Anonymous | Login | Signup for a new account | 2024-03-29 07:36 UTC |
Main | My View | View Issues | Change Log | Roadmap |
View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||
ID | Project | Category | View Status | Date Submitted | Last Update | ||||
0000503 | Taste | [All Projects] Ocarina - Ada code generation | public | 2016-05-23 14:26 | 2021-05-21 08:25 | ||||
Reporter | maxime | ||||||||
Assigned To | hugues | ||||||||
Priority | normal | Severity | minor | Reproducibility | have not tried | ||||
Status | closed | Resolution | fixed | ||||||
Platform | OS | OS Version | |||||||
Summary | 0000503: Ocarina does not detect issue in AADL when targetting Ada as it does when targetting C | ||||||||
Description | check the attachement There are two directories : Ada_Bug and C_OK In both of them run:
In C an error is reported because the interface "pulse" is not connected to anything (and it is not cyclic) In Ada this error is not reported and the system can be built... | ||||||||
Tags | No tags attached. | ||||||||
Attached Files | ocarina_bug.tgz [^] (25,855 bytes) 2016-05-23 14:26 | ||||||||
Notes | |
(0002647) hugues (administrator) 2016-05-23 14:38 |
Thanks for the small reproducer, I could reproduce it. Patch is quite trivial and is being tested Note that the model is not wrong by construction, it is "just" inconsistent or incomplete. Hence the question: should this be rejected by the VT ? or by Ocarina ? In the Ocarina case, we may accept it, even though we create unreachable code. PolyORB-HI/C backend is overreacting to this situation, it may simply report a warning and not an error |
(0002648) maxime (administrator) 2016-05-23 15:05 |
Good point. The problem is that if it is just a warning, it goes undetected during the build (lost in the noise), and if it comes from a user mistake it will be extremely difficult to spot (both to detect and to fix). So I think it is better to keep it as an error in Ocarina, but improve the error message as it is now almost impossible to relate to the user mistake. Saying something like "Thread input port is not cyclic and not connected to any other thread" On my side I can add the following behaviour to buildsupport: (1) in debug mode, change the unconnected PIs from Sporadic to Unprotected, so that they will not appear in the concurrency view, thus not raise the error (to let users work on an incomplete system) (2) in release mode, don't change -> the error from ocarina will raise |
(0003753) maxime (administrator) 2021-05-21 08:25 |
With kazoo, unconnected PIs are allowed and ignored (not populated to the Concurrency view) Therefore closing the ticket, which becomes obsolete |
Issue History | |||
Date Modified | Username | Field | Change |
2016-05-23 14:26 | maxime | New Issue | |
2016-05-23 14:26 | maxime | Status |
new => assigned |
2016-05-23 14:26 | maxime | Assigned To |
=> hugues |
2016-05-23 14:26 | maxime | File Added: ocarina_bug.tgz | |
2016-05-23 14:38 | hugues | Note Added: 0002647 | |
2016-05-23 15:05 | maxime | Note Added: 0002648 | |
2017-09-29 08:15 | maxime | Relationship added |
related to 0000668 |
2021-05-21 08:25 | maxime | Note Added: 0003753 | |
2021-05-21 08:25 | maxime | Status |
assigned => closed |
2021-05-21 08:25 | maxime | Resolution |
open => fixed |
Copyright © 2000 - 2011 MantisBT Group |