An authors-transform.txt file for converting to git

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

An authors-transform.txt file for converting to git

Karl O. Pinc
Hi,

Converting from svn to git requires an author
mapping.  Somebody besides me will have to
fill out the remaining entries.

An authors-transform.txt file can be found at:

  https://gist.github.com/kpinc/9139445

Note also that "no author" sounds like it
could someday cause legal copyright related
headaches.


Meanwhile, I'm going to see about coming up
with a conversion process and will post the result
to this list.


Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Hilmar Lapp-3
I've made some changes. Unfortunately, I can't see how to create a pull or patch request from these, so I'm resorting to emailing you the link:


  -hilmar


On Fri, Feb 21, 2014 at 12:57 PM, Karl O. Pinc <[hidden email]> wrote:
Hi,

Converting from svn to git requires an author
mapping.  Somebody besides me will have to
fill out the remaining entries.

An authors-transform.txt file can be found at:

  https://gist.github.com/kpinc/9139445

Note also that "no author" sounds like it
could someday cause legal copyright related
headaches.


Meanwhile, I'm going to see about coming up
with a conversion process and will post the result
to this list.


Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema



--
Hilmar Lapp -:- lappland.io


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
Hi Hilmar,

Thanks.

On 02/22/2014 09:50:02 AM, Hilmar Lapp wrote:
> I've made some changes. Unfortunately, I can't see how to create a
> pull or
> patch request from these, so I'm resorting to emailing you the link:

I don't know how to do it either.  I can clone the repo
and push patches.  (Maybe anytbody can and there's
no authorization because it's a gist?)

> https://gist.github.com/hlapp/9156786

As far as I can tell there's no way to frob the
revision history and correct author names without
re-writing history (vis. git filter-branch or
some such).  Since this rewrites history it'd
be an awful pain -- better to get the author
list right as part of the conversion process.

But there seems to be a lot of authors
that need fixing.  Has there been any work
done in this direction already as part of
some prior git conversion attempt?



Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Hilmar Lapp-3

On Sat, Feb 22, 2014 at 3:18 PM, Karl O. Pinc <[hidden email]> wrote:
I can clone the repo and push patches.  (Maybe anytbody can and there's
no authorization because it's a gist?)

I can't push to your gist, so there does seem to be authorization. Maybe that means you will need to copy over the changes yourself? (short of me emailing you a patch)
 
As far as I can tell there's no way to frob the revision history and correct author names without re-writing history (vis. git filter-branch or some such).

I didn't mean for you to rewrite history. Normally you do an initial migration just to obtain the user IDs that need mapping. Once you've gathered all (or all for which someone still knows the email), you rerun the migration.

   -hilmar

--
Hilmar Lapp -:- lappland.io


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
On 02/22/2014 04:27:40 PM, Hilmar Lapp wrote:

> On Sat, Feb 22, 2014 at 3:18 PM, Karl O. Pinc <[hidden email]> wrote:
>
> > I can clone the repo and push patches.  (Maybe anytbody can and
> there's
> > no authorization because it's a gist?)
> >
>
> I can't push to your gist, so there does seem to be authorization.
> Maybe
> that means you will need to copy over the changes yourself? (short of
> me
> emailing you a patch)

I bet my new gist will have a new url anyway, so people may as
well work from your url from this point.

By the by, I got my author mapping by grepping
my archive of the gmod-schema
list.  However my list archive is very
limited.  If anybody has
a complete copy of the list archive this
would provide a pretty good path forward
to a complete author mapping.  (What about
the "no author" ones?  Could those be the
db dumps?)

>
>
> > As far as I can tell there's no way to frob the revision history
> and
> > correct author names without re-writing history (vis. git
> filter-branch
> > or some such).
>
>
> I didn't mean for you to rewrite history. Normally you do an initial
> migration just to obtain the user IDs that need mapping. Once you've
> gathered all (or all for which someone still knows the email), you
> rerun
> the migration.

Yes.  But re-running the migration changes all the git hashes,
so any changes made to the original git repo will map on to the
new HEAD.  Or something.  It's like the whole git repo's
been rebased.  Not so bad if no changes were made to the
svn repo, between the 2 migrations, but not something that
works well over any period of time.  Better to get the
author mapping right, move to git, and be done with it.


Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Scott Cain
If somebody wouldn't mind making the change, my github email address is [hidden email]

Thanks,
Scott

Sent from my iPhone

> On Feb 22, 2014, at 6:36 PM, "Karl O. Pinc" <[hidden email]> wrote:
>
>> On 02/22/2014 04:27:40 PM, Hilmar Lapp wrote:
>>> On Sat, Feb 22, 2014 at 3:18 PM, Karl O. Pinc <[hidden email]> wrote:
>>>
>>> I can clone the repo and push patches.  (Maybe anytbody can and
>> there's
>>> no authorization because it's a gist?)
>>
>> I can't push to your gist, so there does seem to be authorization.
>> Maybe
>> that means you will need to copy over the changes yourself? (short of
>> me
>> emailing you a patch)
>
> I bet my new gist will have a new url anyway, so people may as
> well work from your url from this point.
>
> By the by, I got my author mapping by grepping
> my archive of the gmod-schema
> list.  However my list archive is very
> limited.  If anybody has
> a complete copy of the list archive this
> would provide a pretty good path forward
> to a complete author mapping.  (What about
> the "no author" ones?  Could those be the
> db dumps?)
>
>>
>>
>>> As far as I can tell there's no way to frob the revision history
>> and
>>> correct author names without re-writing history (vis. git
>> filter-branch
>>> or some such).
>>
>>
>> I didn't mean for you to rewrite history. Normally you do an initial
>> migration just to obtain the user IDs that need mapping. Once you've
>> gathered all (or all for which someone still knows the email), you
>> rerun
>> the migration.
>
> Yes.  But re-running the migration changes all the git hashes,
> so any changes made to the original git repo will map on to the
> new HEAD.  Or something.  It's like the whole git repo's
> been rebased.  Not so bad if no changes were made to the
> svn repo, between the 2 migrations, but not something that
> works well over any period of time.  Better to get the
> author mapping right, move to git, and be done with it.
>
>
> Karl <[hidden email]>
> Free Software:  "You don't pay back, you pay forward."
>                 -- Robert A. Heinlein
>
> ------------------------------------------------------------------------------
> Managing the Performance of Cloud-Based Applications
> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
> Read the Whitepaper.
> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
> _______________________________________________
> Gmod-schema mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-schema

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
On 02/22/2014 06:07:07 PM, Scott Cain wrote:
> If somebody wouldn't mind making the change, my github email address
> is [hidden email]

Are you sure?  This is about being credited as a
chado patch author and has nothing to do with
authentication, authorization, or ancillary
github related information.  Anybody with
any email address can have patches pushed to
github by authorized committers and
that's the sort of email address being collected.

Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Hilmar Lapp-3
In reply to this post by Karl O. Pinc

On Sat, Feb 22, 2014 at 6:36 PM, Karl O. Pinc <[hidden email]> wrote:
Yes.  But re-running the migration changes all the git hashes, so any changes made to the original git repo will map on to the new HEAD.

Perhaps I was missing something. I meant you don't actually publish the initial migration. Or if you do, it is only temporary; you trash it before the "real" migration.

So, no, there would be no change of hashes, rebase, or other things. Just as rerun of the migration once author map is complete, and the result of that becomes the blessed repo.

Am I making sense? Sorry if this is counter to some plan that was floated earlier and that I missed.

  -hilmar
--
Hilmar Lapp -:- lappland.io


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Hilmar Lapp-3
In reply to this post by Scott Cain

On Sat, Feb 22, 2014 at 7:07 PM, Scott Cain <[hidden email]> wrote:
If somebody wouldn't mind making the change, my github email address is [hidden email]

This should be unnecessary. Any email you have registered with your github account will be recognized as mapping to your account. If the one in the mapping isn't, you can just register it.

So the only thing to worry about is that the author map doesn't use inactive email addresses for people who are still active in some shape or form, because registering an email address on Github requires you to verify it, which you can't if it doesn't exist anymore. 

  -hilmar

--
Hilmar Lapp -:- lappland.io


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
In reply to this post by Hilmar Lapp-3
On 02/23/2014 11:14:17 AM, Hilmar Lapp wrote:

> On Sat, Feb 22, 2014 at 6:36 PM, Karl O. Pinc <[hidden email]> wrote:
>
> > Yes.  But re-running the migration changes all the git hashes, so
> any
> > changes made to the original git repo will map on to the new HEAD.
> >
>
> Perhaps I was missing something. I meant you don't actually publish
> the
> initial migration. Or if you do, it is only temporary; you trash it
> before
> the "real" migration.
>
> So, no, there would be no change of hashes, rebase, or other things.
> Just
> as rerun of the migration once author map is complete, and the result
> of
> that becomes the blessed repo.
>
> Am I making sense? Sorry if this is counter to some plan that was
> floated
> earlier and that I missed.

This makes plenty of sense.  The issue is more at my end.

After having converted to git I'd like to immediately
start keeping my patches in git in order to
keep track of things.  I'm only really justified
in spending time at this if it's going to help
me with my chado work.

Re-converting later throws a kink into my plans.



Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
In reply to this post by Hilmar Lapp-3
On 02/23/2014 11:18:20 AM, Hilmar Lapp wrote:

> So the only thing to worry about is that the author map doesn't use
> inactive email addresses for people who are still active in some
> shape
> or
> form, because registering an email address on Github requires you to
> verify
> it, which you can't if it doesn't exist anymore.

I don't believe we need to worry about this either.

The email addresses that are associated with
git patches have nothing to do with github.  We'd like
email addresses in the git change log that reflect
who made the change.  The addresses may not even
have to be current or unique per individual.  If an individual
has changed jobs and had different email addresses
over time then the address associated with different
patches can reflect that, even if some of the
addresses are stale.  

The utility of having stale email addresses in the
log is in the tracking of ownership and granting of
copyright.  For example, some institutions,
not only businesses but also public Universities
like the University of Illinois, require
employees (and students) to sign over ownership
of all "intellectual property" to the institution.
Therefore should you ever need to show that
Chado is truly licensed as advertised it
may be necessary to contact not only the
individual authors but also the institution
with which they were associated at the
time each patch was written.

And, of course, it's nice to attribute people's
work.

While we could probably get away with mapping
all the svn author data to emails associated
with, e.g., the example.com domain it'd be
a lame git conversion that did so.

To reiterate, the email addresses on file
with github have to do with supporting
github processes, pull requests, etc., etc.,
and have nothing to do with the mapping
we are attempting to create between the patch
author data logged in git and and the patch
author data we want logged in git.

Regards,

Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Fields, Christopher J
In reply to this post by Hilmar Lapp-3
On Feb 23, 2014, at 11:18 AM, Hilmar Lapp <[hidden email]> wrote:


On Sat, Feb 22, 2014 at 7:07 PM, Scott Cain <[hidden email]> wrote:
If somebody wouldn't mind making the change, my github email address is [hidden email]

This should be unnecessary. Any email you have registered with your github account will be recognized as mapping to your account. If the one in the mapping isn't, you can just register it.

So the only thing to worry about is that the author map doesn't use inactive email addresses for people who are still active in some shape or form, because registering an email address on Github requires you to verify it, which you can't if it doesn't exist anymore. 

  -hilmar

Yep, we ran into this a bit w/ the bioperl migration.  There are several authors who aren’t mapped correctly.  This can be fixed but requires git filter-branch IIRC.

chris

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Hilmar Lapp-3
In reply to this post by Karl O. Pinc
I don't think there's anything that prevents you from that. You migrate, this gives you a git repo (which may or may not push up to a remote host), and then you branch and apply your changes.

Since a later final migration won't create any code  changes, you can then just add that as a remote and cherry-pick your changes over.

That's just off the top of my head, there may even be better ways. But the point is, you don't need to wait on some final migration having happened.

 -hilmar 


On Sun, Feb 23, 2014 at 2:19 PM, Karl O. Pinc <[hidden email]> wrote:
On 02/23/2014 11:14:17 AM, Hilmar Lapp wrote:
> On Sat, Feb 22, 2014 at 6:36 PM, Karl O. Pinc <[hidden email]> wrote:
>
> > Yes.  But re-running the migration changes all the git hashes, so
> any
> > changes made to the original git repo will map on to the new HEAD.
> >
>
> Perhaps I was missing something. I meant you don't actually publish
> the
> initial migration. Or if you do, it is only temporary; you trash it
> before
> the "real" migration.
>
> So, no, there would be no change of hashes, rebase, or other things.
> Just
> as rerun of the migration once author map is complete, and the result
> of
> that becomes the blessed repo.
>
> Am I making sense? Sorry if this is counter to some plan that was
> floated
> earlier and that I missed.

This makes plenty of sense.  The issue is more at my end.

After having converted to git I'd like to immediately
start keeping my patches in git in order to
keep track of things.  I'm only really justified
in spending time at this if it's going to help
me with my chado work.

Re-converting later throws a kink into my plans.



Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein



--
Hilmar Lapp -:- lappland.io


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
In reply to this post by Fields, Christopher J
On 02/23/2014 01:41:15 PM, Fields, Christopher J wrote:
> On Feb 23, 2014, at 11:18 AM, Hilmar Lapp
> <[hidden email]<mailto:[hidden email]>> wrote:

> So the only thing to worry about is that the author map doesn't use
> inactive email addresses for people who are still active in some
> shape
> or form, because registering an email address on Github requires you
> to verify it, which you can't if it doesn't exist anymore.
>
>   -hilmar
>
> Yep, we ran into this a bit w/ the bioperl migration.  There are
> several authors who aren’t mapped correctly.  This can be fixed but
> requires git filter-branch IIRC.
>
> chris


The problem with git filter-branch is that it basically rebases
the whole repo, requiring anybody who's cloned the repo (and
anybody who's in turn cloned those repos, etc.) to rebase
also.

In other words, fixing after the fact can be very painful.
It's probably easier to commit a document that describes
the corrections that should be made and let anybody who
really cares do the remapping themselves.


Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
In reply to this post by Hilmar Lapp-3
On 02/23/2014 01:52:52 PM, Hilmar Lapp wrote:

> I don't think there's anything that prevents you from that. You
> migrate,
> this gives you a git repo (which may or may not push up to a remote
> host),
> and then you branch and apply your changes.
>
> Since a later final migration won't create any code  changes, you can
> then
> just add that as a remote and cherry-pick your changes over.
>
> That's just off the top of my head, there may even be better ways.
> But
> the
> point is, you don't need to wait on some final migration having
> happened.


I'm hoping that this is the case.  But your assumption that
a later migration won't create any code changes
is surly false; if enough time passes there will
be more svn commits.



Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Fields, Christopher J
In reply to this post by Karl O. Pinc
On Feb 23, 2014, at 2:06 PM, Karl O. Pinc <[hidden email]> wrote:

> On 02/23/2014 01:41:15 PM, Fields, Christopher J wrote:
>> On Feb 23, 2014, at 11:18 AM, Hilmar Lapp
>> <[hidden email]<mailto:[hidden email]>> wrote:
>
>> So the only thing to worry about is that the author map doesn't use
>> inactive email addresses for people who are still active in some
>> shape
>> or form, because registering an email address on Github requires you
>> to verify it, which you can't if it doesn't exist anymore.
>>
>>  -hilmar
>>
>> Yep, we ran into this a bit w/ the bioperl migration.  There are
>> several authors who aren’t mapped correctly.  This can be fixed but
>> requires git filter-branch IIRC.
>>
>> chris
>
>
> The problem with git filter-branch is that it basically rebases
> the whole repo, requiring anybody who's cloned the repo (and
> anybody who's in turn cloned those repos, etc.) to rebase
> also.
>
> In other words, fixing after the fact can be very painful.
> It's probably easier to commit a document that describes
> the corrections that should be made and let anybody who
> really cares do the remapping themselves.
>
>
> Karl <[hidden email]>

Frankly, we didn’t even bother with additional filter-branch correction. Only one instance I recall where it was a problem, and it was corrected by simply having the developer add the email to their github account.  It’s nice if you can get everyone’s updates, but most contribs don’t even seem to really notice the nice link to their account from the commits.

chris


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema
Reply | Threaded
Open this post in threaded view
|

Re: An authors-transform.txt file for converting to git

Karl O. Pinc
On 02/23/2014 02:44:01 PM, Fields, Christopher J wrote:
> On Feb 23, 2014, at 2:06 PM, Karl O. Pinc <[hidden email]> wrote:
> > On 02/23/2014 01:41:15 PM, Fields, Christopher J wrote:

> >> Yep, we ran into this a bit w/ the bioperl migration.  There are
> >> several authors who aren’t mapped correctly.

> > In other words, fixing after the fact can be very painful.

> Frankly, we didn’t even bother with additional filter-branch
> correction. Only one instance I recall where it was a problem, and it
> was corrected by simply having the developer add the email to their
> github account.  It’s nice if you can get everyone’s updates, but
> most
> contribs don’t even seem to really notice the nice link to their
> account from the commits.

While crediting contributors is nice
I see lawyers as the real problem.
Who would have thought that, e.g., SCO would
claim to own the rights to Linux and spend
10s of millions of dollars trying to prove it,
in the process tracing the provenance of every
line of code back to an author?  Without
a record of who did what you're a lot more
vulnerable.  No doubt this influenced the
design of git's patch attribution and sign-off
features.  

I get the feeling that some simple
processes, like patch sign-off and, e.g.,
the PostgreSQL contributor agreement,
provides a reasonable level of
confidence in defense against legal
harpies.  But if you ignore the issue
you might someday find others claiming
ownership of chado and have problems
defending your claim.  

I'm reminded
of the java model railroad interface
and claims thereto, although I don't know
that this case speaks directly to
author attribution issues since I think
the problem there was that the Artistic
license is not well crafted.
http://en.wikipedia.org/wiki/Jacobsen_v._Katzer
My point being, I guess, is that if
you look weak you can be targeted.
Especially if real money is
at stake, and who knows when that
might be the case?

The issue raised is: why attribute
patch authorship?  I'd like to clearly
point out that there's more to the answer
than the crediting of patch-fu.

Regards,

Karl <[hidden email]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

P.S. Of course there's always craziness
that can't be avoided, like Oracle's claim
against Google where Oracle claimed to
own any language that complied with
the Java specifications.
------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Gmod-schema mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-schema