New subject: Getting to git

10 Jan 2017

Now I hope that I will also be heard, even if I am not an Inkscape
developer (but help with the website, which is also supposed to move,
even to move first).

I've been using github for a couple of small projects (gitlab didn't
really exist yet when I opened that account) for a couple of
years, and
also collaborated on a handful of other projects there, so had a
chance
to get acquainted with their interface.

The Inkscape project, however, I'd prefer to see on gitlab, a fellow
open source project. They are making an impact not only on the
web, but
also on people's own servers, which may be a reason why their own
project is the most-loved and most-discussed one in their listing. If
Inkscape moves there, we could really make an impact for them, too.

As ideology is a very un-practical argument (but nonetheless very
important to me), I've spent yesterday night to look into gitlab
and to
investigate what they offer. These are my main findings:

Github -> Gitlab user flow
--------------------------
- One of the main things to note is that you do *not* even need a new
account for it, if you're a github user. You can use OAuth, and login
with your github user name.

- Importing your own projects from github works just as easily. Allow
gitlab to access them, select from a list, wait a couple of
seconds, and
it's all there, including issues.

Interface
---------

- I didn't like their interface, when I was logged out, because it
didn't seem suitable for working on a desktop computer. After
logging in
and exploring, I found it's configurable, and I don't need to use the
100% width, but can have a menu at the left.

- They seem to like and support open formats, like SVG. This shows in
them displaying uploaded SVGs in img tags in comments. This also shows
in SVG user avatars being allowed - I've never before used a web
service
where that works.

- I like that they support text formatting almost everywhere (I
could do
without the emojis, but they seem to be en vogue).

Functionality
-------------

- The functionality that invites 'drive-by' edits seems to be
identical
to github's (i.e. forking + their code editor). Diff pages look
the same
to me. Comments on commit diffs work, too.

- Repos appear to have a 10GB limit.

- Here's a list of things that they may charge for, which are
currently
free / unlimited:
https://about.gitlab.com/gitlab-com/
<https://about.gitlab.com/gitlab-com/> (at the bottom)
I don't have enough info to tell if any of that can become a problem
(maybe the build minutes).
Also, I don't have github data to compare.

- I know that bug reports are not supposed to move along with Inkscape
(although I don't know the exact reasons, and think it's dangerous to
introduce this kind of divide between users and developers), but
looked
at that, too. They have a little less functionality than the very
elaborate filtering that is possible on lp. Nothing that would prevent
me from using those for inkscape-web, though. The things that aren't
available could be replaced by tags. They support the creation of
different report templates. Ordering by 'likes' (which could be the
equivalent of the 'affects me, too' button on lp) is possible, too.
Their 'issue board' looks like half of a Trello board.

- As someone stated before, their CI seems to develop. This is a thing
that I can't really investigate, because I don't know enough about
what
is needed (I only know that after Johan left, the existing Jenkins
server at http://jenkins.inkscape.org/ wasn't cared for by anyone on a
regular basis).

Could someone with the required knowledge do a re-evaluation?
Start here: https://gitlab.com/help/ci/quick_start/README
<https://gitlab.com/help/ci/quick_start/README>
(does the option to upload docker images mean that one could use
anything that one could possibly want to use?)
Here someone used it, but could have been on their private
installation:
http://ghostlyrics.net/building-and-deploying-a-c-library-with-gitlab.html
<http://ghostlyrics.net/building-and-deploying-a-c-library-with-gitlab.html>

- I also couldn't test their groups/permission management.

- They do have an API, which seems to allow to do everything from the
command line, or to write one's own scripts that do anything that
their
interface doesn't (https://docs.gitlab.com/ee/api/README.html
<https://docs.gitlab.com/ee/api/README.html>)

Conclusion
----------

So, in conclusion, for inkscape-web, gitlab seems to be more than
sufficient. For Inkscape-the-program, I'd like to see it there rather
than on github. The things that I can compare are at least equivalent,
and it has the additional benefit of being open source, but I cannot
evaluate all the features properly.

Maybe contacting them would also be interesting. Who knows if they
wouldn't want to offer help for a project as popular as Inkscape?

Maren

Am 10.01.2017 um 19:25 schrieb Ben Scholzen:
> Just putting my two cents in here as well.
>
> From my experience, probably most open source developers have a
GitHub
> account, which means that forking and working on Inkscape would
be less
> pain if it'd be hosted there.
>
> On 10.01.2017 19:14, Bryce Harrington wrote:
>> ----- Forwarded message from Will Entriken
<fulldecent@...400... <mailto:fulldecent@...400...>> -----
>>
>> Date: Mon, 9 Jan 2017 22:04:20 -0500
>> From: Will Entriken <fulldecent@...400...
<mailto:fulldecent@...400...>>
>> To: Bryce Harrington <bryce@...961...
<mailto:bryce@...961...>>
>> Subject: Re: [Inkscape-devel] Getting to git]
>>
>> Bryce,
>>
>> Thanks for keeping me in the loop.
>>
>> I am very glad to hear about this update. And I fully expect
much more
>> participation in the project as soon as we open this floodgate.
With a
>> good amount of focus and direction, we should be able to get to
1.0.0
>> quickly.
>>
>> Here's my two cents on the GitLab / GitHub choice. GitHub is my
>> default choice because it is the most popular and it works well
>> enough. The largest project I contribute to there is
>> https://github.com/chartjs/Chart.js
<https://github.com/chartjs/Chart.js> and GitHub has no pain points in
>> dealing with that project. I am more than happy to pick
>> indie/non-popular choices if they are better. And if attracting
>> developers and casual contributions is the goal then I do not think
>> GitLab is better.
>>
>> More thorough discussion:
>>  * Top 10 projects of all time on GitLab:
>> https://gitlab.com/explore/projects/starred
<https://gitlab.com/explore/projects/starred>
>>  * Top 10 projects this week on GitHub:
>> https://github.com/trending?since=weekly
<https://github.com/trending?since=weekly>
>>  * Random projects I started:
>>
https://github.com/fulldecent?language=&page=2&q=&tab=repositories&type=source&utf8=%E2%9C%93
<https://github.com/fulldecent?language=&page=2&q=&tab=repositories&type=source&utf8=%E2%9C%93>
>>
>>
>> Projects got more exposure (measured in stars) this week on GitHub
>> than the whole history of GitLab. And my random pet projects (which
>> are way less notable than Inkscape) would rank displace the #1,
#2 and
>> #6 projects on GitLab. Therefore I think Inkscape would be a
big fish
>> in a small pond there.
>>
>> Please feel free to copy this to the mailing list.
>>
>> Thank you,
>> Will
>>
>>
>> William Entriken
>> 267-738-4201
>>
>>
>>
>> On Fri, Jan 6, 2017 at 2:32 PM, Bryce Harrington
>> <bryce@...961... <mailto:bryce@...961...>>
wrote:
>>> Hi fulldecent,
>>>
>>> I know you've been very active in maintaining the Inkscape
mirror on
>>> github.  You'll be pleased to hear the project is finally
getting around
>>> to its git migration planning.
>>>
>>> We're debating gitlab vs. github as our way forward, and while
we're
>>> still pretty early in discussion currently consensus appears to be
>>> favoring gitlab.  I'd like to make sure we loop you in (and
any others
>>> you've worked with on the github mirror) to make sure we have
plenty of
>>> input and information before reaching any decisions.
>>>
>>> Feel free to either chime in on the mailing list with your
thoughts, or
>>> just reply to me and I'll pass them along.
>>>
>>> Thanks,
>>> Bryce
>>>
>>> ----- Forwarded message from Bryce Harrington
>>> <bryce@...961... <mailto:bryce@...961...>>
-----
>>>
>>> Date: Fri, 6 Jan 2017 11:21:28 -0800
>>> From: Bryce Harrington <bryce@...961...
<mailto:bryce@...961...>>
>>> To: Ted Gould <ted@...11... <mailto:ted@...11...>>
>>> Cc: inkscape-devel@lists.sourceforge.net
<mailto:inkscape-devel@lists.sourceforge.net>
>>> Subject: Re: [Inkscape-devel] Getting to git
>>>
>>> On Fri, Jan 06, 2017 at 08:56:50AM -0600, Ted Gould wrote:
>>>> I think where this mindshare comes into play is less with
drive-by
>>>> contributions as much as integrations with other services.
There are a
>>>> lot of cool tools built that "just work" with Github. Some
only support
>>>> that as a backend (even though they probably just grab git
and work on
>>>> it, they just don't have an option in the UI) What I'm not
sure about
>>>> is whether we'd end up using any of those services anyway.
>>>
>>> True, good points.  And yeah, most of the services I actually
have in
>>> mind for us just require git, so should work equally well
regardless of
>>> what we pick.
>>>
>>> I've used both gitlab and github for medium-sized projects,
and found
>>> them both to be usable and performant.  Inkscape is larger and
more
>>> complex, but so far I haven't seen any reason that either of those
>>> options wouldn't do the basic job as well as the other.
>>>
>>>>> We also liked the integration between
>>>>> bzr and the LP bug tracker -- we'll lose this, although
github/gitlab
>>>>> provide different integration opportunities that might
compensate a
>>>>> bit.
>>>>
>>>> Since the last time we've discussed it LP's Git features have
matured a
>>>> lot. I don't think that they're on par with Github/Gitlab
yet, but we
>>>> should probably put into consideration just moving to Git and
sticking
>>>> with LP, as it might be a simpler transition.
>>>
>>> You're right it should be included as an option.  If nothing
else it has
>>> familiarity and inertia to change working in its favor.  I'm
not sure
>>> how compelling a case can be made for it beyond that though.
>>>
>>> A large chunk of our transition pain will be developers
adapting to git,
>>> and we'll have that regardless of any of the options.  The
other chunk
>>> is the web interface and I suppose there's some advantage with
Launchpad
>>> of sticking with the familiar.
>>>
>>> However, Launchpad has remained understaffed for years, and
>>> unfortunately it really shows.  The chances of reaching parity
with
>>> github/gitlab seem slim to none.  And once their development
attention
>>> moves on to the next feature development effort, ongoing
maintenance of
>>> the service is going to become an issue, as it already has in
so many
>>> other places.
>>>
>>> Like gitlab, launchpad is open source.  I've contributed LP code
>>> myself.  I don't know if we're likely to invest development
energy here,
>>> but both gitlab and LP seem on the same level here.  gitlab
seems to
>>> have a more active development community though.
>>>
>>>>> Let me know your thoughts, and help us drive towards a
consensus for
>>>>> github or gitlab.
>>>>
>>>> For me, right now, my feelings are towards Gitlab. I think
that we
>>>> should stick with a FOSS solution overall and Github doesn't have
>>>> enough advantages to override that.
>>>
>>> There was a ticket on our github account with people pushing
for github
>>> adoption.  I'll touch base with them to gather their input,
but at this
>>> point I agree with you it seems gitlab is looking the better
option for
>>> Inkscape.
>>>
>>> Bryce
>>>
>>>
------------------------------------------------------------------------------
>>>
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
>>> _______________________________________________
>>> Inkscape-devel mailing list
>>> Inkscape-devel@lists.sourceforge.net
<mailto:Inkscape-devel@lists.sourceforge.net>
>>> https://lists.sourceforge.net/lists/listinfo/inkscape-devel
<https://lists.sourceforge.net/lists/listinfo/inkscape-devel>
>>>
>>> ----- End forwarded message -----
>>
>> ----- End forwarded message -----
>>
>>
------------------------------------------------------------------------------
>>
>> Developer Access Program for Intel Xeon Phi Processors
>> Access to Intel Xeon Phi processor-based developer platforms.
>> With one year of Intel Parallel Studio XE.
>> Training and support from Colfax.
>> Order your platform today. http://sdm.link/xeonphi
>> _______________________________________________
>> Inkscape-devel mailing list
>> Inkscape-devel@lists.sourceforge.net
<mailto:Inkscape-devel@lists.sourceforge.net>
>> https://lists.sourceforge.net/lists/listinfo/inkscape-devel
<https://lists.sourceforge.net/lists/listinfo/inkscape-devel>
>>
>
>
>
>
>
------------------------------------------------------------------------------
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi
>
>
>
> _______________________________________________
> Inkscape-devel mailing list
> Inkscape-devel@lists.sourceforge.net
<mailto:Inkscape-devel@lists.sourceforge.net>
> https://lists.sourceforge.net/lists/listinfo/inkscape-devel
<https://lists.sourceforge.net/lists/listinfo/inkscape-devel>
>

------------------------------------------------------------------------------
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
_______________________________________________
Inkscape-devel mailing list
Inkscape-devel@lists.sourceforge.net
<mailto:Inkscape-devel@lists.sourceforge.net>
https://lists.sourceforge.net/lists/listinfo/inkscape-devel
<https://lists.sourceforge.net/lists/listinfo/inkscape-devel>

Re: [Inkscape-devel] Getting to git

Bryce Harrington

Ben Scholzen

Maren Hachmann

Victor Westmann

Github -> Gitlab user flow

Interface

Functionality

Conclusion

Mark Schafer

Maren Hachmann

tags (0)

participants (5)