Replace libdparse with DMD in BodyOnDisabledFuncsCheck #127
Conversation
Vladiwostok
force-pushed
the
body-on-disabled
branch
from
9977a60
to
9e04e59
Compare
| return hasDisabledMemberFunctionAttribute(dec.memberFunctionAttributes); | ||
| } | ||
| bool isCurrentDisabled = isDisabled || (node.storage_class & STC.disable) != 0; | ||
| if (isCurrentDisabled) |
There was a problem hiding this comment.
This is the only use of this variable. You could simply inline the condition.
| if (attr.atAttribute !is null && attr.atAttribute.identifier.text == "disable") | ||
| return true; | ||
| bool isCurrentDisabled = isDisabled || (node.storage_class & STC.disable) != 0; | ||
| if (isCurrentDisabled && node.fbody !is null) |
| ^^ [warn]: Function marked with '@disabled' should not have a body +/ | ||
| ~this() {} /+ | ||
| ^^ [warn]: Destructor marked with '@disabled' should not have a body +/ | ||
| this() {} // [warn]: Constructor marked with '@disabled' should not have a body |
There was a problem hiding this comment.
I'm a bit worried that currently we are simply using the line to output the warning whereas upstream dscanner seems to also have some column information. This indeed makes it easier to follow where the problem is. I know that this is just for internal use, but this still might meet some opposition when trying to integrate to upstream.
There was a problem hiding this comment.
I know, however this isn't the only check where warnings are not verified for column number. I plan on going through all visitors and fixing various diffs that are still present even after the rebase, such as column verification, nolint config, autofix
Vladiwostok
force-pushed
the
body-on-disabled
branch
from
3c0cc1b
to
ca45bb6
Compare
Vladiwostok
force-pushed
the
body-on-disabled
branch
from
ca45bb6
to
6b1826f
Compare
This check warns for disabled functions that have a disabled body. Disabled aggregates are ignored.