Holger Vogt

I have pushed some update to the state code mode (branch pre-master-47): Improve the error messages upon syntax errors, and bail out when reading the state file has not been successful.

Weird problem with meas including an index it shouldn't

Please send your file per email. It should be runnable out-of-the box (include all includes) and show the bug.

Next point is: 2 0s 1s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0 1 -> 4 What is 0s 0 1 ->? The first header line differs 0s 0 -> You have one vector element of the input defined [Shift_in], so0 1is wrong.

Please note (from the manual): A line that is not one of the above will cause a file-loading error. Note that in the example shown, whitespace (any combination of blanks, tabs, commas) is used to separate values, and that the character -> is used to underline the state transition implied by the input preceding it. 0 0s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0 ->0 1 ->1 1 1s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0 ->2 1 ->3 2 0s 1s 0s 0s 0s 0s 0s 0s 0s 0s 0s 0 1 ->4 1 ->5 should become 0 0s 0s 0s 0s 0s 0s 0s 0s 0s...

Justin, If case sensitivity is required, command 'setcs' may replace command 'set'. The pre_ prefix to command 'osdi' has a special meaning, see manual chapter 13.5.56. What has been the problem, causing the extra code of the commit? https://sourceforge.net/p/ngspice/ngspice/ci/02aa3418c7be55d116cda11a5d08642ae7fc46aa/

The self-heating code in the diode also uses integration ... and has the same problem. Should we exclude this code (dioload.c, lines 696 ff) as well? The general reason for this dependency is: When running a transient simulation, in dctran.c, line 682, NIcomCof(ckt) is called. This function sets ckt->CKTag[0] = 1/ckt->CKTdelta; ckt->CKTag[1] = -1/ckt->CKTdelta; When starting with dc, both CKTag are 0, when dc is applied after tran, both have a delta given by the final delta of the tran sim. The numbers...

The self-heating code in the diode also uses integration ... and has the same problem. Should we exclude this code (dioload.c, lines 696 ff) as well? The general reason for this dependency is: When running a transient simulation, in dctran.c, line 682, NIcomCof(ckt) is called. This function sets ckt->CKTag[0] = 1/ckt->CKTdelta; ckt->CKTag[1] = -1/ckt->CKTdelta; When starting with dc, both are 0, when dc is applied after tran, both have the numbers given above. The numbers are never changed anymore,...