TRZ reader and writer are compatible with python 3 #690
Conversation
Work around the merge of int and long int in python 3. Note that there is no test for that work around as it would require a heavy TRZ file.
The length of the title in a TRZ file is limited to 80 characters.
TRZ titles shorter than 80 characters are read with trailing spaces. These spaces are striped as they are not expected by the users.
| seeksize = framesize * nframes | ||
| # On python 2, seek has issues with long int. This is solve in python 3 | ||
| # where there is no longer a distinction between int and long int. | ||
| if six.PY2: |
There was a problem hiding this comment.
This is not tested as it would require a very large file. Any idea to test that without such a large file is welcome. If somebody as a large file to test this at least once, it may be worth to try.
There was a problem hiding this comment.
I'll see if I have one around for this..
There was a problem hiding this comment.
And as for without a large file, I think technically we could mock up a fake file that returns byte offsets as if it were a large file or something... And then raised an Error if it was given a long int or something? I'll just check this manually I think :D
There was a problem hiding this comment.
I guess a Mock could do. But it would have to fail where an actual large file would fail, and it would have to capture how the behaviour differs among different versions of python. This would be tricky.
|
Thanks for doing this, all looks good except comments |
Following @richardjgowers comment on MDAnalysis#690
TRZ reader and writer are compatible with python 3
TRZ coordinate reader and writer are now compatible with python 3. A tests are added to make sure the title is read and write as expected. Also, the reader strips tailing spaces from short titles. Finally, the error massage raised by the writer when the title is too long is now friendlier.
Fix #689
See #260