[sword-devel] DVCS (was Re: Project "Free Scriptures" started)

[Greg Hellings]

On Wed, Feb 26, 2014 at 8:24 AM, Troy A. Griffitts <scribe at crosswire.org>wrote:

> Quickly, regarding DVCS, the argument to keeping SVN over a DVCS isn't
> that a DVCS isn't 'better' in some ways or as Peter has suggested, wouldn't
> let us maintain tight review of our code or that we simply don't ever want
> to change anything. We eventually switched from CVS to SVN :) The reason
> the core engine uses SVN is because it was the best and most popular choice
> 12 years ago or so when we chose to use it, and basically I don't like git
> (yet). It's pretty simple really. Other CrossWire projects use various
> choices for source control. You can argue about the popularity of DVCS all
> you want, but the truth is, there is not one ubiquitous opensource
> development choice today like there was 12 years ago when we switched to
> SVN, so I'm not yet motivated by popularity. And the counter argument is,
> if you like git, use it. Many contributors to the core engine do. Sure you
> don't get absolutely all the benefits like pull requests, but if you're
> concerned about that, attach your patch to the bug report. VCS is a choice,
> I'm typically conservative (if you haven't guessed) and haven't chosen to
> change the part of CrossWire I maintain. It is not keeping contributors
> from contributing.
>

I've been successfully using git-svn for a while now on my work with
CrossWire. It has made it at least one order of magnitude more difficult to
collaborate with others than if Sword used git directly, but it has
simultaneously made it one order of magnitude easier to collaborate than if
git-svn wasn't an option. But I still have to revert back to the actual svn
commands for maintaining the sword-1-7-x branch because the overlays that
git-svn does are incompatible with svn's concept of branching.

The claim that there isn't one dominant DVCS in use by open source projects
does not hold up with the reading and research I've done. The share info
I've seen shows the relative popularity among open source projects with
Subversion hosting about 45% of projects, git hosting about 37% of
projects, CVS hosting about 10% with Mercurial and Bazaar each hosting
about 1-1.5% of open source projects in surveys by Ohloh. So the community
has spoken - the DVCS of choice is git. But the songs of Subversion's
demise are just a little premature at this point. Its share of projects is
falling, but it still ranks as the most popular choice.

But don't get me wrong. I very strongly wish Sword and Xiphos would join
the bandwagon, even if it is a bit noisy up here between the trumpets and
snare drums. I say that as an occasional committer, but mainly as the
person maintaining an official branch and an unofficial branch (for the
Xiphos Windows Unicode patch which I've still not quite gotten a chance to
mainline yet - it was still having some trouble with the UTF-8/UTF-16
conversion methods).


>
> Though as an aside, just from a personal query, how do you guys get
> Windows and GUI Mac users to use git? I have at least 2 ancillary projects
> with SVN repos on CrossWire's server where the users are not programmers,
> the command line scares them, and they had no clue what a versioning system
> was before I installed TortoiseSVN for them. They are HTML designers, flash
> programmers, transcription editors, etc. It was a real task to get them to
> even begin to use SVN, but the integration right into the OS helped (and we
> found some Mac-y GUI thingy that was simple to use on OSX). I don't want to
> train them about a stash as well. That's just one more obstacle. Are there
> really mature GUI clients like these for the simple VC users. I know there
> are power GUIs. That is not my question. I'm asking about tools integrated
> into the OS which make VC simpler for the simple use?
>

Windows has TortoiseGit, which integrates directly into the shell just like
TortoiseSVN does. I'd say it cuts down about 5% of the Git learning curve.
If you're dealing with non-technical people, you're still best to go with
SVN. On Mac, I haven't touched GUI tools for git because I only work with
Mac because it has a solid command line. I know there are somewhat popular
GUI Git tools out there for it, but I couldn't hope to point you in the
right direction.

--Greg
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140226/7ad067bf/attachment.html>

[Peter von Kaehne]

> On Wed, Feb 26, 2014 at 8:24 AM, Troy A. Griffitts
> <scribe at crosswire.org> wrote:
> Quickly, regarding DVCS, the argument to keeping SVN over a
> DVCS isn't that a DVCS isn't 'better' in some ways or as Peter
> has suggested, wouldn't let us maintain tight review of our
> code or that we simply don't ever want to change anything. We
> eventually switched from CVS to SVN :) The reason the core
> engine uses SVN is because it was the best and most popular
> choice 12 years ago or so when we chose to use it, and
> basically I don't like git (yet).

FWIW I use git a lot module making and for maintaining utilities/scripts
during module making. I do most of my module making etc when I am off,
and a lot of times where I have little access to the net.

I use it as if I had several version control servers in series. I have
two computers, one luggable, one small and truly portable. And a central
server.

I want to work on both laptops in a seamless fashion. So I have a
central Git server to where I push whenever I can and where I pull from
whenever i can.

I have on each laptop all relevant projects under git.

So, as a result I can version and branch around on both laptops, knowing
I can always go back to where I started and can whenever i have a chance
bring both laptops into sync, including all trial branches etc.

Is it easy to use? If you use nothing but add, commit, push and pull,
you have in essence a svn like makup with little or no difference apart
from the benefit of local version control. But from there you can expand
with essentially no limit. SVN is a straightjacket in comparison.

Could a noob learn it? I did. I find it hugely helpful. There are
concepts I have not understood, but I did not need them so far either.

The book I found most helpful is Git Magic. It is simple and
straightforward. Your non-programmer friends could learn it, not that
they would need to. If their projects are not sword, what stops them
from continuing on svn.


> It is not keeping contributors from contributing.

And I think this is simply not true. You got Nic stating that he holds
things back, you got Greg stating that he has stuff not committed, you
have Jaak who clearly has grander plans. We have Kostya who has quietly
worked away on mapping until he got nearly blasted by Chris (though that
seems to be now ironed out). And my own stuff on the filters is still
not in either. And none of us are really able to collaborate in a
sensible fashion other than by exchanging patches or throwing up
alternative locations. Which causes difficulties in its own right.

So things _are_ held back. Maybe not by SVN per se but by the inability
to commit to a central location where others can comment and approve and
share and improve - without impacting on the central branch which needs
full approval. And while Git vs SVN would not be the thing which makes
you approve things faster, Git would allow all of us others to
collaborate speedily towards new releases instead of waiting that every
single small bit gets approved.

On Wed, 2014-02-26 at 08:43 -0600, Greg Hellings wrote:

> I've been successfully using git-svn for a while now on my work with
> CrossWire.

And I looked at it and despaired. I am a noob, admittedly. I find these
things hard. And yet I got git to work and work well for my purposes.

Peter
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140226/212a3763/attachment.sig>

[Troy A. Griffitts]

While I enjoy hearing all the the praises of git from its fanbase, I don't agree with your deductions. There are at least two sword git repositories publicly available which I know about. The point is, git is a DVCS. Collaboration happens between distributed clones regardless if the source from which they fork even still exists. It shouldn't matter to a git user if the source is an svn repository, except as a matter of religion or politics. :)

When git gits to 65+% popularity among open source projects, you can start waiving the popularity flag. I am using git for some projects, even programmatically integrating into some. I don't hate git. I am just not yet comfortable enough and convinced that the pain to change the core repository exceeds the benefit, especially marked by the fact that git users can git away in their distributed repos to their hearts' content right now.

Nic, you make it sound like I've repeatedly dropped multiple patches you've tried to submit. I hope I've not been that irresponsible.

Peter, the issue with your recent commit problems is permission to directly commit freely to specific resources in the authoritative repo. I'd have this same problem setting up your perms to push directly from git. In fact, I'd have more problems in git because I don't know how :) At least I know how an svn access file works. I just need to finish your permissions properly.



On February 26, 2014 8:43:25 AM MST, Peter von Kaehne <refdoc at gmx.net> wrote:
>
>
>> On Wed, Feb 26, 2014 at 8:24 AM, Troy A. Griffitts
>> <scribe at crosswire.org> wrote:
>> Quickly, regarding DVCS, the argument to keeping SVN over a
>> DVCS isn't that a DVCS isn't 'better' in some ways or as
>Peter
>> has suggested, wouldn't let us maintain tight review of our
>> code or that we simply don't ever want to change anything. We
>> eventually switched from CVS to SVN :) The reason the core
>> engine uses SVN is because it was the best and most popular
>> choice 12 years ago or so when we chose to use it, and
>> basically I don't like git (yet).
>
>FWIW I use git a lot module making and for maintaining
>utilities/scripts
>during module making. I do most of my module making etc when I am off,
>and a lot of times where I have little access to the net.
>
>I use it as if I had several version control servers in series. I have
>two computers, one luggable, one small and truly portable. And a
>central
>server.
>
>I want to work on both laptops in a seamless fashion. So I have a
>central Git server to where I push whenever I can and where I pull from
>whenever i can.
>
>I have on each laptop all relevant projects under git.
>
>So, as a result I can version and branch around on both laptops,
>knowing
>I can always go back to where I started and can whenever i have a
>chance
>bring both laptops into sync, including all trial branches etc.
>
>Is it easy to use? If you use nothing but add, commit, push and pull,
>you have in essence a svn like makup with little or no difference apart
>from the benefit of local version control. But from there you can
>expand
>with essentially no limit. SVN is a straightjacket in comparison.
>
>Could a noob learn it? I did. I find it hugely helpful. There are
>concepts I have not understood, but I did not need them so far either.
>
>The book I found most helpful is Git Magic. It is simple and
>straightforward. Your non-programmer friends could learn it, not that
>they would need to. If their projects are not sword, what stops them
>from continuing on svn.
>
>
>> It is not keeping contributors from contributing.
>
>And I think this is simply not true. You got Nic stating that he holds
>things back, you got Greg stating that he has stuff not committed, you
>have Jaak who clearly has grander plans. We have Kostya who has quietly
>worked away on mapping until he got nearly blasted by Chris (though
>that
>seems to be now ironed out). And my own stuff on the filters is still
>not in either. And none of us are really able to collaborate in a
>sensible fashion other than by exchanging patches or throwing up
>alternative locations. Which causes difficulties in its own right.
>
>So things _are_ held back. Maybe not by SVN per se but by the inability
>to commit to a central location where others can comment and approve
>and
>share and improve - without impacting on the central branch which needs
>full approval. And while Git vs SVN would not be the thing which makes
>you approve things faster, Git would allow all of us others to
>collaborate speedily towards new releases instead of waiting that every
>single small bit gets approved.
>
>On Wed, 2014-02-26 at 08:43 -0600, Greg Hellings wrote:
>
>> I've been successfully using git-svn for a while now on my work with
>> CrossWire.
>
>And I looked at it and despaired. I am a noob, admittedly. I find these
>things hard. And yet I got git to work and work well for my purposes.
>
>Peter
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>sword-devel mailing list: sword-devel at crosswire.org
>http://www.crosswire.org/mailman/listinfo/sword-devel
>Instructions to unsubscribe/change your settings at above page

--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140226/62db088f/attachment.html>

[Peter Von Kaehne]

> Von:?"Troy A. Griffitts" <scribe at crosswire.org>ptures" started)

> While I enjoy hearing all the the praises of git from its fanbase, I don't agree with your deductions. There are at least two sword git repositories publicly available which I know about. The point is, > git is a DVCS. Collaboration happens between distributed clones regardless if the source from which they fork even still exists. It shouldn't matter to a git user if the source is an svn repository, > except as a matter of religion or politics. :)

True, but politics matter and religion too. If the project is perceived to be too stagnant then one of the git clones will start acting like a new master. And if there is an added difficulty in that the main contributor/controller does not participate in the git interplay then this becomes a lot more likely. or even more likely we have people turning away.
?
> When git gits to 65+% popularity among open source projects, you can start waiving the popularity flag.

I think the question is whether DVCS or non DVCS. As far as popularity within the DVCS camp goes - git is it.

> Nic, you make it sound like I've repeatedly dropped multiple patches you've tried to submit. I hope I've not been that irresponsible.

Troy, if you look at the mad rush and long delays before the last release, I think, without wishing to be insulting, it can be fairly said you are dropping things. Again and again. And if people persist, they get things in (I got my bibliography stuff in), if they do not, they do not. I am around for long enough and trusted well enough that I get my stuff done, eventually, but others do not. What about Kostya's stuff?

Peter

[Chris Burrell]

Even though it is a dvcs, it is usually the norm to want to push back to a
master branch especially for redistributable libraries.
Chris
On 26 Feb 2014 17:23, "Peter Von Kaehne" <refdoc at gmx.net> wrote:

> > Von: "Troy A. Griffitts" <scribe at crosswire.org>ptures" started)
>
> > While I enjoy hearing all the the praises of git from its fanbase, I
> don't agree with your deductions. There are at least two sword git
> repositories publicly available which I know about. The point is, > git
> is a DVCS. Collaboration happens between distributed clones regardless if
> the source from which they fork even still exists. It shouldn't matter to a
> git user if the source is an svn repository, > except as a matter of
> religion or politics. :)
>
> True, but politics matter and religion too. If the project is perceived to
> be too stagnant then one of the git clones will start acting like a new
> master. And if there is an added difficulty in that the main
> contributor/controller does not participate in the git interplay then this
> becomes a lot more likely. or even more likely we have people turning away.
>
> > When git gits to 65+% popularity among open source projects, you can
> start waiving the popularity flag.
>
> I think the question is whether DVCS or non DVCS. As far as popularity
> within the DVCS camp goes - git is it.
>
> > Nic, you make it sound like I've repeatedly dropped multiple patches
> you've tried to submit. I hope I've not been that irresponsible.
>
> Troy, if you look at the mad rush and long delays before the last release,
> I think, without wishing to be insulting, it can be fairly said you are
> dropping things. Again and again. And if people persist, they get things in
> (I got my bibliography stuff in), if they do not, they do not. I am around
> for long enough and trusted well enough that I get my stuff done,
> eventually, but others do not. What about Kostya's stuff?
>
> Peter
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140226/217bb636/attachment.html>

[Ryan Hiebert]

Above and beyond the benefits of using a DVCS, using GitHub provides a
place for collaboration and code review that a mailing list cannot provide.
There are some things that a mailing list is still well suited for, but
when you're wishing to discuss code and collaborate around code, GitHub
really is amazing for that.

I would personally consider the choice of whether to use GitHub (or another
code collaboration site if you must) to be a strong indicator as to whether
you want to have more people involved and helping.

Ryan Hiebert
a lurker
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140226/458a571d/attachment-0001.html>

[Nic Carter]

On 27 Feb 2014, at 3:44 am, Troy A. Griffitts <scribe at crosswire.org> wrote:

> Nic, you make it sound like I've repeatedly dropped multiple patches you've tried to submit. I hope I've not been that irresponsible.

I have both submitted patches and also emailed asking if I should submit a patch when I have modified things, and I've had no response.
I have also submitted patches that have been accepted, though! :) [ insert comment about HTML parsing :D ]

I'm happy keeping my own fork, although it does waste some of my time when I then need to manually merge things when SWORD is updated.

My only comment about "irresponsible" is that I wouldn't say that!!! I would comment that we are all volunteers and there's a big lack of time and resources. (And here I am largely thinking of myself right now! One could just as easily comment that I have been irresponsible in how long it's taken for the latest version of PS (which is late in the beta cycle) to be officially released.)
:(


Thanks, ybic
nic... :)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140227/7047f97c/attachment.html>

[Matej Cepl]

On Thu, Feb 27, 2014 at 06:58:26PM +1100, Nic Carter wrote:
>I then need to manually merge things when SWORD is updated.

git svn rebase ????

Mat?j
--
http://www.ceplovi.cz/matej/, Jabber: mcepl<at>ceplovi.cz
GPG Finger: 89EF 4BC6 288A BF43 1BAB 25C3 E09F EF25 D964 84AC

My opinions may have changed, but not the fact that I am right.
--Ashleigh Brilliant

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 190 bytes
Desc: not available
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140227/6189f049/attachment.sig>

[Troy A. Griffitts]

Look guys, if you only read and consider any of my emails in their
entirely, make it this one.

1) I don't hate DVCS.
2) I believe git is the dominant opensource DVCS (nb: 'D') winner and
choice.
3) I have never said we WON'T switch to git at some time.
4) I have never said that SVN is superior to git.
5) I don't want anything to be 'closed' or 'hidden'.
____________________________________

The reason things aren't switching right now:

1) Contrary to popular belief, the project is not stagnant. You can
check the commit logs. I also currently have multiple working copies on
my personal box with code changes I am testing and improving before I
commit. I personally am still more comfortable in SVN than git-- in
both using daily and more importantly administrating. I make the
majority of the commits to the repository, second probably Chris, then
probably DM, and then likely GHellings after that (sorry if I've gotten
the order wrong). Only one of these people is pushing for changing the
central repository to git (GHellings).

2) If we switched the main repository to use git, this would effectively
change one thing:

git svn clone https://crosswire.org/svn/sword -s (plus: git svn rebase,
git svn dcommit) to:
git clone https://crosswire.org/svn/sword.git

This in itself will not facilitate us working together better or faster
or prevent 'losing' patches.
In summary, it is technically possible to already do these things with
git now and we are not.

3) We have a bug tracker with the ability to even integrate bugs and
patches with our VCS right now. We don't. We are not going to be
magically more disciplined if we switch to a distributed version control
system.
__________________________

Final thoughts and suggestions:

1) I like the integrated tools around git from projects like github
which I certainly do see the value of facilitating work together. I
would like to investigate a tool suite, like these provide, which we
could host on our servers. Think 'longevity'. I am not as concerned
about the coolest, newest thing as I am concerned about that toolset
still being a dominant force in 10 years.

2) I DON'T HATE GIT.

3) I certainly HEAR the cries for change.

4) We will probably change when I start using git personally for my own
SWORD work because I just can't live without its killer features.

5) If there are any outstanding patches which I do not yet know about,
please (*cough*) create a PCR and attach your patch. This should be
reasonably equivalent to a pull request.

6) Share and work together using git now. Show me how much progress
you've made collaborating together on a new SWORD feature and how great
your experience has been. Convince me that using a DVCS will change the
way we collaborate and you will go a long way to moving it up on my todo
list.

Sincerely (no, really, "Sincerely"),

Troy









On 02/27/2014 12:58 AM, Nic Carter wrote:
>
> On 27 Feb 2014, at 3:44 am, Troy A. Griffitts <scribe at crosswire.org
> <mailto:scribe at crosswire.org>> wrote:
>
>> Nic, you make it sound like I've repeatedly dropped multiple patches
>> you've tried to submit. I hope I've not been that irresponsible.
>
> I have both submitted patches and also emailed asking if I should
> submit a patch when I have modified things, and I've had no response.
> I have also submitted patches that have been accepted, though! :) [
> insert comment about HTML parsing :D ]
>
> I'm happy keeping my own fork, although it does waste some of my time
> when I then need to manually merge things when SWORD is updated.
>
> My only comment about "irresponsible" is that I wouldn't say that!!! I
> would comment that we are all volunteers and there's a big lack of
> time and resources. (And here I am largely thinking of myself right
> now! One could just as easily comment that I have been irresponsible
> in how long it's taken for the latest version of PS (which is late in
> the beta cycle) to be officially released.)
> :(
>
>
> Thanks, ybic
> nic... :)
>
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140227/fdd46657/attachment.html>

[Chris Burrell]

If you were after a tool to foster collaboration/tracking of patches, etc.
then Atlassian's Crucible is possibly the best one on the market for that.
You can then review commits or patches, branches, etc. People can upload
patches which you can then view against the source code.

It has a VERY and friendly interface, and does integrate with JIRA if you
want it to. It's very easy to use. You can also browse, view previous
revisions, compare any two things, etc.

It basically allows you to do what DM, Martin and I have been using GitHub
to do but is (D)VCS agnostic.

You could ask Atlassian for their On-Demand version if you didn't want to
spend the effort of setting things up initially in a stand-alone way. (STEP
has been using the On-Demand suite and been very pleased with it). Both
On-demand and standalone installs are available for free with an open
source license.

Chris



On 27 February 2014 19:32, Troy A. Griffitts <scribe at crosswire.org> wrote:

> Look guys, if you only read and consider any of my emails in their
> entirely, make it this one.
>
> 1) I don't hate DVCS.
> 2) I believe git is the dominant opensource DVCS (nb: 'D') winner and
> choice.
> 3) I have never said we WON'T switch to git at some time.
> 4) I have never said that SVN is superior to git.
> 5) I don't want anything to be 'closed' or 'hidden'.
> ____________________________________
>
> The reason things aren't switching right now:
>
> 1) Contrary to popular belief, the project is not stagnant. You can check
> the commit logs. I also currently have multiple working copies on my
> personal box with code changes I am testing and improving before I commit.
> I personally am still more comfortable in SVN than git-- in both using
> daily and more importantly administrating. I make the majority of the
> commits to the repository, second probably Chris, then probably DM, and
> then likely GHellings after that (sorry if I've gotten the order wrong).
> Only one of these people is pushing for changing the central repository to
> git (GHellings).
>
> 2) If we switched the main repository to use git, this would effectively
> change one thing:
>
> git svn clone https://crosswire.org/svn/sword -s (plus: git svn rebase,
> git svn dcommit) to:
> git clone https://crosswire.org/svn/sword.git
>
> This in itself will not facilitate us working together better or faster or
> prevent 'losing' patches.
> In summary, it is technically possible to already do these things with git
> now and we are not.
>
> 3) We have a bug tracker with the ability to even integrate bugs and
> patches with our VCS right now. We don't. We are not going to be
> magically more disciplined if we switch to a distributed version control
> system.
> __________________________
>
> Final thoughts and suggestions:
>
> 1) I like the integrated tools around git from projects like github which
> I certainly do see the value of facilitating work together. I would like
> to investigate a tool suite, like these provide, which we could host on our
> servers. Think 'longevity'. I am not as concerned about the coolest,
> newest thing as I am concerned about that toolset still being a dominant
> force in 10 years.
>
> 2) I DON'T HATE GIT.
>
> 3) I certainly HEAR the cries for change.
>
> 4) We will probably change when I start using git personally for my own
> SWORD work because I just can't live without its killer features.
>
> 5) If there are any outstanding patches which I do not yet know about,
> please (*cough*) create a PCR and attach your patch. This should be
> reasonably equivalent to a pull request.
>
> 6) Share and work together using git now. Show me how much progress
> you've made collaborating together on a new SWORD feature and how great
> your experience has been. Convince me that using a DVCS will change the
> way we collaborate and you will go a long way to moving it up on my todo
> list.
>
> Sincerely (no, really, "Sincerely"),
>
> Troy
>
>
>
>
>
>
>
>
>
>
> On 02/27/2014 12:58 AM, Nic Carter wrote:
>
>
> On 27 Feb 2014, at 3:44 am, Troy A. Griffitts <scribe at crosswire.org>
> wrote:
>
> Nic, you make it sound like I've repeatedly dropped multiple patches
> you've tried to submit. I hope I've not been that irresponsible.
>
>
> I have both submitted patches and also emailed asking if I should submit a
> patch when I have modified things, and I've had no response.
> I have also submitted patches that have been accepted, though! :) [ insert
> comment about HTML parsing :D ]
>
> I'm happy keeping my own fork, although it does waste some of my time
> when I then need to manually merge things when SWORD is updated.
>
> My only comment about "irresponsible" is that I wouldn't say that!!! I
> would comment that we are all volunteers and there's a big lack of time and
> resources. (And here I am largely thinking of myself right now! One could
> just as easily comment that I have been irresponsible in how long it's
> taken for the latest version of PS (which is late in the beta cycle) to be
> officially released.)
> :(
>
>
> Thanks, ybic
> nic... :)
>
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
>
>
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140227/ad2c3cf1/attachment-0001.html>

[Chris Little]

On 2/27/2014 11:32 AM, Troy A. Griffitts wrote:
> Look guys, if you only read and consider any of my emails in their
> entirely, make it this one.
>
> 1) I don't hate DVCS.
> 2) I believe git is the dominant opensource DVCS (nb: 'D') winner and
> choice.
> 3) I have never said we WON'T switch to git at some time.
> 4) I have never said that SVN is superior to git.
> 5) I don't want anything to be 'closed' or 'hidden'.
> ____________________________________
>
> The reason things aren't switching right now:
>
> 1) Contrary to popular belief, the project is not stagnant. You can
> check the commit logs. I also currently have multiple working copies on
> my personal box with code changes I am testing and improving before I
> commit. I personally am still more comfortable in SVN than git-- in
> both using daily and more importantly administrating. I make the
> majority of the commits to the repository, second probably Chris, then
> probably DM, and then likely GHellings after that (sorry if I've gotten
> the order wrong). Only one of these people is pushing for changing the
> central repository to git (GHellings).

Based on ohloh's statistics, Sword development exhibits "Moderate
activity", which puts us in the top 3.8%-1.2% of projects they track.

Top committers (in the sword repository ONLY--which is not necessarily a
representative statistic for everyone), who are still active in the last
12 months are, in order: Troy, me, Greg, Manfred, DM, & Peter.

Commit stats:
https://www.ohloh.net/p/sword/contributors?sort=latest_commit



FWIW, I do in fact support a move to git. I'm just not religious about
it. I see it as inevitable. I see it as a positive move for CrossWire. I
don't personally like *using* git all that much, but that's largely
because I use it seldom & don't know it all that well. But I'll
learn--all of us who don't know git that well will learn.

--Chris

[David &quot;Judahs Shadow&quot; Blue]

On Thursday, February 27, 2014 12:32:31 PM Troy A. Griffitts wrote:
> 4) We will probably change when I start using git personally for my own
> SWORD work because I just can't live without its killer features.

Speaking of features of git. I'm shocked no one has mentioned the bisect
commend yet. From what I understand, it helps you determine which commit broke
things.

[Matěj Cepl]

On Wed, 2014-03-05 at 20:04 -0500, David "Judahs Shadow" Blue wrote:
> Speaking of features of git. I'm shocked no one has mentioned the bisect
> commend yet. From what I understand, it helps you determine which commit broke
> things.

There are many advantages of git over SVN, too many to enumerate here.
However, I don't that's the issue. Everybody involved admit that git is
better, faster, more popular etc., but there is not enough time/will to
do the switch on side of people who can do it (yes, and another
advantage of DVCS is that there would be a less single-point of failure
situations).

Mat?j

--
http://www.ceplovi.cz/matej/, Jabber: mcepl at ceplovi.cz
GPG Finger: 89EF 4BC6 288A BF43 1BAB 25C3 E09F EF25 D964 84AC

This message has been composed of recycled electrons. None of
these electrons has been harmed or injured in the creation and
transmission of this message but they have been shamelessly
exploited for this use.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: This is a digitally signed message part
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140306/7069d998/attachment.sig>

[Baiju M]

Using git-svn, it is possible to use git as a client for svn:
http://blogs.atlassian.com/2013/12/git-svn-tips-and-tricks/

Regards,
Baiju M
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140313/79705bee/attachment.html>

[Caleb Maclennan]

As I understand it the issue here is that most SWORD stuff is in CVS, not
SVN and it's a little harder to use git as a front end client for CVS than
it is for SVN. It can be done (see http://stackoverflow.com/a/586225/313192),
but isn't easy. There is a high barrier to entry even for those familiar
with _both_ systems.

I know the subject is being beaten to death lately but to add one more
voice: there are those of us out here just waiting to contribute and
participate when the toolset becomes less arcane. Non-distributed version
control simply makes it too hard and too frustrating to experiment and then
cleanup those experiments for contribution. The result is fewer
contributions and less quality review and cooperation on the ones that do
happen.

I have quite a bit of git know-how including migrations from other VCS
systems and would be willing to help anybody on the SWORD team with
migration issues or just coming up to speed with DVS in general. I also
have experience hosting git servers including Gitlab if they would like to
run something like that internally to coordinate (and control)
contributions. In fact I would be willing to donate hosting a Gitlab
instance if that's what it takes. This would allow fairly fine grained
control over repos and even branches, specifying who is and isn't allowed
to push to which ones. At the same time as keeping a tight reign on the
quality of the master fork and branch, folks could also freely have their
own copies of the repo and easily submit pull requests for for review
similar to how the Github workflow functions.

Feel free to contact me on or off list about any of the above and know that
there will be much rejoicing out here when migration does happen!

Caleb


On Thu, Mar 13, 2014 at 8:34 AM, Baiju M <baiju.m.mail at gmail.com> wrote:

> Using git-svn, it is possible to use git as a client for svn:
> http://blogs.atlassian.com/2013/12/git-svn-tips-and-tricks/
>
> Regards,
> Baiju M
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140319/a833fcf0/attachment.html>

[Peter Von Kaehne]

Von:?"Caleb Maclennan" <caleb at alerque.com>

> As I understand it the issue here is that most SWORD stuff is in CVS, not SVN
?

That is a misunderstanding. For historical reasons our source commit mailing list is still called sword-cvs, but we use svn for many many years.

Peter

[Caleb Maclennan]

On Wed, Mar 19, 2014 at 6:59 PM, Peter Von Kaehne <refdoc at gmx.net> wrote:

> That is a misunderstanding. For historical reasons our source commit
> mailing list is still called sword-cvs, but we use svn for many many years.


My mistake. The last time I mucked around with it it was CVS and recent
discussions have included indications that some parts still were. In that
event a transition to DVCS will be more gain for less pain.

Caleb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140320/8ffbca68/attachment.html>

[Peter von Kaehne]

On Thu, 2014-03-20 at 22:45 +0200, Caleb Maclennan wrote:
> On Wed, Mar 19, 2014 at 6:59 PM, Peter Von Kaehne <refdoc at gmx.net>
> wrote:
> That is a misunderstanding. For historical reasons our source
> commit mailing list is still called sword-cvs, but we use svn
> for many many years.
>
> My mistake. The last time I mucked around with it [..]

That would have been before 2006 then.

[rj_qgsous_sword]

I saw recently that there has been a good deal of conversation on
this mailing list about Git. Personally, I think that Mercurial (Hg)
is much better than Git, and I wanted to mention a few things
regarding this:



Although (as someone recently mentioned) Mozilla Firefox OS
switched from using Hg to Git, Mozilla still has a quite active (4
repositories changed so far today) Hg source code repository at &lt;http://hg.mozilla.org&gt;.
When I was considering whether to use Git or Hg for my own
single-person project, I was overwhelmed by the complexity of Git, with
139
commands in version 1.5.0. A quick look at screenshots of
TortoiseGit's context menu versus TortoiseHg's will show this
difference as well. While Git may be faster than Hg for certain
operations on Linux, it has a reputation for being difficult to
learn, and its Windows support (as of 2010) was
reported to be much worse than Hg's. (Some of this information
is from a book that comes with TortoiseHg, Mercurial: The Definitive Guide, which might appear to be biased,
but its comparison of different VCSs seems objective to me.)
This book also mentions that Git repositories require frequent
"repacks" of metadata, which can decrease performance, take
large amounts of disk space and server bandwidth, and take many
hours to complete for large repositories. It goes on to say that
Git uses shell or Perl scripts to implement several of its
commands, which can ignore certain errors that should be fatal.




I have been using Mercurial (through TortoiseHg) for 6 months now
and am very pleased with it. (I used TortoiseSVN before that, and
had some frustrations with it.) Recently I found that a file in my
repository had somehow gotten corrupted, but after reading the
page on Mercurial's website about repository corruption, I was
able to successfully recover my repository.




Since switching to a different VCS can be quite an ordeal, I
wanted to mention these things. I admit that I am not very
knowledgeable about VCSs, so please correct me if I am wrong about
anything.




-- Timothy

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140315/2b53eaf3/attachment.html>

[Matej Cepl]

On Sat, Mar 15, 2014 at 07:25:15PM +0100, rj_qgsous_sword wrote:
> and its Windows support (as of 2010) was reported to be much
> worse than Hg's.

Let me emphasize ?as of 2010? ? I don?t use Windows myself (see
the email address), but I?ve heard that http://msysgit.github.io/
is quite useful.

Best,

Mat?j

--
http://www.ceplovi.cz/matej/, Jabber: mcepl<at>ceplovi.cz
GPG Finger: 89EF 4BC6 288A BF43 1BAB 25C3 E09F EF25 D964 84AC

They that can give up essential liberty to obtain a little
temporary safety deserve neither liberty nor safety.
-- Benjamin Franklin, Historical Review
of Pennsylvania, 1759.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 190 bytes
Desc: not available
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140316/b76687eb/attachment.sig>

[Ryan Hiebert]

On Sun, Mar 16, 2014 at 4:14 PM, Matej Cepl <mcepl at redhat.com> wrote:

> On Sat, Mar 15, 2014 at 07:25:15PM +0100, rj_qgsous_sword wrote:
>
>> and its Windows support (as of 2010) was reported to be much worse than
>> Hg's.
>>
>
> Let me emphasize "as of 2010" ... I don't use Windows myself (see the email
> address), but I've heard that http://msysgit.github.io/
> is quite useful.
>
> A quick comment to address this: Windows support for git with is quite
good. It was with msysgit, and now there is an official distribution from
git-scm (for a while now actually). Git is even being integrated into
Visual Studio and Team Foundation Server.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20140316/660f1367/attachment.html>

[Dimitri John Ledkov]

On 15 March 2014 18:25, rj_qgsous_sword <rj_qgsous_sword at objectmail.com> wrote:
> I saw recently that there has been a good deal of conversation on this
> mailing list about Git. Personally, I think that Mercurial (Hg) is much
> better than Git, and I wanted to mention a few things regarding this:
>
> Although (as someone recently mentioned) Mozilla Firefox OS switched from
> using Hg to Git, Mozilla still has a quite active (4 repositories changed so
> far today) Hg source code repository at <http://hg.mozilla.org>.
>
> When I was considering whether to use Git or Hg for my own single-person
> project, I was overwhelmed by the complexity of Git, with 139 commands in
> version 1.5.0.. A quick look at screenshots of TortoiseGit's context menu
> versus TortoiseHg's will show this difference as well. While Git may be
> faster than Hg for certain operations on Linux, it has a reputation for
> being difficult to learn, and its Windows support (as of 2010) was reported
> to be much worse than Hg's. (Some of this information is from a book that
> comes with TortoiseHg, Mercurial: The Definitive Guide, which might appear
> to be biased, but its comparison of different VCSs seems objective to me.)
> This book also mentions that Git repositories require frequent "repacks" of
> metadata, which can decrease performance, take large amounts of disk space
> and server bandwidth, and take many hours to complete for large
> repositories. It goes on to say that Git uses shell or Perl scripts to
> implement several of its commands, which can ignore certain errors that
> should be fatal.
>
> I have been using Mercurial (through TortoiseHg) for 6 months now and am
> very pleased with it. (I used TortoiseSVN before that, and had some
> frustrations with it.) Recently I found that a file in my repository had
> somehow gotten corrupted, but after reading the page on Mercurial's website
> about repository corruption, I was able to successfully recover my
> repository.
>
> Since switching to a different VCS can be quite an ordeal, I wanted to
> mention these things. I admit that I am not very knowledgeable about VCSs,
> so please correct me if I am wrong about anything.
>


git storage format is better than Mercurial's.
git UI/GUI are constantly improving.
msysgit / git cheetah (Tortoise-like plugin) are very performant.
Which i've facilitate in deployments to large and mostly Windows based
developers.
git has more traction behind it / more people and potential developers
know/use git.

If you preffer hg client, you can use http://hg-git.github.io/ against
git repositories without anybody finding out =)

--
Regards,

Dimitri.