View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||||||
ID | Project | Category | View Status | Date Submitted | Last Update | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0000040 | mercury | Bug | public | 2008-02-01 15:15 | 2010-02-02 16:23 | ||||||||
Reporter | juliensf | ||||||||||||
Assigned To | |||||||||||||
Priority | normal | Severity | minor | Reproducibility | always | ||||||||
Status | confirmed | Resolution | open | ||||||||||
Product Version | |||||||||||||
Target Version | Fixed in Version | ||||||||||||
Summary | 0000040: RTTI looks for the name of a solver value return repn. type name | ||||||||||||
Description | type_name(type_of(X)) where X is a solver type returns the name of the representation type not the solver type itself. This will occurs even if the solver type is imported from another module. The attached test case demonstrates this problem. | ||||||||||||
Tags | No tags attached. | ||||||||||||
Attached Files |
|
Notes | |
juliensf (administrator) 2008-02-01 17:24 Last edited: 2008-02-03 16:45 |
I think this is occurring because the implementation of type_of/1 calls MR_collapse_equivalences and (for some reason) solver types are being treated as equivalences in this respect. Unfortunately, there is nothing in the current RTTI data structures that would allow us to distinguish that a type_info is for a solver type and avoid collapsing the equivalences. (Actually, for solver types and solver type equivalences you would need to collapse equivalences, but only as far back as the last solver type.) In order to fix this I think we will need to add something to type_infos that distinguishes solver types from non-solver types. |