Beancount issues

[Martin Blais]

I've quickly archived all the Beancount issues and the repo in prevision of the Mercurial shutdown next week.

New changes will be lost.

It's possible to reimport them into another Mercurial project I think, that's the easiest avenue.

I still need to figure out where this is all going to go.

As you can imagine, with all that's going on, I have even less time.

I'm not sure yet where the repo will live, will advise

[Uwe Ziegenhagen]

Hi Martin,

if you need help, please let us know.

If you want to keep it simple:

a) just archive the Mercurial stuff

b) import the latest version only to Github

If you have the issues in a readable format ask us, your followers, to import them. I am sure some of us have plenty of time now. :-(

Uwe

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/CAK21%2BhNsp2S%2BfHX8bb0wuWTM%2Bb8Zk%2Bhgh7gia0pyLPvUeweamQ%40mail.gmail.com.

--

Dr. Uwe Ziegenhagen

0179-7476050
<http://www.uweziegenhagen.de>

[Tristan Kohl]

I do not know about importing archived Issues but a Fossil Repository contains the issues alongside everything else (wiki, issues/bugtracker,...). Maybe merging and hosting a Fossil repository might be a solution?

[Martin Michlmayr]

* Uwe Ziegenhagen <ziege...@gmail.com> [2020-03-22 16:26]:

> If you have the issues in a readable format ask us, your followers, to
> import them. I am sure some of us have plenty of time now. :-(

https://github.com/jeffwidman/bitbucket-issue-migration

And someone already said he created a Git repo.

--
Martin Michlmayr
https://www.cyrius.com/

[Martin Blais]

We already hvae a github repo here:

https://github.com/beancount/beancount

It's just not up-to-date because I don't use it, and sync'ing fails regularly (and I don't have time to fix it)

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20200322160427.GA25753%40jirafa.cyrius.com.

[Martin Blais]

I just attempted to update the github mirror now and it fails again.

I'm spending more time fixing this up than I have, I just can't, have to figure out trades for tomorrow's session

I'd contemplate switching over if I knew it was possible to switch back (demonstrated) without losing any data

[Stefano Zacchiroli]

On Sun, Mar 22, 2020 at 01:05:20PM -0400, Martin Blais wrote:
> I just attempted to update the github mirror now and it fails again.
> I'm spending more time fixing this up than I have, I just can't, have to
> figure out trades for tomorrow's session
> I'd contemplate switching over if I knew it was possible to switch back
> (demonstrated) without losing any data

I believe it is not your intention, but this comes out as a rant and is
not actionable by others. Do you need help and would you like someone
else to try doing the sync/migration for you? If so, please say it
explicitly and I'm confident people will step in to help out with this.
(I'm not volunteering myself because forced home schooling is now
absorbing my entire free time, but I'm optimist that others can help, as
long as it's clear that help with this is wanted/welcome.)

--
Stefano Zacchiroli . za...@upsilon.cc . upsilon.cc/zack . . o . . . o . o
Computer Science Professor . CTO Software Heritage . . . . . o . . . o o
Former Debian Project Leader & OSI Board Director . . . o o o . . . o .
« the first rule of tautology club is the first rule of tautology club »

[Kirill Goncharov]

I just updated my github mirror with your latest commit: https://github.com/xuhcc/beancount/commits/master

The tool that I'm using is git-remote-hg (https://github.com/mnauw/git-remote-hg). It allows both pull and push operations to remote mercurial repos, but local repo needs to be in git format.

I can help with keeping the mirrors in sync if it's needed.

On Sunday, March 22, 2020 at 8:05:35 PM UTC+3, Martin Blais wrote:

I just attempted to update the github mirror now and it fails again.

I'm spending more time fixing this up than I have, I just can't, have to figure out trades for tomorrow's session

I'd contemplate switching over if I knew it was possible to switch back (demonstrated) without losing any data

On Sun, Mar 22, 2020 at 12:13 PM Martin Blais <bl...@furius.ca> wrote:

We already hvae a github repo here:

https://github.com/beancount/beancount

It's just not up-to-date because I don't use it, and sync'ing fails regularly (and I don't have time to fix it)

On Sun, Mar 22, 2020 at 12:04 PM Martin Michlmayr <t...@cyrius.com> wrote:

* Uwe Ziegenhagen <ziege...@gmail.com> [2020-03-22 16:26]:
> If you have the issues in a readable format ask us, your followers, to
> import them. I am sure some of us have plenty of time now. :-(

https://github.com/jeffwidman/bitbucket-issue-migration

And someone already said he created a Git repo.

--
Martin Michlmayr
https://www.cyrius.com/

--
You received this message because you are subscribed to the Google Groups "Beancount" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bean...@googlegroups.com.

[Martin Blais]

On Sun, Mar 22, 2020 at 1:20 PM Stefano Zacchiroli <za...@upsilon.cc> wrote:

On Sun, Mar 22, 2020 at 01:05:20PM -0400, Martin Blais wrote:
> I just attempted to update the github mirror now and it fails again.
> I'm spending more time fixing this up than I have, I just can't, have to
> figure out trades for tomorrow's session
> I'd contemplate switching over if I knew it was possible to switch back
> (demonstrated) without losing any data

I believe it is not your intention, but this comes out as a rant and is
not actionable by others. Do you need help and would you like someone
else to try doing the sync/migration for you? If so, please say it
explicitly and I'm confident people will step in to help out with this.

Sorry I don't mean to rant.

I meant that I'm now ruling out a permanent hg-git solution.

It just breaks too much, that's not going to work.

So either I find another Hg hosting (doesn't seem to be anything foss friendly with issue tracking out there I feel confident about), or try a move to github.

Okay, so for github: in order to give it an honest try, I need to find out:

- If converted, are all the branches still present and usable? I need to try other export methods.

- Will it be possible to convert back to Hg later? (probably not moving w/out that possiiblity)

- Are exported issues from bitbucket importable into github?

These are the things I'd need to figure out before I try moving this project to github as the main repo.

I need to run tests on these things.

If someone wants to help answer these questions LMK.

Short of that the project will just expire on bitbucket and eventually reappear on my own fickle server at hg.furius.ca.

 

(I'm not volunteering myself because forced home schooling is now
absorbing my entire free time, but I'm optimist that others can help, as
long as it's clear that help with this is wanted/welcome.)

Similar situation here, plus it's tax time

These are VERY busy days

 

--
Stefano Zacchiroli . za...@upsilon.cc . upsilon.cc/zack . . o . . . o . o
Computer Science Professor . CTO Software Heritage . . . . . o . . . o o
Former Debian Project Leader & OSI Board Director  . . . o o o . . . o .
« the first rule of tautology club is the first rule of tautology club »

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20200322172049.puxm7tnwxomfbsoj%40upsilon.cc.

[Stefano Zacchiroli]

On Sun, Mar 22, 2020 at 01:58:53PM -0400, Martin Blais wrote:
> So either I find another Hg hosting (doesn't seem to be anything foss
> friendly with issue tracking out there I feel confident about), or try
> a move to github.

It's known™ that I'd personally favor a move to a git-based hosting.
But. In case it would make things easier for you, I've checked with a
friend who's also an upstream Mercurial developer what he recommends as
an alternative to Bitbucket. He pointed that this hosting:

https://foss.heptapod.net/

offers free hosting for FOSS projects. It's GitLab, but extended with
Mercurial software. It's developed and maintained by Mercurial upstream
devs, so it should be reasonably well integrated and safe mid-term (not
that I've tested it myself though...).

I've no idea how to migrate issues from bitbucket to GitLab, but I can
ask.

Hope this helps,
Cheers.

[Martin Blais]

Kirill: Thanks, I managed to re-export from scratch using git-remote-log.

Despite this being a completely separate export, it did manage to update the github.com/beancount/beancount repo, so it must be using stable hashes or some other way to avoid allocating new hashes when exporting (which is great).

I can't see if all my branches are up-to-date though.

Stefano: I saw that a while back, I thought it didn't have issues, but having a second look now, it does not nice.

I'll kick the tires real quick just now, this could be a quick and dirty option.

Thank you,

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20200322192957.37lc7iluz6gkptes%40upsilon.cc.

[Martin Blais]

On Sun, Mar 22, 2020 at 8:48 PM Martin Blais <bl...@furius.ca> wrote:

Kirill: Thanks, I managed to re-export from scratch using git-remote-log.

Despite this being a completely separate export, it did manage to update the github.com/beancount/beancount repo, so it must be using stable hashes or some other way to avoid allocating new hashes when exporting (which is great).

I can't see if all my branches are up-to-date though.

Stefano: I saw that a while back, I thought it didn't have issues, but having a second look now, it does not nice.

I'll kick the tires real quick just now, this could be a quick and dirty option.

Ahh yes, now I know why. It's just the software, it's not hosting.

I need a hosted solution.

[Stefano Zacchiroli]

On Sun, Mar 22, 2020 at 10:06:01PM -0400, Martin Blais wrote:
> Ahh yes, now I know why. It's just the software, it's not hosting.
> I need a hosted solution.

They do hosting too. The only restriction is that the project should be
under a FOSS license, which is the case for Beancount. I think you can
just create an account here https://foss.heptapod.net/explore and then
create a project. (But, again, I haven't tried it myself.) If you want
to explore this and encounter troubles, let me know and I can put you in
touch with the maintainers.

Cheers

[Daniele Nicolodi]

On 23/03/2020 02:30, Stefano Zacchiroli wrote:
> On Sun, Mar 22, 2020 at 10:06:01PM -0400, Martin Blais wrote:
>> Ahh yes, now I know why. It's just the software, it's not hosting.
>> I need a hosted solution.
>
> They do hosting too. The only restriction is that the project should be
> under a FOSS license, which is the case for Beancount. I think you can
> just create an account here https://foss.heptapod.net/explore and then
> create a project. (But, again, I haven't tried it myself.) If you want
> to explore this and encounter troubles, let me know and I can put you in
> touch with the maintainers.

To further explore this possibility, I went ahead and filled a hosting
request ticket on https://foss.heptapod.net

https://foss.heptapod.net/heptapod/foss.heptapod.net/issues/49

The only requirement on the project, other than being released under a
FOSS license is to acknowledge support from Octobus and Clever Cloud for
hosting on the project webpage.

Martin, please let me know if those seems like acceptable terms and how
you would like a Gitlab-like hosting. In particular, this is the
workflow that Heptapod promotes (I haven't read it in detail yet):

https://octobus.net/blog/2019-09-04-heptapod-workflow.html

As an experimental physicists, I foresee running out soon of useful work
to do at home, thus I can dedicate some time to following the migration.

Cheers,
Dan

[Daniele Nicolodi]

On 27/03/2020 12:11, Daniele Nicolodi wrote:
> To further explore this possibility, I went ahead and filled a hosting
> request ticket on https://foss.heptapod.net
>
> https://foss.heptapod.net/heptapod/foss.heptapod.net/issues/49
>
> The only requirement on the project, other than being released under a
> FOSS license is to acknowledge support from Octobus and Clever Cloud for
> hosting on the project webpage.

The folks at Heptapod responded very quickly and approved my request for
hosting the beancount repository under a beancount group (organization).

Martin, if you would like to have them import the Bitbucket repository,
issues, and merge requests to Heptapod, you should invite
georges...@octobus.net to have (read only) repository membership on
Bitbucket (this is a requirement of the import mechanism).

The only limitation I see is that ownership of issues and merge requests
will most likely be lost. This is a know issue, unlikely to be resolved
before Bitbucket shuts down mercurial projects.

Heptapod admins are processing a backlog of hosting requests and should
be able to run the import by the end of next week (and releasing a new
Heptapod RC).

Cheers,
Dan

[Martin Blais]

Thank you Daniele.

Just as I had pretty much decided to give in to the Github...

I had ruled out Heptapod as it seemed to have only a few projects it itself owned just even last week (and now I see the list growing, so it's morphing into a hosting solution).

Let me think about this a bit more. 

[Daniele Nicolodi]

Hi Martin,

Sure. I didn't want to force your hand one way or the other (although,
if you decide to move to git, I would prefer Gitlab over GitHub). I was
just investigating what it would take to migrate to Heptapod. I didn't
expect things to move so quickly.

Cheers,
Dan

On 28/03/2020 13:10, Martin Blais wrote:
> Thank you Daniele.
> Just as I had pretty much decided to give in to the Github...
> I had ruled out Heptapod as it seemed to have only a few projects it
> itself owned just even last week (and now I see the list growing, so
> it's morphing into a hosting solution).
> Let me think about this a bit more. 
>
>
>
> On Fri, Mar 27, 2020 at 5:20 PM Daniele Nicolodi <dan...@grinta.net
> <mailto:dan...@grinta.net>> wrote:
>
> On 27/03/2020 12:11, Daniele Nicolodi wrote:
> > To further explore this possibility, I went ahead and filled a hosting
> > request ticket on https://foss.heptapod.net
> >
> > https://foss.heptapod.net/heptapod/foss.heptapod.net/issues/49
> >
> > The only requirement on the project, other than being released under a
> > FOSS license is to acknowledge support from Octobus and Clever
> Cloud for
> > hosting on the project webpage.
>
> The folks at Heptapod responded very quickly and approved my request for
> hosting the beancount repository under a beancount group (organization).
>
> Martin, if you would like to have them import the Bitbucket repository,
> issues, and merge requests to Heptapod, you should invite

> georges...@octobus.net <mailto:georges...@octobus.net> to

[Red S]

Just curious: is https://bitbucket.org/blais/beancount/src/default/ still the official repo? I can still seem to check out from it. Is mercurial on bitbucket going to remain alive?

[Stefano Zacchiroli]

On Sat, Mar 28, 2020 at 03:10:52PM -0400, Martin Blais wrote:
> I had ruled out Heptapod as it seemed to have only a few projects it
> itself owned just even last week (and now I see the list growing, so it's
> morphing into a hosting solution).

Looks like this has been made official now:

https://heptapod.host/

the non-payware part for FOSS projects is at:

https://foss.heptapod.net/

[Martin Blais]

Nothing changed.

I'm too busy to deal with it, a few weeks ago I was set to move the whole thing to github manually but now it looks like there's a bunch of easier solutions are coming around (converters to github that will also import issues, heptapod to stick with mercurial, etc.).

Atlassian also delayed the deadline.

I'm waiting until the very last moment, just got too much going on

On Sat, Apr 25, 2020 at 2:21 AM Red S <redst...@gmail.com> wrote:

Just curious: is https://bitbucket.org/blais/beancount/src/default/ still the official repo? I can still seem to check out from it. Is mercurial on bitbucket going to remain alive?

--

You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/13374d31-594b-4a63-b594-8b8a87461f43%40googlegroups.com.

[Red S]

Hi Martin,

Okay, good to know, and no rush.

As others suggested, please do feel free to lean on this community for anything we can assist you with. If you want help with testing the handful of bitbucket-to-github solutions that exist, for instance, do let us know. I'm sure several of us would be glad to help :).

Related: I'm wondering if git and github would make beancount more accessible to folks wanting to make contributions, if that's something you consider to be an advantage as you think of reducing the scope of beancount so you can focus on its depth and leave things like the query layer to contributors in the larger beancount ecosystem.

On Saturday, April 25, 2020 at 7:48:35 AM UTC-7, Martin Blais wrote:

Nothing changed.

I'm too busy to deal with it, a few weeks ago I was set to move the whole thing to github manually but now it looks like there's a bunch of easier solutions are coming around (converters to github that will also import issues, heptapod to stick with mercurial, etc.).

Atlassian also delayed the deadline.

I'm waiting until the very last moment, just got too much going on

On Sat, Apr 25, 2020 at 2:21 AM Red S <redst...@gmail.com> wrote:

Just curious: is https://bitbucket.org/blais/beancount/src/default/ still the official repo? I can still seem to check out from it. Is mercurial on bitbucket going to remain alive?

--
You received this message because you are subscribed to the Google Groups "Beancount" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bean...@googlegroups.com.

[Martin Blais]

On Sat, Apr 25, 2020 at 7:22 PM Red S <redst...@gmail.com> wrote:

Hi Martin,

Okay, good to know, and no rush.

As others suggested, please do feel free to lean on this community for anything we can assist you with. If you want help with testing the handful of bitbucket-to-github solutions that exist, for instance, do let us know. I'm sure several of us would be glad to help :).

Related: I'm wondering if git and github would make beancount more accessible to folks wanting to make contributions, if that's something you consider to be an advantage as you think of reducing the scope of beancount so you can focus on its depth and leave things like the query layer to contributors in the larger beancount ecosystem.

I know. That's the main draw. I already figured how to do that conversion, it's just missing importing the tickets.

 

 

On Saturday, April 25, 2020 at 7:48:35 AM UTC-7, Martin Blais wrote:

Nothing changed.

I'm too busy to deal with it, a few weeks ago I was set to move the whole thing to github manually but now it looks like there's a bunch of easier solutions are coming around (converters to github that will also import issues, heptapod to stick with mercurial, etc.).

Atlassian also delayed the deadline.

I'm waiting until the very last moment, just got too much going on

On Sat, Apr 25, 2020 at 2:21 AM Red S <redst...@gmail.com> wrote:

Just curious: is https://bitbucket.org/blais/beancount/src/default/ still the official repo? I can still seem to check out from it. Is mercurial on bitbucket going to remain alive?

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bean...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/13374d31-594b-4a63-b594-8b8a87461f43%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Beancount" group.

To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/424d067c-df90-4b4e-84fb-976e3f2949ec%40googlegroups.com.

[Kirill Goncharov]

Just tried this issue migration tool: https://github.com/jeffwidman/bitbucket-issue-migration

Here's the result: https://github.com/xuhcc/beancount/issues

And Gitlab apparently has a pretty good built-in importer: https://docs.gitlab.com/ee/user/project/import/bitbucket.html

(Regarding the platform choice - Gitlab has less users than Github but it is open-source and we can migrate to self-hosted instance at any time; both seem OK to me. I would agree with Red S that git is preferrable because everyone is familiar with it.)

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/424d067c-df90-4b4e-84fb-976e3f2949ec%40googlegroups.com.

[Martin Michlmayr]

* Kirill Goncharov <kdgon...@gmail.com> [2020-04-26 05:49]:

> Just tried this issue migration tool:
> https://github.com/jeffwidman/bitbucket-issue-migration
>
> Here's the result: https://github.com/xuhcc/beancount/issues

Thanks so much for doing this! I was going to do a test migration but
I'm happy you did it already. ;)

I hope we can move to GitHub since that's the most widely used
platform. Kirill has shown how to migrate from mercurial to Git and
how to migrate the issues.

I have some patches I'd like to submit to beancount and I'll commit to
creating several pull requests if beancount moves to GitHub. ;)

Kirill: in some tickets the Markdown or something seems broken, e.g.
https://github.com/xuhcc/beancount/issues/428

[Kirill Goncharov]

Kirill: in some tickets the Markdown or something seems broken, e.g.
https://github.com/xuhcc/beancount/issues/428

I think we can simply remove profile links and keep only a reference to original issue with a display name:

https://github.com/xuhcc/bitbucket-issue-migration/commit/aa39e88a7e43c719065553c336af804a294dbc20

This kind of username matching is not reliable anyway

[Martin Michlmayr]

* Kirill Goncharov <kdgon...@gmail.com> [2020-04-26 07:02]:

> I think we can simply remove profile links and keep only a reference to
> original issue with a display name:

I kinda like the original links.

> This kind of username matching is not reliable anyway

I can work on a mapping of user names.

[Daniele Nicolodi]

As anticipated a couple of weeks ago, the folks at Heptapod ran a test
migration of beancount from Bitbucket to their Gitlab-like platform
with Mercurial support. It looks good at a first glance. Currently the
bitbucket project on Heptapod is in private mode, but we can switch it
to public any time, if we decide to go with Heptapod as an hosting solution.

Cheers,
Dan

> <https://groups.google.com/d/msgid/beancount/13374d31-594b-4a63-b594-8b8a87461f43%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> You received this message because you are subscribed to the Google
> Groups "Beancount" group.
> To unsubscribe from this group and stop receiving emails from it,

> send an email to beancount+...@googlegroups.com
> <mailto:beancount+...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/beancount/424d067c-df90-4b4e-84fb-976e3f2949ec%40googlegroups.com
> <https://groups.google.com/d/msgid/beancount/424d067c-df90-4b4e-84fb-976e3f2949ec%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> You received this message because you are subscribed to the Google
> Groups "Beancount" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to beancount+...@googlegroups.com
> <mailto:beancount+...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/beancount/CAK21%2BhOszDL4MF2zcNZCXOLMf%3D9ypnDuJQ_sErpfDEwFBga0Gg%40mail.gmail.com
> <https://groups.google.com/d/msgid/beancount/CAK21%2BhOszDL4MF2zcNZCXOLMf%3D9ypnDuJQ_sErpfDEwFBga0Gg%40mail.gmail.com?utm_medium=email&utm_source=footer>.

[tinot...@gmail.com]

 

Hello,

As everybody I'm ready to help with anything for the migration. And I would like to vote (I know that is not a voting decision but anyway) for Github first and Gitlab second.

I think is going to be easier to work all together later even if we loss some information in the migration.

Regards.

[Martin Blais]

Sorry I lost that thread.

I just pulled it up now and looked at the tickets, looks very nice indeed, thanks for this test run Kirill.

Did you do anything beyond running https://github.com/jeffwidman/bitbucket-issue-migration?

I see you have your own fork.

It's a bummer the author of all tickets ends up being who runs the script, but we can live with that and manually Cc the github author when a ticket is updated.

I checked on some of the tickets, most links to the original submitters seem to work.

This is good enough.

Martin: Did you want to go beyond this and create a mapping?

--
You received this message because you are subscribed to the Google Groups "Beancount" group.

To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20200426142749.GA6947%40jirafa.cyrius.com.

[Kirill Goncharov]

>I see you have your own fork.

In some cases the script creates invalid markup, for example here: https://github.com/xuhcc/beancount/issues/428

I looked into the code and found that it all comes down to account matching procedure and it's hard to do it right automatically, so I removed it altogether in my fork. But then Martin Michlmayr suggested to create a mapping manually, I think it's a good approach too.

On Monday, May 18, 2020 at 11:24:42 AM UTC+3, Martin Blais wrote:

Sorry I lost that thread.

I just pulled it up now and looked at the tickets, looks very nice indeed, thanks for this test run Kirill.

Did you do anything beyond running https://github.com/jeffwidman/bitbucket-issue-migration?

I see you have your own fork.

It's a bummer the author of all tickets ends up being who runs the script, but we can live with that and manually Cc the github author when a ticket is updated.

I checked on some of the tickets, most links to the original submitters seem to work.

This is good enough.

Martin: Did you want to go beyond this and create a mapping?

On Sun, Apr 26, 2020 at 10:28 AM Martin Michlmayr <t...@cyrius.com> wrote:

* Kirill Goncharov <kdgon...@gmail.com> [2020-04-26 07:02]:
> I think we can simply remove profile links and keep only a reference to
> original issue with a display name:

I kinda like the original links.

> This kind of username matching is not reliable anyway

I can work on a mapping of user names.

--
Martin Michlmayr
https://www.cyrius.com/

--
You received this message because you are subscribed to the Google Groups "Beancount" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bean...@googlegroups.com.

[Martin Michlmayr]

* Kirill Goncharov <kdgon...@gmail.com> [2020-05-18 01:43]:

> But then Martin Michlmayr suggested to create a mapping manually, I
> think it's a good approach too.

I can take a look tomorrow.

[Martin Michlmayr]

* Martin Blais <bl...@furius.ca> [2020-05-18 04:24]:

> It's a bummer the author of all tickets ends up being who runs the script,

Unfortunately the GitHub API doesn't allow setting the submitter.
I ran into the same issue with the ledger migration from Bugzilla.

> but we can live with that and manually Cc the github author when a ticket
> is updated.

Yeah, agreed.

[Martin Michlmayr]

* Martin Blais <bl...@furius.ca> [2020-05-18 04:24]:

> Martin: Did you want to go beyond this and create a mapping?

I have a mapping now for some users. For quite a few people I cannot
find a mapping. I think once we've done the migration, we can update
the BitBucket issues and let them know about the move to GitHub.

Anyway, I emailed some people asking for their GitHub username, so
please give me a few days.

[Martin Blais]

SGTM, I'll wait for your mapping and then I'll test it with git-remote-hg.

Thanks a lot for your help and thanks to Kirill Goncharov for figuring out the issues migration.

I re-ran the hg-fast-export conversion and git-remote-hg ones last night to see if there's any benefit to using one vs. the other.

The former fails with an exception; the latter works well.  It'll be a git-remote-log conversion.

I diffed all the heads of branches to make sure nothing's lost; seems to have worked perfectly.

I also spot-checked some of the per-branch logs, they also fine.

The tags are also present and matching.

There are a few final relatively easy things I need to figure out:

- I have some repo lying around with local changes that needs to get merged before I make the final conversion. I'll merge those changes in hg and reimport.

- Mercurial has the concept of "closing" a branch (typically when it gets merged). This results over time in a large set of "inactive" and a smaller set of "active" branches. The conversion merely creates branches, for both active and inactive ones. The git-remote-hg docs mention "Closed branches are not supported; they are not shown and you can’t close or reopen. Additionally in certain rare situations a synchronization issue can occur (Bug #65)."  I want to figure a way to keep the branch refs / history yet have them not show up in the list of branches (so that the little github menu doesn't show all these closed branches as if they were work in progress). I think I may use a tag in under archive/ like some people do to differentiate those from tags for released versions.

- Dominik Aumayr created the github.com/beancount/beancount repository originally and gave me admin rights. While I'm able to pull the output of a fresh git-remote-hg conversion on top of that repo (with a minor merge on a tags file), I'd prefer to empty it or remove it and recreate it from scratch with the latest version of the git-remote-hg converter.  Is this going to cause problems? I suspect I would not since I can pull the recent conversion on top, the checksums are probably the same and the repos would be compatible. Have you done that before? I wonder if it's easiest to delete and recreate the repo project in github (does github make that possible or if I delete the repo would the beancount/beancount name be unavailable forever?).

After I refresh the code and migrate the tickets, I'd love to integrate Kirill's .md docs and ask if people can remove other static docs to there's only one copy out there and change all pointers to the repo.

[Kirill Goncharov]

>After I refresh the code and migrate the tickets, I'd love to integrate Kirill's .md docs

How would you like to do that?

I think it's better to keep the parser and resulting markdown files in a separate repo unless there's a compelling reason to do otherwise, as this allows frequent updates (I currently do re-parsing once a week).

That being said, I would happily transfer my beancount-docs repo to https://github.com/beancount/ organization and add you as collaborator.

[Martin Blais]

On Wed, May 20, 2020 at 6:34 AM Kirill Goncharov <kdgon...@gmail.com> wrote:

>After I refresh the code and migrate the tickets, I'd love to integrate Kirill's .md docs

How would you like to do that?

I think it's better to keep the parser and resulting markdown files in a separate repo unless there's a compelling reason to do otherwise, as this allows frequent updates (I currently do re-parsing once a week).

That being said, I would happily transfer my beancount-docs repo to https://github.com/beancount/ organization and add you as collaborator.

Come to think of it, this makes the most sense. This way there's no need to pepper the code repo with documentation updates.

I'd love for you to move that repo under "http://github.com/beancount/docs" if you don't mind, that would be ideal.

It would make a https://beancount.github.io/docs/ root for that documentation site, which is a pretty nice URL.

Two more things:

- Feel free to stick your name in the generated docs somewhere as the maintainer and get some credit for all that nice work.  Alternatively, I have a credits file somewhere in the source repo that I can convert to a Google doc, that would be an easy way to make this appear in your generated docs.

- Your conversikon seems like it covers all of the static docs generated by Dominik, here:

https://aumayr.github.io/beancount-docs-static/

Is this correct? Does it cover everything that is there?

At a quick glance, I notice Dominik's has a full index in the API reference, and not a hierarchical one (not sure this matters):

https://aumayr.github.io/beancount-docs-static/api_reference/index.html

Also your API reference is missing the "projects" subdirectory.

Let me know if I help with anything to make this happen,

 

 

On Tuesday, May 19, 2020 at 4:57:14 PM UTC+3, Martin Blais wrote:

On Tue, May 19, 2020 at 5:28 AM Martin Michlmayr <t...@cyrius.com> wrote:

* Martin Blais <bl...@furius.ca> [2020-05-18 04:24]:
> Martin: Did you want to go beyond this and create a mapping?

I have a mapping now for some users.  For quite a few people I cannot
find a mapping.  I think once we've done the migration, we can update
the BitBucket issues and let them know about the move to GitHub.

 

Anyway, I emailed some people asking for their GitHub username, so
please give me a few days.

SGTM, I'll wait for your mapping and then I'll test it with git-remote-hg.

Thanks a lot for your help and thanks to Kirill Goncharov for figuring out the issues migration.

I re-ran the hg-fast-export conversion and git-remote-hg ones last night to see if there's any benefit to using one vs. the other.

The former fails with an exception; the latter works well.  It'll be a git-remote-log conversion.

I diffed all the heads of branches to make sure nothing's lost; seems to have worked perfectly.

I also spot-checked some of the per-branch logs, they also fine.

The tags are also present and matching.

There are a few final relatively easy things I need to figure out:

- I have some repo lying around with local changes that needs to get merged before I make the final conversion. I'll merge those changes in hg and reimport.

- Mercurial has the concept of "closing" a branch (typically when it gets merged). This results over time in a large set of "inactive" and a smaller set of "active" branches. The conversion merely creates branches, for both active and inactive ones. The git-remote-hg docs mention "Closed branches are not supported; they are not shown and you can’t close or reopen. Additionally in certain rare situations a synchronization issue can occur (Bug #65)."  I want to figure a way to keep the branch refs / history yet have them not show up in the list of branches (so that the little github menu doesn't show all these closed branches as if they were work in progress). I think I may use a tag in under archive/ like some people do to differentiate those from tags for released versions.

- Dominik Aumayr created the github.com/beancount/beancount repository originally and gave me admin rights. While I'm able to pull the output of a fresh git-remote-hg conversion on top of that repo (with a minor merge on a tags file), I'd prefer to empty it or remove it and recreate it from scratch with the latest version of the git-remote-hg converter.  Is this going to cause problems? I suspect I would not since I can pull the recent conversion on top, the checksums are probably the same and the repos would be compatible. Have you done that before? I wonder if it's easiest to delete and recreate the repo project in github (does github make that possible or if I delete the repo would the beancount/beancount name be unavailable forever?).

After I refresh the code and migrate the tickets, I'd love to integrate Kirill's .md docs and ask if people can remove other static docs to there's only one copy out there and change all pointers to the repo.

--
You received this message because you are subscribed to the Google Groups "Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/fac47049-267d-4205-88f3-6e030e7d2039%40googlegroups.com.

[Kirill Goncharov]

I'd love for you to move that repo under "http://github.com/beancount/docs" if you don't mind, that would be ideal.

Sure, I'll rename it.

According to the GitHub documentation, I need a permission to create repositories in organization in order to initiate the transfer. Can you grant me such permission?

Feel free to stick your name in the generated docs somewhere as the maintainer and get some credit for all that nice work.  Alternatively, I have a credits file somewhere in the source repo that I can convert to a Google doc, that would be an easy way to make this appear in your generated docs.

We have a history of commits, I think that's enough.

 

At a quick glance, I notice Dominik's has a full index in the API reference, and not a hierarchical one (not sure this matters):

https://aumayr.github.io/beancount-docs-static/api_reference/index.html

I didn't found a way to build such an index automatically with mkdocs.

Also your API reference is missing the "projects" subdirectory.

This subdirectory is not present in the beancount package downloaded from PyPI.

Also, the documentation generator fails to read 'tools' subdirectory, I can't figure out why. But I added documentation pages for 'loader', 'prices' and 'web' which were missing.

To unsubscribe from this group and stop receiving emails from it, send an email to bean...@googlegroups.com.

[Martin Blais]

On Wed, May 20, 2020 at 2:35 PM Kirill Goncharov <kdgon...@gmail.com> wrote:

I'd love for you to move that repo under "http://github.com/beancount/docs" if you don't mind, that would be ideal.

Sure, I'll rename it.

According to the GitHub documentation, I need a permission to create repositories in organization in order to initiate the transfer. Can you grant me such permission?

I'm not able to grant at the moment, I'm merely a Member of the org; I just asked Dominik if he'll promote me to Owner.

In any case, I've created an empty repo here:

https://github.com/beancount/docs

Are you able to push it over that (along with the history)?

If you'd prefer to move it (I didn't know it was possible to move a repo across users) you can wait, if I have Owner ACLs I'd be happy to make that happen.

 

Feel free to stick your name in the generated docs somewhere as the maintainer and get some credit for all that nice work.  Alternatively, I have a credits file somewhere in the source repo that I can convert to a Google doc, that would be an easy way to make this appear in your generated docs.

We have a history of commits, I think that's enough.

 

At a quick glance, I notice Dominik's has a full index in the API reference, and not a hierarchical one (not sure this matters):

https://aumayr.github.io/beancount-docs-static/api_reference/index.html

I didn't found a way to build such an index automatically with mkdocs.

No worries.

  

Also your API reference is missing the "projects" subdirectory.

This subdirectory is not present in the beancount package downloaded from PyPI.

Also, the documentation generator fails to read 'tools' subdirectory, I can't figure out why. But I added documentation pages for 'loader', 'prices' and 'web' which were missing.

Strange. Maybe there's something hardcoded in the generator?

 

To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/bbc6da8c-20a5-4c0d-9292-ce9a112b8c99%40googlegroups.com.

[Kirill Goncharov]

If you'd prefer to move it (I didn't know it was possible to move a repo across users) you can wait, if I have Owner ACLs I'd be happy to make that happen.

Yes, I would prefer transfer as it preserves issues, watchers and stars. Let's wait for Dominik's response.

Strange. Maybe there's something hardcoded in the generator?

I don't know. It says that `beancount.tools` is a module that has no attribute `sheets_upload` (full traceback: https://github.com/xuhcc/beancount-docs/issues/14).

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/bbc6da8c-20a5-4c0d-9292-ce9a112b8c99%40googlegroups.com.

[Martin Blais]

On Tue, May 19, 2020 at 9:56 AM Martin Blais <bl...@furius.ca> wrote:

On Tue, May 19, 2020 at 5:28 AM Martin Michlmayr <t...@cyrius.com> wrote:

* Martin Blais <bl...@furius.ca> [2020-05-18 04:24]:
> Martin: Did you want to go beyond this and create a mapping?

I have a mapping now for some users.  For quite a few people I cannot
find a mapping.  I think once we've done the migration, we can update
the BitBucket issues and let them know about the move to GitHub.

 

Anyway, I emailed some people asking for their GitHub username, so
please give me a few days.

SGTM, I'll wait for your mapping and then I'll test it with git-remote-hg.

Thanks a lot for your help and thanks to Kirill Goncharov for figuring out the issues migration.

I re-ran the hg-fast-export conversion and git-remote-hg ones last night to see if there's any benefit to using one vs. the other.

The former fails with an exception; the latter works well.  It'll be a git-remote-log conversion.

I diffed all the heads of branches to make sure nothing's lost; seems to have worked perfectly.

I also spot-checked some of the per-branch logs, they also fine.

The tags are also present and matching.

There are a few final relatively easy things I need to figure out:

- I have some repo lying around with local changes that needs to get merged before I make the final conversion. I'll merge those changes in hg and reimport.

That's done.

- Mercurial has the concept of "closing" a branch (typically when it gets merged). This results over time in a large set of "inactive" and a smaller set of "active" branches. The conversion merely creates branches, for both active and inactive ones. The git-remote-hg docs mention "Closed branches are not supported; they are not shown and you can’t close or reopen. Additionally in certain rare situations a synchronization issue can occur (Bug #65)."  I want to figure a way to keep the branch refs / history yet have them not show up in the list of branches (so that the little github menu doesn't show all these closed branches as if they were work in progress). I think I may use a tag in under archive/ like some people do to differentiate those from tags for released versions.

That's done (tagged as archive/).

 

- Dominik Aumayr created the github.com/beancount/beancount repository originally and gave me admin rights. While I'm able to pull the output of a fresh git-remote-hg conversion on top of that repo (with a minor merge on a tags file), I'd prefer to empty it or remove it and recreate it from scratch with the latest version of the git-remote-hg converter.  Is this going to cause problems? I suspect I would not since I can pull the recent conversion on top, the checksums are probably the same and the repos would be compatible. Have you done that before? I wonder if it's easiest to delete and recreate the repo project in github (does github make that possible or if I delete the repo would the beancount/beancount name be unavailable forever?).

I'm just missing Owner ACLs to the Beancount organization, from either Dominik or Jakob so I can move the repository.

After I refresh the code and migrate the tickets, I'd love to integrate Kirill's .md docs and ask if people can remove other static docs to there's only one copy out there and change all pointers to the repo.

Ditto, same thing, once I'm one of the owners, Kirill's docs repo will be moved at beancount/docs.

There's another task before we can complete it: I need to handle the pending 16 pull requests.

I'll try to review/merge as many of the small ones as possible, and the larger ones I'll try to save as branches in the converted git repo.

It's a decision, the Beancount repo will be moving to github soon.

[Martin Blais]

I've either merged (in hg), or converted to branches all the pending pull requests and closed them all. 

I named them prXXX_<user>_<name>.

All I'm missing to get the job finished is "Owner" ACLs to the organization, so I can move the repo and help Kirill move docs into it too. 

Jakob? Dominik?

[Jakob Schnitzer]

Hi Martin,

thanks for the direct CC. I'm happy that the migration is happening :)

On Thu, May 21, 2020 at 12:41:22AM -0400, Martin Blais wrote:
>> - Dominik Aumayr created the github.com/beancount/beancount
>> repository
>>> originally and gave me admin rights. While I'm able to pull the output of a
>>> fresh git-remote-hg conversion on top of that repo (with a minor merge on a
>>> tags file), I'd prefer to empty it or remove it and recreate it from
>>> scratch with the latest version of the git-remote-hg converter. Is this
>>> going to cause problems? I suspect I would not since I can pull the recent
>>> conversion on top, the checksums are probably the same and the repos would
>>> be compatible. Have you done that before? I wonder if it's easiest to
>>> delete and recreate the repo project in github (does github make that
>>> possible or if I delete the repo would the beancount/beancount name be
>>> unavailable forever?).

Doing a force push with git (git push --force) will ensure that the
branch that you're pushing gets completely synced with your local state
(and basically corresponds to "recreating" it). That should be slightly
quicker than delete-and-recreate (I'm also not sure if the deletion
could cause any problems).

>All I'm missing to get the job finished is "Owner" ACLs to the
>organization, so I can move the repo and help Kirill move docs into it too.
>Jakob? Dominik?

Seems like I was slightly too late, you're already an owner of the
Beancount org on Github now.

Best,
Jakob

[Martin Blais]

On Wed, May 20, 2020 at 3:45 PM Kirill Goncharov <kdgon...@gmail.com> wrote:

If you'd prefer to move it (I didn't know it was possible to move a repo across users) you can wait, if I have Owner ACLs I'd be happy to make that happen.

Yes, I would prefer transfer as it preserves issues, watchers and stars. Let's wait for Dominik's response.

Done.

I deleted the empty github/beancount/docs repository I had created and invited you to join the org, and checked that it should allow you to create new repos.

Let me know if you encounter any issues.

Thank you Kirill, and looking fwd to your Beancount docs at that location.

 

 

To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/a5714ee2-852a-4a19-acba-26c5c1ab6fc3%40googlegroups.com.

[Kirill Goncharov]

Thanks! I successfuly transferred the repo: https://github.com/beancount/docs

Documentation site is at https://beancount.github.io/docs/

Now we need to change site url in repo description (https://github.com/beancount/docs). I guess it requires admin privileges.

Contributions page also needs an update:

https://docs.google.com/document/d/1Z37bQ45wDtjTPaMQ_x-f33p1trH9fNosEAUgbQXwp30/edit#heading=h.wcyqutvazcon

To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/a5714ee2-852a-4a19-acba-26c5c1ab6fc3%40googlegroups.com.

[Martin Blais]

On Fri, May 22, 2020 at 5:27 AM Kirill Goncharov <kdgon...@gmail.com> wrote:

Thanks! I successfuly transferred the repo: https://github.com/beancount/docs

Documentation site is at https://beancount.github.io/docs/

Now we need to change site url in repo description (https://github.com/beancount/docs). I guess it requires admin privileges.

Done!

 

Contributions page also needs an update:

https://docs.google.com/document/d/1Z37bQ45wDtjTPaMQ_x-f33p1trH9fNosEAUgbQXwp30/edit#heading=h.wcyqutvazcon

Done!

Thank you Kirill, your help is much appreciated

  

To unsubscribe from this group and stop receiving emails from it, send an email to beancount+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/a9139221-84a7-4efe-9f6e-029b9a92754a%40googlegroups.com.