Re: Devs: please run unit tests, pylint and pyflakes more often

On Thursday, November 16, 2017 at 10:51:25 AM UTC-6, Edward K. Ream wrote: A few notes about reinstalling Anaconda, copied from a recent issue: The following worked to upgrade Anaconda for me, and now the qtconsole works for 2 and 3: conda2 update --all # python 2 conda update --all # python S

Re: Devs: please run unit tests, pylint and pyflakes more often

On Thu, 16 Nov 2017 12:07:51 -0600 "Edward K. Ream" wrote: > ​You will quickly wonder how you ever lived without it. It is, by > far, the best python tool available. I use pyflakes, just didn't recall the Leo setting. It seems mostly to save you from running code that's going to fail, how big a

Re: Devs: please run unit tests, pylint and pyflakes more often

On Thu, Nov 16, 2017 at 11:54 AM, Terry Brown wrote: > It would also be good to run pylint -a before all commits. > > I'll give pylint a try, although my impression in the past is that it > whines more than it helps. The changes I see in Leo source to > accommodate it seem that way to me. Do yo

Re: Devs: please run unit tests, pylint and pyflakes more often

On Thu, 16 Nov 2017 08:51:25 -0800 (PST) "Edward K. Ream" wrote: > I've put a lot of work into making running unit tests faster and more > convenient. Please run them before all commits. The speed up and focus neutrality certainly make running them a lot less onerous. I'm still not sure how m

Re: Back to fixing bugs

On Friday, November 10, 2017 at 3:24:57 AM UTC-6, Edward K. Ream wrote: This filter [is:issue is:open label:First label:Bug] shows that only 4 high-priority bugs remain. I w

Devs: please run unit tests, pylint and pyflakes more often

I've put a lot of work into making running unit tests faster and more convenient. Please run them before all commits. Terry, your concerns about environmental concerns are valid, but until #577 Create a new TestCase subclass for reformatting unit tests

4d01ec92: Fixed a hangnail involving unit tests

The checkin log: QQQ Fixed an irritating hangnail in the unit tests by fixing a bug in LM.computeWorkbookFileName. LM.computeWorkbookFileName now returns None only if: 1. The workbook does not exist. 2. We are unit testing or in batch mode. Previously, the method returned None reg

Re: launching unit tests

On Wed, Nov 15, 2017 at 10:51 AM, Terry Brown wrote: ​​ > for example a test of paragraph rejustification might be easier to > write if you didn't have to account for different user preferences on > text width - that sort of thing. > ​I agree​. Happily, there is a straightforward way. The @tes