steps to reproduce:

virtualenv ./venv
. ./venv/bin/activate
pip install --upgrade pip
pip install --upgrade setuptools
hg clone http://hg.code.sf.net/p/roundup/code roundup-code
cd roundup-code
pip install --editable .

result

error: [Errno 2] No such file or directory:
'/home/user/src/roundup-code/roundup-mailgw'

I can confirm that I am seeing a similar issue when using the
-e/--editable option with pip. Invoking pip without the -e option does
not result in an error.

So there seems to be some issue with "development mode"[1] that needs
looking into a bit closer.


[1] https://pythonhosted.org/setuptools/setuptools.html#development-mode

a worked solution in attachement:

works with "pip install ." and "pip install -e ."

tested with:
pip (6.0.7)
setuptools (12.0.5)

This patch works for me as well.

In regard to the patch the use of the map function makes it non-obvious
what is going on (ie. requires more brain power to figure what is
happening). It may be simpler to make the equivalent changes in the
scriptname() function instead.

I have slightly reworked the patch an push the changes to the repository:

  http://hg.code.sf.net/p/roundup/code/rev/cdfb1a3fb56f

Thanks for the original patch tonich

This change breaks scripts installation with `setup.py install`. Working 
on a proper fix for this.

The problem is that "pip install --editable ." converts setup.py from 
distutils to setuptools script and ignores build_scripts step.

setuptools instead calls its own install_egg_scripts() on "develop", which 
just copies the stuff verbatim.

@techtonik also see issue2550898

@jerrykan aha, thanks for the pointer. I was about to report the same 
results. =)

Another try

ping. Please take a look on the second patch. It solves both problems
(this and #2550898).

Jerrykan or anatoly, can one of you evaluate the second path
on this issue and see if it solves the problem?

If so I'll get it integrated.


-- rouilj

ping John, Anatoly

have either of you evaluated the second patch to see if it fixed the issues?

-- rouilj

Re-ping does anybody understand what's happening here and if
the second patch fixes it?

-- rouilj

Triple ping JerryKan or Anatoly can either of you validate the patch? I 
would like to get it into 1.6.0, but we are late in the process. I don't 
want to chance breaking something and I don't understand what's happening 
here enough to make an assessment.

Hello John,

Great to hear a v1.6.0 release might be coming soon.

This issue really only affects developers/contributors who are working 
on the roundup codebase so I don't really see it as being a blocker for 
a v1.6.0 release.

I would suggest that if any patch were to be merged, it might be safer 
to do it after the v1.6.0 release given that changes could potentially 
cause problems for end-users (see #2550898 for an example of where this 
has happened before).

SeeYa,
John.

On 01/06/18 11:59, John Rouillard wrote:
> 
> John Rouillard added the comment:
> 
> Triple ping JerryKan or Anatoly can either of you validate the patch? I
> would like to get it into 1.6.0, but we are late in the process. I don't
> want to chance breaking something and I don't understand what's happening
> here enough to make an assessment.
> 
> ________________________________________________
> Roundup tracker <issues@roundup-tracker.org>
> <http://issues.roundup-tracker.org/issue2550866>
> ________________________________________________
>

I am closing this. Developers can run from the tarball without the need to install as we can now run python roundup/scripts/roundup_server with what looks like live reload.

I don't think that existed when anatoly wanted live reload in the demo.py script for his jinja2 work.

There is already a reference to this ticket in issue2550899,
one of the distutils -> setuptools tickets. 
The work to make sure the patches are polyglot etc. can be done there.