Open chlookaburra opened 7 months ago
While @chlookaburra and I were looking at the code, we noticed two occasions where we're using a fixed-size buffer: https://github.com/SimVascular/svZeroDSolver/blob/7c722a51d58fa2784418b0cc012d103024d27344/src/solve/csv_writer.cpp#L50-L52 https://github.com/SimVascular/svZeroDSolver/blob/7c722a51d58fa2784418b0cc012d103024d27344/src/solve/csv_writer.cpp#L182-L184 I could see a full buffer causing this problem (I don't know if those sizes are hard-coded for performance). As a first step, try increasing them and see if that solves the problem. If yes, we could either dynamically allocate the size (if that doesn't reduce performance) or throw an error if things would be written incorrectly.
Output should use iostream
with iomanip
not fixed-size char arrays.
Description
If the string name of a blood vessel is too long, then it will append to the pressure_out column, instead of at the start of a new row.
See the attached file, where the first row has the usual time, flow_in, flow_out, pressure_in, pressure_out columns. However, the second and third rows have the name of the next blood vessel in the pressure_out column, which can cut off the value of pressure_out (as in the third row).
Reproduction
One could try naming a blood vessel "RightCommonCarotidArtery," which will reproduce this result.
Expected behavior
We would like svZeroDSolver to throw an error if we enter a blood vessel name that is too long or accept long strings.
Additional context
No response
Code of Conduct