Solves #1061 and #1656

[lsaffre]

This solves #1061 and #1656, at least for me and as far as i tried.
See also http://lino-framework.org/blog/2015/0103.html
Sorry for an accidental changeset (and a second one to revert it).

[shimizukawa]

Thanks for contribution! However, the change makes errors. Please check the test result.

[birkenfeld]

This is the cause of the test failures: the angle brackets are added compared to the old return value.

I think it's better readable with angle brackets, but:

• test should be adapted
• angle brackets should also be added in the next line for consistency

[lsaffre]

Thanks Georg for finding the reason (i had no idea where to start searching). Yes, brackets might be more readable, but for me it works without them.

[lsaffre]

Now the tests fail only on Python 3 where they say "nose.proxy.NameError: global name 'unicode' is not defined". Any suggestions how to solve this?

[shimizukawa]

This 3 lines change is a cause of test failing on Py3. text_type is a compatibility type for py2 unicode and py3 str. please revert this 3 lines.

[lsaffre]

Thanks, Takayuki, for your hint!

[birkenfeld]

Looks good to me now.

[birkenfeld]

BTW, now with git it's easier to remove unwanted changesets. Just to a git rebase -i, remove the offending commits, and then push with -f. Github will automatically update the pull request.

[lsaffre]

Thanks, Georg, for your hint about |git rebase|. I tried it but without success... must say that this would have been my first rebase... details in my blog http://lino-framework.org/blog/2015/0105.html

[birkenfeld]

Well, looking at your blog entry, there are two snags:

• git needs EDITOR to return only when the edit is completed. I don't think this can be changed, but I assume that it's possible to set a different editor executable.
• the argument to the rebase is the head of the commits that should not be touched. Commits between that and HEAD will be offered for rebase. When you do rebase -i master on the master branch, it's a noop since there are no commits between master and master :) Usually the argument will be origin/master (rebase/edit commits made relative to upstream repository's branch), or a different branch (rebase commits onto the specified branch).

[birkenfeld]

Merged manually, thanks!

[birkenfeld]

See 21b8384

Note that it retained your author info, another nice bit in git.

[lsaffre]

Nice, I am part of Sphinx's history :-) Thanks, Georg and Takayuki, for
your help and sorry for my slowness (so many other urgent projects).

Luc

On 06/01/15 19:19, Georg Brandl wrote:

See 21b8384
21b8384

Note that it retained your author info, another nice bit in git.

—
Reply to this email directly or view it on GitHub
#1663 (comment).