Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 321651

Summary:

[compiler][otre] array-translations (lifting/lowering) in nested team not properly translated

Product:

[Tools] Objectteams

Reporter:

Stephan Herrmann <stephan.herrmann>

Component:

OTJ

Assignee:

Stephan Herrmann <stephan.herrmann>

Status:

VERIFIED FIXED

QA Contact:

Severity:

normal

Priority:

Version:

0.7

Target Milestone:

0.7.1

Hardware:

Other

OS:

Linux

Whiteboard:

Attachments:

Description	Flags
fixes	none

Description Stephan Herrmann

2010-08-03 16:19:04 EDT

When a nested team performs lowering/lifting for arrays of its (nested) role,
both compiler and otre currently fail, mainly because lifting/lowering
requires the use of an appropriate team-method, but those methods are searched
for in the enclosing team, rather in the nested team itself.

Although r671 already fixed an easy issue (witnessed by test2215_arrayLifting6)
additional test test2215_arrayLifting7() shows more problems in this area.

Comment 1 Stephan Herrmann

2010-08-03 16:28:15 EDT

Created attachment 175808 [details]
fixes

Various fixes as needed for test2215_arrayLifting7():

* callin-wrapper didn't find lift method because search started at the
  enclosing team
  (rather than expecting role to be a team with lift-methods, too)

* array-lowering of base-call result didn't find a suitable call target,
  because lowering as inserted by InsertTypeAdjustmentsVisitor didn't have
  enough context (recall that callin wrappers are out off its scope).
  * insert an explicit PotentialLowerExpression
  * feed the appropriate team expression into the PotentialLowerExpression
    and from there into ArrayLowering

* array-lowering in a nested team should resolve to the interface method for 
  easier use at the call-site

* OTRE adjustment for base-call surrogates 
  (these seem to be independent of the issue of arrays):
  * select this or this$n
  * use the correct receiver type for the lifting call.

Comment 2 Stephan Herrmann

2010-08-03 16:29:49 EDT

Patch has been committed as r672, r673.

Comment 3 Stephan Herrmann

2010-08-03 17:53:57 EDT

r674 fixes a regression in test1129_nestedCallin2() caused by this patch:
More work was needed to determine the correct receiver for array-lowering
of a callin result.

Comment 4 Stephan Herrmann

2010-09-23 08:56:27 EDT

Verified using I201009211735