The first two commits are related to the issue with a locked database that has been discussed on the mailing list. Even though it is now possible use postgres instad of sqlite as the database backend, this simple workaround may be useful for some. It simply keeps trying to save the record in case there is a conflict with another simulation run (until a maximum number of attempts is reached). This doesn't always work, but it's fixed the problem for me in 95% of the cases.
The other two commits are minor changes which others may (or may not) find useful. Regarding the second of those: personally, I often run simulations which produce lots of output files, but I'm usually mostly interestd in the parameter values for the runs. However, if many output files are present it's tedious having to scroll down to the 'parameters' section in the web interface. Therefore the proposed change to display the paramters before the input/output files. Others' mileage may vary, though, and I'm happy to remove this commit (or one of the others) from the PR if it doesn't seem to be generally useful.
Thanks for these fixes. I'm not convinced about re-ordering the parameters and data files - for my work, I am more likely to have a few dozen parameters and only one or two data files - however, I will accept the pull request and see what feedback we get.
Thanks for merging and the comment! As I said, feel free to revert any of these commits if they don't seem to be of general usefulness. I agree that most people might not have that many output files, but I figured that the parameter values are still of more interest than the data files (which I presume would usually be analysed externally, although I might be wrong). That's why I thought it might help to display the parameters first. But again, I'm happy to just apply that change in my own repository. Or maybe that's even easily configurable somewhere?