You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Follow-up from #1716: it should be possible to externally add libraries with @Owning / @NotOwning in suitable locations.
I will try to blend this into the existing support for .eea files.
eea for null analysis uses this encoding:
0 = @Nullable
1 = @NonNull
Without modifying the general format, we still have the digits 2-9 for encoding more annotations. While there is no intuitive mapping like 0/1, I think using 8/9 is a pragmatically valid option, that still leaves us with the option to use 2 for another null related annotation etc. (Should we ever run out of digits, we can still change the format to let digits start a multi char encoding etc. - but not for now).
Technically we need to check if the existing integration via TypeAnnotationWalker can suitably "simulate" owning annotations in declaration positions. Let's see whether having @Owning specify TYPE_USE as one of its targets works for or against this strategy.
The text was updated successfully, but these errors were encountered:
fixeseclipse-jdt#2258
Implement minimal tests borrowing infra from existing ones
+ add AutoCloseable to jclMin21.jar
Feed eea into tagBits of method, parameter, field
Update analysis to consider assignment to @owning field of other class
Follow-up from #1716: it should be possible to externally add libraries with
@Owning
/@NotOwning
in suitable locations.I will try to blend this into the existing support for .eea files.
eea for null analysis uses this encoding:
0
=@Nullable
1
=@NonNull
Without modifying the general format, we still have the digits 2-9 for encoding more annotations. While there is no intuitive mapping like 0/1, I think using 8/9 is a pragmatically valid option, that still leaves us with the option to use
2
for another null related annotation etc. (Should we ever run out of digits, we can still change the format to let digits start a multi char encoding etc. - but not for now).Technically we need to check if the existing integration via
TypeAnnotationWalker
can suitably "simulate" owning annotations in declaration positions. Let's see whether having@Owning
specifyTYPE_USE
as one of its targets works for or against this strategy.The text was updated successfully, but these errors were encountered: