Amend NameAndTypeResolution expected messages

[izgzhen]

Fixes #2613.

[axic]

Thanks! A few tests fail, see here for a list: https://travis-ci.org/ethereum/solidity/jobs/256799732

[axic]

Oh I see the probable reason. @izgzhen these aren't messages displayed to the user during testing, but the messages the compiler emits when compiling those code snippets. Our tests need to ensure we check that the right message was emitted by the compiler.

[izgzhen]

Right ... so I think another half of the work is still missing right?

[izgzhen]

If so, although not quite the "up-for-grab" level, I'll give a try to some of them at least 😆

[chriseth]

@izgzhen so the idea is that you copy-paste the source code into a file, run solc on the file, check that there is exactly one error message and just copy the error message into the test file. If you want to verify, run soltest -- --no-ipc.

[izgzhen]

Fixed, and for some cases, I found that no matter what I filled in, it is always okay (I filled in "x")

[axic]

for some cases, I found that no matter what I filled in, it is always okay (I filled in "x")

Are you sure the message doesn't contain a single letter x in those cases?

It tries a text submatch.

[izgzhen]

Are you sure the message doesn't contain a single letter x in those cases?
It tries a text submatch.

Well, it is interesting .. I'll try it again. Thanks!

[izgzhen]

@axic Fixed, sorry that I was a bit busy these days thus I didn't carefully go through the existing code before diving into it. Thanks a lot for your patience

[axic]

Thanks! Is this still partially solving #2613 or does it change every message?

[izgzhen]

Thanks! Is this still partially solving #2613 or does it change every message?

Every originally "" msg that I can find.

[axic]

Looks good to me.

[axic]

@chriseth please double check, but would be nice to merge soon

[axic]

Actually the real test case here is "Override changes function to modifier.".

[axic]

Actually the real test case here is "Override changes modifier to function."

[axic]

The real test case here is "Only one fallback function is allowed."

[axic]

@izgzhen sorry it took so long, but could you change the above three things? I'd like to get this merged asap.

[izgzhen]

@izgzhen sorry it took so long, but could you change the above three things? I'd like to get this merged asap.

It is okay :) I've updated the code, hoping it works.

[axic]

Thanks!

Seems like we need to improve our testing infrastructure to handle those properly, because it always takes the first error:

Expected message "Override changes function to modifier." but found "Identifier already declared.".
/home/travis/build/ethereum/solidity/test/libsolidity/SolidityNameAndTypeResolution.cpp(1120): error in "modifier_overrides_function": check searchErrorMessage(err, ("Override changes function to modifier.")) failed
Expected message "Override changes function to modifier." but found "Identifier already declared.".
/home/travis/build/ethereum/solidity/test/libsolidity/SolidityNameAndTypeResolution.cpp(1131): error in "function_overrides_modifier": check searchErrorMessage(err, ("Override changes function to modifier.")) failed
Expected message "Only one fallback function is allowed." but found "Function with same name and arguments defined twice.".
/home/travis/build/ethereum/solidity/test/libsolidity/SolidityNameAndTypeResolution.cpp(1369): error in "fallback_function_twice": check searchErrorMessage(err, ("Only one fallback function is allowed.")) failed

Can you just revert those 3 to what you had and I'll address it in a separate PR?

[axic]

@izgzhen realised I can push to your branch and did the changes myself

[axic]

@izgzhen thanks!