Google Code Issue migration

[Ivan Zhakov]

I'm currently working on JSON import data to migration issues from
Google Code to ASF hosted JIRA. Relevant ASF Infra ticket:
https://issues.apache.org/jira/browse/INFRA-10186

Since ASF hosted JIRA uses different user accounts than Google Code
users from Google Code should be mapped to ASF JIRA accounts.

So please provide me pair of your Google Code user name and ASF JIRA
account. You may signup for new user account if you don't have it
already:
https://issues.apache.org/jira/secure/Signup!default.jspa

All other user names will be converted as 'serf-importer'.

--
Ivan Zhakov

[C. Michael Pilato]

I should be 'cmpilato' in both universes, I think.

--
Ivan Zhakov

--
You received this message because you are subscribed to the Google Groups "Serf Development List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to serf-dev+u...@googlegroups.com.
To post to this group, send email to serf...@googlegroups.com.
Visit this group at http://groups.google.com/group/serf-dev.
For more options, visit https://groups.google.com/d/optout.

[Bert Huijben]

I used the following hack to transform the usernames for our repository import.
[[
case $author in
gstein)
continue
;;
jerenkrantz)
continue
;;
ivan)
continue
;;

bhuijben)
new_author=rhuijben
;;

b.huijben)
new_author=rhuijben
;;

chemodax)
new_author=ivan
;;

justin.erenkrantz)
new_author=jerenkrantz
;;

cmpi...@gmail.com)
new_author=cmpilato
;;

danielrall)
new_author=
;;

lieven.govaerts)
new_author=lgo
;;

paul.querna)
new_author=pquerna
;;

paul....@gmail.com)
new_author=pquerna
;;

rooneg)
continue
;;

roy.fielding)
new_author=fielding
;;

st...@elego.de)
new_author=stsp
;;

andreas...@gmx.de)
new_author=astieger
;;

breser)
continue
;;

bre...@gmail.com)
new_author=breser
;;

lieven.govaerts)
new_author=lgo
;;

lieven....@gmail.com)
new_author=lgo
;;

paul.querna)
new_author=pquerna
;;

paul....@gmail.com)
new_author=pquerna
;;

rooneg)
continue;;

roy.fielding)
new_author=fielding
;;
]]

Bert

[Ivan Zhakov]

On 25 August 2015 at 18:07, Bert Huijben <be...@qqmail.nl> wrote:
>
>
>> -----Original Message-----
>> From: serf...@googlegroups.com [mailto:serf...@googlegroups.com] On
>> Behalf Of Ivan Zhakov
>> Sent: dinsdag 25 augustus 2015 16:05
>> To: serf...@googlegroups.com; d...@serf.apache.org
>> Subject: [serf-dev] Google Code Issue migration
>>
>> I'm currently working on JSON import data to migration issues from
>> Google Code to ASF hosted JIRA. Relevant ASF Infra ticket:
>> https://issues.apache.org/jira/browse/INFRA-10186
>>
>> Since ASF hosted JIRA uses different user accounts than Google Code
>> users from Google Code should be mapped to ASF JIRA accounts.
>>
>> So please provide me pair of your Google Code user name and ASF JIRA
>> account. You may signup for new user account if you don't have it
>> already:
>> https://issues.apache.org/jira/secure/Signup!default.jspa
>>
>> All other user names will be converted as 'serf-importer'.
>
> I used the following hack to transform the usernames for our repository import.

ASF JIRA has different accounts than ASF accounts. For example my ASF
nickname (ivan) is already occupied so I've registered as "ivan.z"


--
Ivan Zhakov

[Lieven Govaerts]

Hi Ivan,

My Jira name = "lgo"

A question: is it possible to make an easy mapping between google code
issue number and Jira issue number? So that we can rewrite svn log
messages / code comments easily (if we want to) ?

thanks, Lieven

> --
> Ivan Zhakov
>

[Ivan Zhakov]

On 25 August 2015 at 20:39, Lieven Govaerts <l...@mobsol.be> wrote:
> Hi Ivan,
>
> On Tue, Aug 25, 2015 at 4:04 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:
>> I'm currently working on JSON import data to migration issues from
>> Google Code to ASF hosted JIRA. Relevant ASF Infra ticket:
>> https://issues.apache.org/jira/browse/INFRA-10186
>>
>> Since ASF hosted JIRA uses different user accounts than Google Code
>> users from Google Code should be mapped to ASF JIRA accounts.
>>
>> So please provide me pair of your Google Code user name and ASF JIRA
>> account. You may signup for new user account if you don't have it
>> already:
>> https://issues.apache.org/jira/secure/Signup!default.jspa
>>
>> All other user names will be converted as 'serf-importer'.
>>
>
> My Jira name = "lgo"
>

Thanks!

> A question: is it possible to make an easy mapping between google code
> issue number and Jira issue number? So that we can rewrite svn log
> messages / code comments easily (if we want to) ?
>

It should be possible to get the same issue numbers in imported data.


--
Ivan Zhakov
CTO | VisualSVN | http://www.visualsvn.com

[Bert Huijben]

My Jira name is rhuijben... But you probably already had that. Accounts on google code are probably b.huijben, bhuijben and be...@qqmail.nl.

Bert

---

From: Ivan Zhakov
Sent: ‎25-‎8-‎2015 21:29
To: serf...@googlegroups.com
Cc: d...@serf.apache.org
Subject: Re: [serf-dev] Google Code Issue migration

[Greg Stein]

gstein.

Thanks Ivan!

[Ivan Zhakov]

On 26 August 2015 at 03:52, Greg Stein <gst...@gmail.com> wrote:
> gstein.
>
Ack.

Btw it seems that you have another ASF JIRA "gstein at gmail.com".

> Thanks Ivan!
>
I've submitted test import data to ASF INFRA team yesterday.

--
Ivan Zhakov

[Bert Huijben]

> -----Original Message-----
> From: serf...@googlegroups.com [mailto:serf...@googlegroups.com] On
> Behalf Of Ivan Zhakov
> Sent: dinsdag 25 augustus 2015 16:05
> To: serf...@googlegroups.com; d...@serf.apache.org

For the final version:
Can you add the original name to the issues as part of the comment for those issues that are now from 'serf-importer'?

Bert

[Greg Stein]

On Wed, Aug 26, 2015 at 4:37 AM, Ivan Zhakov <iv...@visualsvn.com> wrote:

On 26 August 2015 at 03:52, Greg Stein <gst...@gmail.com> wrote:
> gstein.
>
Ack.

Btw it seems that you have another ASF JIRA "gstein at gmail.com".

username "gstein", with address "gstein@apache"

thx,

-g

[Ivan Zhakov]

It's should be possible. Do you suggest add original name only for
comments or also for first comment, that will be used as issue
description? What format do you propose: just add line before comment
like this "Original author: someone" ?

--
Ivan Zhakov

[Bert Huijben]

Something like that would be very useful... +1

And I think we should also do it for the first comment... Without that we would have to go back to some google code archive to find the original issue submitter.
(Although I would guess most of the issues are by the committers)

Bert

[Ivan Zhakov]

Ok, I'll try to implement this.

> (Although I would guess most of the issues are by the committers)
>

I also have an idea to add special custom field with link to original
issue in Google Code Archive. But we need ask INFRA team to add this
field for us. I'm going to do this after successful test migration.

--
Ivan Zhakov

[Greg Stein]

I get the impression that links to code.google.com ("c.g.c") will eventually fail. It may not be worth the effort, or the future failing links, to add a field pointing at c.g.c. 

If you retain all the information, in some way or another, then I don't think links will be needed. If JIRA records all label changes, all authors, all comments, etc … good enough.

(personally, I'm also okay with a loss of fidelity, as long as we can contact the person who originally filed the issue)

Cheers,

-g

[Ivan Zhakov]

On 26 August 2015 at 15:30, Greg Stein <gst...@gmail.com> wrote:
[..]

>>
>> I also have an idea to add special custom field with link to original
>> issue in Google Code Archive. But we need ask INFRA team to add this
>> field for us. I'm going to do this after successful test migration.
>
> I get the impression that links to code.google.com ("c.g.c") will eventually
> fail. It may not be worth the effort, or the future failing links, to add a
> field pointing at c.g.c.
>

I was thinking to add links to Google Code Archive:
https://code.google.com/archive/p/serf/
https://code.google.com/p/support/wiki/ReadOnlyTransition

> If you retain all the information, in some way or another, then I don't
> think links will be needed. If JIRA records all label changes, all authors,
> all comments, etc … good enough.
>

I'm not sure that I'll be able preserve all label changes except
comments, due slightly different data model of Google Code Issue
tracker and ASF JIRA. Current implementation preserve summary,
description, reporter, status and comments.

> (personally, I'm also okay with a loss of fidelity, as long as we can
> contact the person who originally filed the issue)
>

--
Ivan Zhakov

[Bert Huijben]

Google doesn't promise to keep this infrastructure up in the same form for more than a few months, so I'm not entirely sure if that is worth the effort.

The current 'archive format' doesn't even use the last site content :(
So https://code.google.com/archive/p/serf/ doesn't point to the Apache site (yet?)

Bert

[Greg Stein]

On Wed, Aug 26, 2015 at 7:39 AM, Ivan Zhakov <iv...@visualsvn.com> wrote:

On 26 August 2015 at 15:30, Greg Stein <gst...@gmail.com> wrote:
[..]
>>
>> I also have an idea to add special custom field with link to original
>> issue in Google Code Archive. But we need ask INFRA team to add this
>> field for us. I'm going to do this after successful test migration.
>
> I get the impression that links to code.google.com ("c.g.c") will eventually
> fail. It may not be worth the effort, or the future failing links, to add a
> field pointing at c.g.c.
>
I was thinking to add links to Google Code Archive:
https://code.google.com/archive/p/serf/
https://code.google.com/p/support/wiki/ReadOnlyTransition

Up to you... Can't hurt to drop those in, even if they won't live forever (as Bert notes). As issues get closed, the (potentially) broken links will not matter. And the more data you retain in the transfer, the less those links will be needed.

 

> If you retain all the information, in some way or another, then I don't
> think links will be needed. If JIRA records all label changes, all authors,
> all comments, etc … good enough.
>
I'm not sure that I'll be able preserve all label changes except
comments, due slightly different data model of Google Code Issue
tracker and ASF JIRA. Current implementation preserve summary,
description, reporter, status and comments.

Sure. But maybe those label changes can move into the comment text.

Take a look at the c.g.c -> github transfer of an issue:

  https://github.com/gstein/ezt/issues/7

Hopefully, that will give you some ideas on ways to port the content.

Thanks!!

-g

[Bert Huijben]

Ivan,

An initial (test?) import has been completed. See https://issues.apache.org/jira/browse/SERF/

But do note
[[
Gavin commented on INFRA-10186:
-------------------------------

I've made a start on this, some issues this end to go through, matching workflows etc.

So please do not do anything with the SERF Jira until all clear given :)
]]

Bert

[Ivan Zhakov]

On 26 August 2015 at 18:48, Bert Huijben <be...@qqmail.nl> wrote:
> Ivan,
>
> An initial (test?) import has been completed. See https://issues.apache.org/jira/browse/SERF/
>

I see. It's only test import. I've already installed test JIRA on VM
and improved my conversion program.

The first question what workflow we're going to use. Google Code we
have the following statuses:
[[
New
Fixed
WontFix
Accepted
Started
Invalid
Verified
Done
Duplicate
]]

While JIRA is very flexible and could adapted to any workflow I'm
inclined to choose simple way and map all these statuses to "OPEN" and
"RESOLVED":
[[
case "New":
return "OPEN";
case "Fixed":
return "CLOSED";
case "WontFix":
return "CLOSED";
case "Accepted":
return "OPEN";
case "Started":
return "OPEN";
case "Invalid":
return "CLOSED";
case "Verified":
return "CLOSED";
case "Done":
return "CLOSED";
case "Duplicate":
return "CLOSED";
]]]

Please note that CLOSED issues cannot be edited.

With the different "resolutions":
[[[
case "Fixed": return "Fixed";
case "Invalid": return "Invalid";
case "WontFix": return "Won't Fix";
case "Duplicate": return "Duplicate";
case "New":
case "Started":
case "Accepted":
return "";
case "REMIND": return "REMIND";
case "Verified":
return "Fixed";
case "Done":
return "Fixed";
]]]

Does it make sense or we want to customize JIRA workflow to match Google Code?

--
Ivan Zhakov

[Justin Erenkrantz]

Vote for sticking with standard workflow.  -- justin

[Ivan Zhakov]

On 26 August 2015 at 19:22, Justin Erenkrantz

<justin.e...@gmail.com> wrote:
> Vote for sticking with standard workflow. -- justin
>

It seems I was wrong and INFRA team was able to match these extended
statuses to ASF hosted JIRA workflow. For example:
https://issues.apache.org/jira/browse/SERF-173


--
Ivan Zhakov

[Greg Stein]

Yes, please go with the simpler/standard workflow. No customization.

(and don't make me rant about Google Code's status workflow; that pissed me off; 5 steps forward on awesome, but that was 2 steps back)

[Ivan Zhakov]

Status update:
ASF INFRA team performed test import for us:
https://issues.apache.org/jira/browse/SERF

The following data was converted:
1. Summary
2. Reporter and current assignee
3. Comments with original date and author
4. Labels
5. Issue status and resolution
6. Issue creation and last update date

The following data was *not* converted:
1. Attachments -- I didn't find easy way to download them. I'm also
not sure that JIRA supports importing attachments.
2. History of changes.

Things I'm going to do for final migration:
1. Add footer with original comment author for unmapped users.
2. Investigate ways to migrate issue history (source data is not easy to parse).

Bert suggested to remap revision numbers to match ASF repository
revisions. But there is a risk that conversion replace something else
that match rNNN pattern. So I'm inclined to leave them as-is.

--
Ivan Zhakov

[Justin Erenkrantz]

On Wed, Aug 26, 2015 at 2:52 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:
> Status update:
> ASF INFRA team performed test import for us:
> https://issues.apache.org/jira/browse/SERF

Looks good to me! Thanks all! -- justin

[Lieven Govaerts]

On Wed, Aug 26, 2015 at 8:52 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:
> Status update:
> ASF INFRA team performed test import for us:
> https://issues.apache.org/jira/browse/SERF
>
> The following data was converted:
> 1. Summary
> 2. Reporter and current assignee
> 3. Comments with original date and author
> 4. Labels
> 5. Issue status and resolution

Looks good.

I liked the difference between "New" and "Accepted" in the old repo,
is that still possible? Seems like "ACCEPTED" is available in Jira,
atleast in the filter.

> 6. Issue creation and last update date
>
> The following data was *not* converted:
> 1. Attachments -- I didn't find easy way to download them. I'm also
> not sure that JIRA supports importing attachments.
> 2. History of changes.
>
> Things I'm going to do for final migration:
> 1. Add footer with original comment author for unmapped users.
> 2. Investigate ways to migrate issue history (source data is not easy to parse).
>
> Bert suggested to remap revision numbers to match ASF repository
> revisions. But there is a risk that conversion replace something else
> that match rNNN pattern. So I'm inclined to leave them as-is.

You could add a "(asf repo: [rNNN+offset])" right behind the old
number. But I don't think it's that important, these revs are
typically used in issues that are resolved anyway.

Lieven

> --
> Ivan Zhakov
>

[Ivan Zhakov]

On 27 August 2015 at 22:52, Lieven Govaerts <l...@mobsol.be> wrote:
> On Wed, Aug 26, 2015 at 8:52 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:
>> Status update:
>> ASF INFRA team performed test import for us:
>> https://issues.apache.org/jira/browse/SERF
>>
>> The following data was converted:
>> 1. Summary
>> 2. Reporter and current assignee
>> 3. Comments with original date and author
>> 4. Labels
>> 5. Issue status and resolution
> Looks good.
>

Cool!

> I liked the difference between "New" and "Accepted" in the old repo,
> is that still possible? Seems like "ACCEPTED" is available in Jira,
> atleast in the filter.
>

We need custom workflow for this. I prefer to stick we simple standard
workflow at least for migration.

>> 6. Issue creation and last update date
>>
>> The following data was *not* converted:
>> 1. Attachments -- I didn't find easy way to download them. I'm also
>> not sure that JIRA supports importing attachments.
>> 2. History of changes.

>> Things I'm going to do for final migration:
>> 1. Add footer with original comment author for unmapped users.

Done.

>> 2. Investigate ways to migrate issue history (source data is not easy to parse).

Done.

>> Bert suggested to remap revision numbers to match ASF repository
>> revisions. But there is a risk that conversion replace something else
>> that match rNNN pattern. So I'm inclined to leave them as-is.
>
> You could add a "(asf repo: [rNNN+offset])" right behind the old
> number. But I don't think it's that important, these revs are
> typically used in issues that are resolved anyway.
>

It could be possible, but there are still risk to mangle original
comment. I still think that effort doesn't worth the result and
inclined to leave it as is.

I've asked INFRA team to perform another migration for us with
improved import data. I hope it will be final import.

--
Ivan Zhakov

[Ivan Zhakov]

On 27 August 2015 at 23:22, Ivan Zhakov <iv...@visualsvn.com> wrote:
> On 27 August 2015 at 22:52, Lieven Govaerts <l...@mobsol.be> wrote:
>> On Wed, Aug 26, 2015 at 8:52 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:
>>> Status update:
>>> ASF INFRA team performed test import for us:
>>> https://issues.apache.org/jira/browse/SERF
>>>
>>> The following data was converted:
>>> 1. Summary
>>> 2. Reporter and current assignee
>>> 3. Comments with original date and author
>>> 4. Labels
>>> 5. Issue status and resolution
>> Looks good.
>>
> Cool!
>

[...]

> I've asked INFRA team to perform another migration for us with
> improved import data. I hope it will be final import.
>

INFRA team completed another migration for us and I'm inclined to
consider it as final:
https://issues.apache.org/jira/browse/SERF

I think that current state is good enough. May be something could be
improved, but it doesn't worth the efforts. Opinions?

--
Ivan Zhakov

[Lieven Govaerts]

Looks fine by me Ivan, great work!

Will serf committers get a special role in the project in Jira? Or
does anyone with an account can do everything?

> --
> Ivan Zhakov
>

Lieven

[Ivan Zhakov]

Thanks!

> Will serf committers get a special role in the project in Jira? Or
> does anyone with an account can do everything?
>

Typical setup is the following [1]:
[[[
Q: How do permissions work in JIRA?

A: Projects are mapped to a single permission scheme (multiple
projects may share a permission scheme). There are users and groups
(no nested groups). A permission scheme lists possible actions (view
issues, create issues, assign issue, schedule issue, delete issue,
etc...) to one or more users or groups. For the ASF JIRA install, each
ASF project gets one JIRA project per releasable product (so you can
release version 1.6 of tool X and 1.3 of tool Y). Typically we create
one permission scheme for all those products. Typically we also create
one group for that permission scheme. Standard permissions are as
follows:

public can browse project (view issues), create issues, add comments,
link issues, and create attachments. issue reporter can edit an issue
project group (your developers/committers) can also edit issues,
schedule issues, move issues, assign issues, be assigned issues,
resolve issues, close issues, manage releases, and manage components.
one or two people also get added to "JIRA administrators", which is
system-wide access to let you manage who is in your group. Projects
are always free to suggest alternate plans.
]]]

[1] https://wiki.apache.org/general/ApacheJira#line-23

--
Ivan Zhakov