roundup-demo fixes #2

jerrykan · 2021-02-04T05:17:46Z

A couple of small fixes to improve the user's experience when using roundup-demo.

My git-remote-hg setup is currently broken some I'm unable to push directly to the SourceForge Mercurial repo. So I'm creating a pull request so either I remember to merge this fixes when I get git-remote-hg working, or someone else merges them and pushes them to SourceForge.

Some templates like `jinja2` and `responsive` need certain settings in their `config.ini` otherwise the pages won't display correctly. By pulling in these custom settings when running `roundup-demo` we ensure that the pages will display correctly out of the box with no extra changes required.

The `roundup-demo` command assumes that the `user` has a `realname` attribute when it creates the demo user. Without adding the `realname` attribute in the `minimal` template's schema an error will occur when a user tries to create a new demo based on it.

codecov-io · 2021-02-04T05:58:58Z

Codecov Report

Merging #2 (8900803) into master (d217654) will decrease coverage by 0.00%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##           master       #2      +/-   ##
==========================================
- Coverage   70.14%   70.13%   -0.01%     
==========================================
  Files          96       96              
  Lines       20744    20745       +1     
==========================================
  Hits        14550    14550              
- Misses       6194     6195       +1

Impacted Files	Coverage Δ
roundup/demo.py	`0.00% <0.00%> (ø)`

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d217654...8900803. Read the comment docs.

rouilj · 2021-02-04T15:19:03Z

Hi John: In message <roundup-tracker/roundup/pull/[email protected]>, John Kristensen writes:

A couple of small fixes to improve the user's experience when using `roundup-demo`. My `git-remote-hg` setup is currently broken some I'm unable to push directly to the SourceForge Mercurial repo. So I'm creating a pull request so either I remember to merge this fixes when I get `git-remote-hg` working, or someone else merges them and pushes them to SourceForge.

The emailed version of the codecov report for this pull shows a drop of 0.5% in overall coverage. But the one on this pull request shows 0.01%. The email report shows less coverage for files your patch didn't change. I assume this means your repo is behind the head revision of mercurial and thus missing tests?? It's wierd that the online report codecov appended to the pull reports differently from the emailed report though. Also do you have any ideas on running demo.py as part of the tests? At the very least it would probably be good to fire up demo.py with the anydbm backend and each template. That should have caught each of the problems your fixing. Thoughts?

…

-- -- rouilj John Rouillard =========================================================================== My employers don't acknowledge my existence much less my opinions.

rouilj · 2021-02-17T04:45:12Z

Also John, I don't think you need:

Add realname attribute to user in minimal template schema

in demo.py is this bit of code:

    # FIXME: Move tracker-specific demo initialization into the tracker
    # templates.
    if template == 'minimal':
        db.user.create(username='demo', password=password.Password('demo'),
                       roles='User')
    else:
        db.user.create(username='demo', password=password.Password('demo'),
                       realname='Demo User', roles='User')

whish should not try initializing the realname property if using the minimal template. Maybe the bug that needs
to be fixed is that template == 'minimal' is not true for the minimal template???

jerrykan · 2021-02-26T06:03:23Z

whish should not try initializing the realname property if using the minimal template. Maybe the bug that needs
to be fixed is that template == 'minimal' is not true for the minimal template???

It seems that def install_demo(home, backend, template) may have two different types of string value depending on where it is invoked from.

When invoked from roundup/scripts/roundup_demo.py (ie. by running roundup-demo) the install_demo() function is passed a template value that is the full path to the template directory. In this case the template == 'minimal' won't match and will result in an Exception.

When invoked from roundup/demo.py (ie. by running demo.py) the install_demo() function is passed a template value that is just the name of the template. In this case the template == 'minimal' will match the template == 'minimal' condition and work as expected.

I should also note that install_demo() makes some assumptions about he location of the templates, so running demo.py doesn't work unless your working directory is the same as where the demo.py script is located. On the other hand round-admin works correctly when run from any working directory.

rouilj · 2021-03-02T17:24:26Z

Ah thanks for tracking that down. I didn't even realize there was a roundup-demo to tell the truth. I wonder what the use
case for that is. I guess is allows a roundup demo for roundup installed via an OS package?

What do you think about fixing template == 'minimal' check so that it strips the path from template (os.path.basename(template)) and only compares the last component. That should fix the roundup-demo
case when used with the installed version of the distributed minimal template.

I am ok with requiring the current working directory be the same directory where demo.py is located. IIRC all of our
docs have that assumption as well.

Thoughts?

jerrykan · 2021-03-09T01:08:37Z

Ah thanks for tracking that down. I didn't even realize there was a roundup-demo to tell the truth. I wonder what the use
case for that is. I guess is allows a roundup demo for roundup installed via an OS package?

Using roundup-demo is the easiest (only?) way to init a demo instance after doing a pip install roundup. Given that pip is the standard way to install python packages maybe we should be optimising for that use-case, instead of requiring user to checkout the source code to run demo.py?

jerrykan · 2021-03-09T01:10:44Z

Maybe we should move the discussion about the demo stuff to the mailing list? I didn't mean to subvert other communication channels by opening a "placeholder" pull request :D

rouilj · 2021-03-10T02:42:19Z

In message ***@***.***>, John Kristensen writes:

Maybe we should move the discussion about the demo stuff to the mailing list? I didn't mean to subvert other communication channels by opening a "placeholder" pull request :D

We can do that. Do you want to post the initial summary email or should I? Have a great morning.

…

-- -- rouilj John Rouillard =========================================================================== My employers don't acknowledge my existence much less my opinions.

jerrykan · 2021-03-16T09:18:27Z

Discussion moved to the mailing list: https://sourceforge.net/p/roundup/mailman/message/37240835/

jerrykan · 2021-04-12T02:09:26Z

Just adding a note here for when work on these patches get picked-up and worked on again. It might be work looking at using the setuptools ResourceManager API as a more portable/reliable way to accessing template files.

rouilj · 2021-12-09T20:30:04Z

Committed in http://sourceforge.net/p/roundup/code/ci/5a3a386aa8e7

Regarding ResourceTools, I guess I am too dumb to see how that helps deal with the man page, template dir, locale file resources.

jerrykan added 2 commits February 4, 2021 16:12

rouilj closed this Dec 9, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

roundup-demo fixes #2

roundup-demo fixes #2

Uh oh!

jerrykan commented Feb 4, 2021

Uh oh!

codecov-io commented Feb 4, 2021 •

edited

Loading

Uh oh!

rouilj commented Feb 4, 2021 via email

Uh oh!

rouilj commented Feb 17, 2021

Uh oh!

jerrykan commented Feb 26, 2021 •

edited

Loading

Uh oh!

rouilj commented Mar 2, 2021

Uh oh!

jerrykan commented Mar 9, 2021

Uh oh!

jerrykan commented Mar 9, 2021

Uh oh!

rouilj commented Mar 10, 2021 via email

Uh oh!

jerrykan commented Mar 16, 2021

Uh oh!

jerrykan commented Apr 12, 2021 •

edited

Loading

Uh oh!

rouilj commented Dec 9, 2021

Uh oh!

Uh oh!

roundup-demo fixes #2

roundup-demo fixes #2

Uh oh!

Conversation

jerrykan commented Feb 4, 2021

Uh oh!

codecov-io commented Feb 4, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

rouilj commented Feb 4, 2021 via email

Uh oh!

rouilj commented Feb 17, 2021

Uh oh!

jerrykan commented Feb 26, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rouilj commented Mar 2, 2021

Uh oh!

jerrykan commented Mar 9, 2021

Uh oh!

jerrykan commented Mar 9, 2021

Uh oh!

rouilj commented Mar 10, 2021 via email

Uh oh!

jerrykan commented Mar 16, 2021

Uh oh!

jerrykan commented Apr 12, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rouilj commented Dec 9, 2021

Uh oh!

Uh oh!

codecov-io commented Feb 4, 2021 •

edited

Loading

jerrykan commented Feb 26, 2021 •

edited

Loading

jerrykan commented Apr 12, 2021 •

edited

Loading