ID0000924: **This issue was created automatically from Mantis Issue 924. Further discussion may take place here.** --- | **Id** | **Project** | **Category** | **View** | **Due Date** | **Updated** | | --- | --- | --- | --- | --- | --- | | ID0000924 | Frama-C | Plug-in > Eva | public | 2011-08-12 | 2011-08-13 | | | | | | | | | --- | --- | --- | --- | --- | --- | | **Reporter** | Jochen | **Assigned To** | pascal | **Resolution** | no change required | | **Priority** | normal | **Severity** | text | **Reproducibility** | always | | **Platform** | - | **OS** | - | **OS Version** | - | | **Product Version** | Frama-C Carbon-20110201 | **Target Version** | - | **Fixed in Version** | - | ### Description : Analyzing the attached program with "frama-c -val ftest.c -main f -no-unicode -slevel 1000" reports "i IN {0; 1; 2; }", which seems a too small set. When "n" is replaced by e.g. "9" in line 1 (col.19 only) and 4, the computed set "i IN {9; }" seems ok. When instead "a[i]" is replaced by "i" in line 5, the computed set "i IN [0..2147483647]" seems ok, too. (Only in this case, messages like "[value] Semantic level unrolling superposing up to ..." are printed - is this ok?) If arrays with non-constants upper-bound expressions are not supported, a corresponding warning should be emitted for line 1. ## Attachments - [ftest.c](/uploads/0c15bbc66390644d25e8124e8714f21a/ftest.c)