jeo.method includes too many unnecessary empty elements

[yegor256]

This is what I see in XMIR:

<o base="jeo.method" name="j$foo-%28%29%5BB">
  ...         
  <o base="jeo.params"/>
  <o base="jeo.seq.of0" name="annotations-35144683"/>
  <o base="jeo.seq.of11" name="body-1337806054"> ... </o>
  <o base="jeo.seq.of0" name="trycatchblocks-foo-750353843"/>
  <o base="jeo.seq.of0" name="local-variable-table-704374111"/>
</o>

These tailing parts, related to annotations, trycatchblocks and localvars are seen in every method declaration, even when they are empty. How about creating a set of atoms instead:

jeo.method
jeo.method-with-try-catch
jeo.method-with-annotations
jeo.method-with-try-catch-and-annotations
jeo.method-with-everything

Usually (very often), the first one will be used, which is the most compact. Thus, the size of phi expressions will be reduced.

[yegor256]

@volodya-lombrozo please, check

[volodya-lombrozo]

@yegor256 I just can try to omit empty blocks at all

[yegor256]

@volodya-lombrozo keep in mind that we can't have jeo.method with three arguments in one place and then jeo.method with five in another place.

[volodya-lombrozo]

@yegor256 Why not?

[yegor256]

@volodya-lombrozo you will get a warning: objectionary/lints#259 The reason behind this is that there is no reflection in EO: an atom (jeo.method) can't, in runtime, understand what arguments have been passed to it. We do have partial application in EO, but again, it's not applicable in your case -- it's only used for object step-by-step construction, but not for for its "deconstruction" in runtime.