Fix unintended lstrip_blocks behavior

[petee-d]

Introduced in #858. Closes #1138

[petee-d]

I ran a comparison on the resulting parse tree on 5GBs worth of over 100k Jinja templates created by users with various technical backgrounds and saw no difference. So apparently none of these relied on the behavior that this fixes, but it also shows it's unlikely this produces another undesired change of behavior.

I also performance tested this and there is no measurable change in performance. Also including the performance before the first fix that introduced this change.

X axis is log-rounded seconds, Y axis is the count of templates in the bucket.

[petee-d]

Today I got back to this for a bit and realized that the way I was testing it on my 100k sample templates was wrong, I didn't actually enable lstrip_blocks in my environment. So I repeated the test and the result is great. The generated parse trees after #857 were actually different in many cases from the parse trees generated before #857, and the parse trees after #1183 were in all cases the same as the ones before #857. I made the same mistake when comparing the parse trees when implementing #857 so this shows how the change of behavior went unnoticed and it also confirms that this fix returns to the correct behavior while keeping the much better performance.

[davidism]

Awesome work, thanks for following up. I've actually got this checked out right now, rebasing to 2.11.x and adding a changelog. Should be released later today.

[davidism]

I'm not sure how it happened, but you opened a PR from a different user (xponea), so I can't force push to the branch. If there's an option to allow pushes on the PR, please enable it. Otherwise, can you do the following:

git rebase --onto origin/2.11.x origin/master
nano CHANGES.rst  # add changelog entry to bottom of 2.11.2 section, given below
git commit -a --amend --no-edit
git push -f

Changelog entry:

-   Fix whitespace being removed before tags in the middle of lines when
    ``lstrip_blocks`` is enabled. :issue:`1138`

[davidism]

@petee-d ping on getting this rebased

[petee-d]

@davidism Sorry for the delay, I was mostly offline the past few days. Pushed the rebase and change log entry to my fork branch. I don't see the option to let you make edits to this PR created from a fork, as described in https://help.github.com/en/github/collaborating-with-issues-and-pull-requests/allowing-changes-to-a-pull-request-branch-created-from-a-fork. If you need to do more edits, feel free to copy the branch from my fork and recreate the PR.

[davidism]

Looks good, thanks!