inLibrary = kind() != Kind::Event && dynamic_cast<ContractDefinition const&>(*m_declaration->scope()).isLibrary();
+	bool const inLibrary = kind() != Kind::Event && m_declaration->scope() && dynamic_cast<ContractDefinition const&>(*m_declaration->scope()).isLibrary();


I'm wondering whether we should rather assign scopes during parsing - it should be a purely syntactic thing, so the parser should be able to do that easily.

Generally or for this PR?

Just because I'm randomly reading these comments: I actually thought the same when looking at #8834 - the m_free member there could just be derived from the scope, if the parser did this stuff already...

Marenz · 2020-07-23T12:33:33Z

Updated: now missing is the documentation update

Marenz · 2020-07-23T12:36:04Z

libsolidity/interface/CompilerStack.cpp

 	vector<string> contractNames;
-	for (auto const& contract: m_contracts)
-		contractNames.push_back(contract.first);
+


I wonder if I maybe should make this return value based on the imports instead?

I don't understand what you are suggesting here, but do you think it makes sense to just keep the code here and instead initialize m_contracts already right after parsing?

I removed this change completely now. Turns out it was useless.

Marenz · 2020-07-27T10:06:48Z

Documentation is also updated now

axic

Needs to be rebased and changelog moved to 0.7.1.

axic · 2020-08-05T21:51:11Z

libsolidity/ast/ASTJsonConverter.cpp

@@ -731,7 +736,7 @@ bool ASTJsonConverter::visit(FunctionCall const& _node)
 		attributes.emplace_back("isStructConstructorCall", _node.annotation().kind == FunctionCallKind::StructConstructorCall);
 		attributes.emplace_back("type_conversion", _node.annotation().kind == FunctionCallKind::TypeConversion);
 	}
-	else
+	else if (_node.annotation().kind != FunctionCallKind::Unset)


Is this a bugfix?

Well, it fixes a bug that happened as a result from the stop-after-parse change. I think usually you wouldn't get Unset.

solc/CommandLineInterface.cpp

Marenz · 2020-08-13T11:15:42Z

Updated

christianparpart · 2020-08-17T08:35:58Z

Changelog.md

@@ -6,6 +6,7 @@ Language Features:
 Compiler Features:
 * Standard JSON Interface: Do not run EVM bytecode code generation, if only Yul IR or EWasm output is requested.
 * Yul: Report error when using non-string literals for ``datasize()``, ``dataoffset()``, ``linkersymbol()``, ``loadimmutable()``, ``setimmutable()``.
+ * New feature to stop compilation after parsing stage.


Maybe also option the command line option for doing that?

you mean "mention"?

libsolidity/interface/CompilerStack.cpp

christianparpart · 2020-08-17T09:34:39Z

libsolidity/interface/CompilerStack.h

@@ -216,11 +217,11 @@ class CompilerStack: boost::noncopyable

 	/// Parses and analyzes all source units that were added
 	/// @returns false on error.
-	bool parseAndAnalyze();
+	bool parseAndAnalyze(State _stopAfter = State::CompilationSuccessful);


Referring to my above comment, I think this function should not have that parameter, and callers that needs to stop after parse should then use compile(stopAfterStage) or just call parse() directly?

test/stopAfterParseTests.sh

chriseth · 2020-09-02T10:43:48Z

test/libsolidity/ASTJSON/abstract_contract_parseOnly.json

+      "baseContracts": [],
+      "contractDependencies": [],
+      "contractKind": "contract",
+      "fullyImplemented": true,


I think it's totally fine that some members are missing, but I'm worried that wrong information might be very misleading.

chriseth · 2020-09-02T10:46:56Z

The fuzzer somehow interprets a standard-json input as plain solidity.

chriseth · 2020-09-03T10:43:27Z

I think it really makes sense to use std::optional or SetOnce for all annotations. This will also help in other cases.

leonardoalt · 2020-09-24T11:34:27Z

Conflicts

Marenz · 2020-09-29T12:52:57Z

Ready now.

Changelog.md

Marenz · 2020-09-29T14:19:42Z

Update: Moved changelog entry to correct version

chriseth · 2020-09-30T13:07:56Z

Needs another rebase.

Marenz force-pushed the stopafter branch from c1417d3 to c17aa76 Compare July 22, 2020 16:06

chriseth commented Jul 23, 2020

View reviewed changes

Marenz force-pushed the stopafter branch from c17aa76 to 3cef0f2 Compare July 23, 2020 12:33

Marenz reviewed Jul 23, 2020

View reviewed changes

Marenz force-pushed the stopafter branch 3 times, most recently from f9f3e11 to 0a24a29 Compare July 23, 2020 13:01

Marenz self-assigned this Jul 27, 2020

Marenz requested review from christianparpart and leonardoalt July 27, 2020 13:17

axic requested changes Aug 5, 2020

View reviewed changes

axic reviewed Aug 5, 2020

View reviewed changes

solc/CommandLineInterface.cpp Outdated Show resolved Hide resolved

axic reviewed Aug 5, 2020

View reviewed changes

solc/CommandLineInterface.cpp Outdated Show resolved Hide resolved

ekpyron mentioned this pull request Aug 5, 2020

Free functions. #8834

Merged

3 tasks

Marenz force-pushed the stopafter branch from 0a24a29 to 05fefa4 Compare August 13, 2020 11:15

Marenz force-pushed the stopafter branch from 05fefa4 to d4cfd3e Compare August 17, 2020 08:41

christianparpart reviewed Aug 17, 2020

View reviewed changes

Marenz force-pushed the stopafter branch 3 times, most recently from 68add90 to 9c241ac Compare August 17, 2020 14:00

Marenz added the takeover Can be taken over by someone other than label giver label Aug 17, 2020

Marenz force-pushed the stopafter branch from 9c241ac to c3a9d06 Compare August 24, 2020 08:53

Marenz removed the takeover Can be taken over by someone other than label giver label Aug 24, 2020

chriseth force-pushed the stopafter branch 2 times, most recently from 6b488ed to bb3f942 Compare September 2, 2020 10:39

chriseth commented Sep 2, 2020

View reviewed changes

Marenz force-pushed the stopafter branch from bb3f942 to b0faabd Compare September 3, 2020 09:47

Marenz force-pushed the stopafter branch 2 times, most recently from 53424d5 to 73aa7d7 Compare September 3, 2020 16:21

Marenz mentioned this pull request Sep 17, 2020

Make annotations SetOnce or optional where feasible #9839

Merged

Marenz force-pushed the stopafter branch 7 times, most recently from 3c373bd to 96f2605 Compare September 29, 2020 12:08

chriseth commented Sep 29, 2020

View reviewed changes

Changelog.md Outdated Show resolved Hide resolved

Marenz force-pushed the stopafter branch from 96f2605 to 54a7e3e Compare September 29, 2020 14:19

Marenz and others added 2 commits September 30, 2020 16:56

Add name for split-test to prevent failure in other places

4c02cd2

Stop after parsing.

fda8bde

Marenz force-pushed the stopafter branch from 54a7e3e to fda8bde Compare September 30, 2020 14:57

chriseth merged commit 729ffc6 into develop Sep 30, 2020

chriseth deleted the stopafter branch September 30, 2020 16:07

cameel mentioned this pull request Oct 2, 2020

ICE in CompilerStack: contract redefined to inherit from a non-existent type not detected as name clash #9948

Closed

cameel mentioned this pull request Jul 7, 2021

[CLI] Report errors for command line options that have no effect #11629

Closed

cameel mentioned this pull request Jun 14, 2023

Fix incomplete AST in standard json on analysis fail #14328

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Stop after parsing. #9364

Stop after parsing. #9364

chriseth commented Jul 8, 2020 •

edited by Marenz

Loading

chriseth commented Jul 8, 2020

Marenz commented Jul 21, 2020

Marenz commented Jul 22, 2020

chriseth Jul 23, 2020

Marenz Jul 23, 2020

chriseth Jul 23, 2020

ekpyron Aug 5, 2020

Marenz commented Jul 23, 2020

Marenz Jul 23, 2020

chriseth Aug 26, 2020

Marenz Sep 1, 2020

Marenz commented Jul 27, 2020

axic left a comment

axic Aug 5, 2020

Marenz Aug 11, 2020

Marenz commented Aug 13, 2020

christianparpart Aug 17, 2020

Marenz Aug 17, 2020

christianparpart Aug 17, 2020

chriseth Sep 2, 2020

chriseth commented Sep 2, 2020

chriseth commented Sep 3, 2020

leonardoalt commented Sep 24, 2020

Marenz commented Sep 29, 2020

Marenz commented Sep 29, 2020

chriseth commented Sep 30, 2020

Stop after parsing. #9364

Stop after parsing. #9364

Conversation

chriseth commented Jul 8, 2020 • edited by Marenz Loading

chriseth commented Jul 8, 2020

Marenz commented Jul 21, 2020

Marenz commented Jul 22, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Marenz commented Jul 23, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Marenz commented Jul 27, 2020

axic left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Marenz commented Aug 13, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

chriseth commented Sep 2, 2020

chriseth commented Sep 3, 2020

leonardoalt commented Sep 24, 2020

Marenz commented Sep 29, 2020

Marenz commented Sep 29, 2020

chriseth commented Sep 30, 2020

chriseth commented Jul 8, 2020 •

edited by Marenz

Loading