grass.temporal.stds_import: Use pathlib Path.read_text and Path.write_text for proj string

[echoix]

Similar changes for stds_export are not in this PR by purpose. During the weekend, I had made a mistake in stds_export somewhere and didn't find it, but the mismatch in lengths of init file vs list file in stds_import, that wasn't touched yet. That's why I insist so much on separating input/output in different PRs.

Details on a failing test saved because stds_import and stds_export weren't changed together

In the spirit of #4224 (comment), this is only fixes the part where a context manager can be avoided.

See PR description of #4234 for an illustration of the implementation of read_text() and write_text()

[wenzeslaus]

Details on a failing test saved because stds_import and stds_export weren't changed together

That's strange. That should not matter. There might be some end of line counting issue, but maps should be counted, not lines. I didn't see that though with just a quick look into the source code.

[echoix]

Details on a failing test saved because stds_import and stds_export weren't changed together

That's strange. That should not matter. There might be some end of line counting issue, but maps should be counted, not lines. I didn't see that though with just a quick look into the source code.

You can look at here before the revert: echoix#202 or a combined 4 PRs ahead + coverage echoix#206

[echoix]

I suspect that the resourcewarnings are captured in a file that is used somewhere, like if a shell script redirected 2 & 1 together to a file and uses it as an input. That's why I think changing both sides isn't a good idea.

But anyways, this PRs doesn't touch the stds_export with old code, it was some fresh work from tonight, on another file.