Fix compilation error and invalid chunk bug for large simulation data #19
Conversation
When simulation data is large enough to need multiple windows, zero padding is sometimes added between consecutive windows. This fix checks for zero padding between windows and skips ahead if zero padding is detected
- Fix table of contents extraction bug for non-transient simulations - Reconfigure build to use cmake - solves issues with libpsf.la builds - Automate tests with ctest - Use static libpsf.la for python bindings to prevent rpath issue and make build wheels portable
|
@lekez2005 thank you for your great work. I've been using your fork, and it works perfectly, except sometimes the code fails to parse some large transient files (~ 3GB). Do you know what the problem could be? |
|
@yahia3200 Feel free to upload a sample. I am currently unable to work on this due to time constraints but I can take a look the next time I have an opening. Hopefully in the next week or two. |
|
@lekez2005 here is the sample https://drive.google.com/drive/folders/1K9Qk6P8ljT4pTefmIi6TT1Mdi1otls-w?usp=sharing |
|
@yahia3200 I was able to determine the root cause of the crash to be that the file was large enough to overflow int32. I have pushed a fix to the https://github.com/lekez2005/libpsf/tree/uint_32_bugfix branch. This "fix" changes int to unsigned int for some size and offset variables. It successfully parses the data but I don't actually know if the result is correct. Can you test to see if the data is correctly processed? If it is, I'll merge the changes to the main branch. |
No description provided.