Re: [fossil-users] Announce: JSON is in the trunk

[Stephan Beal]

On Fri, Nov 4, 2011 at 11:17 PM, Nolan Darilek wrote:

>  Cool! Is it documented anywhere yet?
>

The current docs are in:

https://docs.google.com/document/d/1fXViveNhDbiXgCuE7QDXQOKeFzf2qNUkBEgiUvoqFN4/edit

and are updated as the code is written.

As the API stabilizes the docs will be moved into the "embedded docs"
section of the main site, but that will not happen very soon (i think we're
at least a couple months away from that)


> If not, how would I go about retrieving the latest X events from the
> timeline? Ideally I'd like to add a "Latest news" section on my homepage as
> rendered HTML, but in the meantime I'll implement that with JQuery.
>

/json/timeline/ci?limit=X
/json/timeline/wiki?limit=X
/json/timeline/ticket?limit=X (ticket support is skeletal/incomplete)
/json/timeline/event?limit=X (not yet implemented, i think)

There are still open questions and problems, to solve, one of which you
touch on here:

> homepage as rendered HTML

Commit messages containing constructs like [artifactid] are not current
parsed by the JSON API. You will get the raw strings. The plan is to
eventually return the raw strings plus additional data containing info
about the references inside the comment string. As a rought example:

"comment": "merged [deadbeef] and [cafebeef].",
"references": [ "deadbeaf", "cafebeef"]

but that's all theoretical at this point - we have bigger fish to fry
before solving that, so there's lots of time for discussion on how/whether
it should be done.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Announce: JSON is in the trunk

[Nolan Darilek]

Cool! Is it documented anywhere yet?

If not, how would I go about retrieving the latest X events from the 
timeline? Ideally I'd like to add a "Latest news" section on my homepage 
as rendered HTML, but in the meantime I'll implement that with JQuery.



On 11/04/2011 04:14 PM, Stephan Beal wrote:

Hi, all!

In a surprising (to me) turn of events:

http://www.fossil-scm.org/fossil/info/796dcfe072

the JSON API is now in the trunk and the json branch has been closed. 
The JSON features are disabled by default and can be enabled by 
passing --enable-json to the configure script. (Without --enable-json 
the related files will still be compiled but they will be effectively 
empty.) It "should" build on Windows and that other popular OS, and is 
known to build on a variety of Unixes/Unicis/Unixis/whatever. Please 
feed back any build difficulties to either fossil-user or fossil-dev 
(the latter, please, if you're already subscribed there).


And since i'm already announcing, here's another one: Caleb Gray has 
joined the project to assist in the JSON effort, and he's already got 
his own JSON Ajax demo online.


Have fun and Happy Hacking!

--
- stephan beal
http://wanderinghorse.net/home/stephan/


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Nolan Darilek]

Just a heads-up:

I did a quick test, and this didn't seem to work:

* Created a new Fossil repository
* Opened it into ./fossil, added README.txt with contents "foo".
* Changed to parent directory, created Git repository in ./git
* Imported Fossil repo and switched to trunk branch
* Modified README.txt to contain "foo\nbar" and committed to Git
* Ran "git fast-export --all |fossil import --incremental ../test.fossil"
* Then ran "cd ../fossil; fossil up"

The result is that my initial commit of README.txt into Fossil appears 
in the timeline twice. The Git commit appears, but "fossil up" won't 
update to it, at least not without being explicitly asked. If you'd like 
to see the resulting repository, it's here:


http://dl.dropbox.com/u/147071/test.fossil

Sorry I didn't test in the first place, I thought that the answer would 
be something like "yes of course that works, that's what --incremental 
does, what a silly question." :) I guess I'll just use Fossil until I 
get pushback, then try promoting it and, if that fails, just export to Git.


Thanks.


On 11/04/2011 03:35 PM, Michael Barrow wrote:
Apologies in advance if this makes no sense. I've only done a tiny bit 
with git and that was some time ago.


What if you have a single directory that both version control systems 
use? You pull from git, commit to Fossil, do your changes and commit 
to Fossil, and then push your changes to git. When you do the new pull 
from git, you just update Fossil and start the cycle again.



--
Michael L. Barrow

On Nov 4, 2011, at 10:54, Nolan Darilek > wrote:



Thanks, but that's not really what I asked.

I totally get Fossil's development model, have used it for over a 
year and think that it'd be a great fit for this particular 
community. I also read this message when it was originally posted. 
But I may be working with people who would rather submit a quick 
change via Github rather than download and install a new piece of 
software. Yes, I get that it's easy, I'm just thinking that it might 
be a barrier here. So let me make the question more explicit:


1. Can I export my project to a Git repository, push that to Github, 
make a few commits, export the changes to Git and push the repository 
again? If so, will it look identical to the repository after the 
first step with a few extra commits such that someone who pulls 
doesn't get told that the repository after the changes is different?


2. If my canonical Fossil repository advances and someone makes 
changes on Github, can I do an incremental import of the Git 
repository and only get their changes without creating an entirely 
new Fossil repository?


3. Has anyone else done this, and how does it work? I'd really rather 
use Fossil, but am worried about losing contributors who don't want 
to learn a new and simpler system. Since we're developing 
applications to meet immediate needs (software for the various 
occupations), the response I may get from people might be to use the 
tools that everyone knows to maximize the community's 
ability/willingness to chip in. If that response comes, hopefully I 
can say that Fossil interacts with Git in the manner I've described 
here and can meet the need.


Thanks.


On 11/04/2011 12:06 PM, Stephan Beal wrote:
On Fri, Nov 4, 2011 at 5:43 PM, Nolan Darilek 
mailto:no...@thewordnerd.info>> wrote:


some pushback from Git users. Is it possible to use Fossil in a
workflow with people who would rather use Git/Github?


Richard wrote a nice summary of that on Oct 16th which i'll paste in 
here:


---
Fossil does not currently support a hierarchical development model 
very well.  It wants everybody to be a peer.  It wants all 
developers to see everything all the time.  Fossil strives to avoid 
a "peeking order" in which some developers are hidden from view 
behind "lieutenants".  This is a more egalitarian model, but also 
one that does not scale as well.


To better support a hierarchy, Fossil would need the ability to sync 
individual branches in addition to its current behavior of always 
syncing everything on every sync request. (Recall that I asked for 
volunteers to implement such a thing a while back.)  But adding that 
feature quickly gets complicated when you then try to figure out how 
to deal with auto-sync.  You could, I suppose, put your local Fossil 
into a mode where it only syncs the branch you are currently working 
on or switching to.  But what about Wiki and Tickets and Events?  Do 
they get synced or not?  Once you leave the comfort of Fossils 
original model of "everybody sees all the code all the time" then 
various operational questions of this kind start to come up.

---

--
- stephan beal
http://wanderinghorse.net/home/stephan/


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


_

Re: [fossil-users] Fossil and Git joint projects?

[Alek Paunov]

On 04.11.2011 22:32, Andreas Kupries wrote:

Does hg have something git's rebase and other history rewriting
operations ?


Yes - hg supports all of them and more (via various bundled [1] and 
endless count of community extensions), but as you know, history 
rewriting kills any repo/repo synchronization in any DVCS. Thus, 
thinking of the partial Fossil/Git bridge, let's focus on ideal case - 
no rebases, full DAG synchronization.




If not, how do hg and the synchronizer (dulwich?) cope when the git
repository they talk to undergoes such a rewrite ?


Dulwich is not a merge-application - dulwich is just a well designed Git 
API (protocol and container format).


In the case of Mercurial/Git bridge, the actual synchronizer is the 
hg-git itself (mercurial extension), which was started as common effort 
with github.com, but these days is actually maintained by hg-subversion 
team (with Augie Fackler [2] as lead)


[1] http://mercurial.selenic.com/wiki/RebaseExtension
[2] https://bitbucket.org/durin42/hg-git/changesets
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] Announce: JSON is in the trunk

[Stephan Beal]

Hi, all!

In a surprising (to me) turn of events:

http://www.fossil-scm.org/fossil/info/796dcfe072

the JSON API is now in the trunk and the json branch has been closed. The
JSON features are disabled by default and can be enabled by passing
--enable-json to the configure script. (Without --enable-json the related
files will still be compiled but they will be effectively empty.) It
"should" build on Windows and that other popular OS, and is known to build
on a variety of Unixes/Unicis/Unixis/whatever. Please feed back any build
difficulties to either fossil-user or fossil-dev (the latter, please, if
you're already subscribed there).

And since i'm already announcing, here's another one: Caleb Gray has joined
the project to assist in the JSON effort, and he's already got his own JSON
Ajax demo online.

Have fun and Happy Hacking!

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Andreas Kupries]

On 11/4/2011 11:48 AM, Alek Paunov wrote:

Hi Nolan, Stephan,

I think that this is doable for a restricted subclass of usage scenarios (these
conforming workflow style cited by Stephan above + bunch of other restrictions).

We can benefit the fantastic opportunity of sqlite as fossil artifacts
container - I think that it is possible to be implemented git protocol enabled
python server with dulwich [1] on top of the fossil db, once we are able to
manage fossil/git commit hashes dictionary in additional db table. Hg-git [2]
also uses similar approach (but client side, It seems to me that server side is
the easier one).


Does hg have something git's rebase and other history rewriting operations ?

If not, how do hg and the synchronizer (dulwich?) cope when the git repository 
they talk to undergoes such a rewrite ?


This claim of 'lossless conversion of changesets' is certainly something of 
interest.


Hm ... given that "This plugin was originally developed by the folks at 
GitHub", maybe we can interest them in doing similar for fossil ?




As Fossil API, for the proof of the concept, the server can use the JSON API,
Stephan ?

Kind Regards,
Alek

[1] http://git.samba.org/?p=jelmer/dulwich.git
[2] http://hg-git.github.com/

P.S. github != git, maybe tickets and wikis can be synchronized too (to some
level) using the gihub API.




--
Andreas Kupries
Senior Tcl Developer
ActiveState, The Dynamic Language Experts

P: 778.786.1122
F: 778.786.1133
andre...@activestate.com
http://www.activestate.com
Get insights on Open Source and Dynamic Languages at www.activestate.com/blog
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Michael Barrow]

Apologies in advance if this makes no sense. I've only done a tiny bit with git 
and that was some time ago.

What if you have a single directory that both version control systems use? You 
pull from git, commit to Fossil, do your changes and commit to Fossil, and then 
push your changes to git. When you do the new pull from git, you just update 
Fossil and start the cycle again.


--
Michael L. Barrow

On Nov 4, 2011, at 10:54, Nolan Darilek  wrote:

> Thanks, but that's not really what I asked.
> 
> I totally get Fossil's development model, have used it for over a year and 
> think that it'd be a great fit for this particular community. I also read 
> this message when it was originally posted. But I may be working with people 
> who would rather submit a quick change via Github rather than download and 
> install a new piece of software. Yes, I get that it's easy, I'm just thinking 
> that it might be a barrier here. So let me make the question more explicit:
> 
> 1. Can I export my project to a Git repository, push that to Github, make a 
> few commits, export the changes to Git and push the repository again? If so, 
> will it look identical to the repository after the first step with a few 
> extra commits such that someone who pulls doesn't get told that the 
> repository after the changes is different?
> 
> 2. If my canonical Fossil repository advances and someone makes changes on 
> Github, can I do an incremental import of the Git repository and only get 
> their changes without creating an entirely new Fossil repository?
> 
> 3. Has anyone else done this, and how does it work? I'd really rather use 
> Fossil, but am worried about losing contributors who don't want to learn a 
> new and simpler system. Since we're developing applications to meet immediate 
> needs (software for the various occupations), the response I may get from 
> people might be to use the tools that everyone knows to maximize the 
> community's ability/willingness to chip in. If that response comes, hopefully 
> I can say that Fossil interacts with Git in the manner I've described here 
> and can meet the need.
> 
> Thanks.
> 
> 
> On 11/04/2011 12:06 PM, Stephan Beal wrote:
>> 
>> On Fri, Nov 4, 2011 at 5:43 PM, Nolan Darilek  wrote:
>> some pushback from Git users. Is it possible to use Fossil in a workflow 
>> with people who would rather use Git/Github?
>> 
>> Richard wrote a nice summary of that on Oct 16th which i'll paste in here:
>> 
>> ---
>> Fossil does not currently support a hierarchical development model very 
>> well.  It wants everybody to be a peer.  It wants all developers to see 
>> everything all the time.  Fossil strives to avoid a "peeking order" in which 
>> some developers are hidden from view behind "lieutenants".  This is a more 
>> egalitarian model, but also one that does not scale as well.
>> 
>> To better support a hierarchy, Fossil would need the ability to sync 
>> individual branches in addition to its current behavior of always syncing 
>> everything on every sync request. (Recall that I asked for volunteers to 
>> implement such a thing a while back.)  But adding that feature quickly gets 
>> complicated when you then try to figure out how to deal with auto-sync.  You 
>> could, I suppose, put your local Fossil into a mode where it only syncs the 
>> branch you are currently working on or switching to.  But what about Wiki 
>> and Tickets and Events?  Do they get synced or not?  Once you leave the 
>> comfort of Fossils original model of "everybody sees all the code all the 
>> time" then various operational questions of this kind start to come up.
>> ---
>> 
>> -- 
>> - stephan beal
>> http://wanderinghorse.net/home/stephan/
>> 
>> 
>> ___
>> fossil-users mailing list
>> fossil-users@lists.fossil-scm.org
>> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
> 
> ___
> fossil-users mailing list
> fossil-users@lists.fossil-scm.org
> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Alek Paunov]

On 04.11.2011 21:01, Stephan Beal wrote:

The biggest catch there is binary data, since JSON has no native way of
handing it. However, my current thinking on binary data is this...

from JSON we serve URL paths which, which called by an HTTP client, will
return the raw content (binary or not) for the given artifact. We could
handle POSTed data the same way, _theoretically_. There is a bit of
proof-of-concept for that in the current timeline output:


This approach (binaries via URL "pointers" on get/direct POST payloads 
on send) seems OK to me ...

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Stephan Beal]

On Fri, Nov 4, 2011 at 7:48 PM, Alek Paunov  wrote:

> As Fossil API, for the proof of the concept, the server can use the JSON
> API, Stephan ?
>
>
The biggest catch there is binary data, since JSON has no native way of
handing it. However, my current thinking on binary data is this...

from JSON we serve URL paths which, which called by an HTTP client, will
return the raw content (binary or not) for the given artifact. We could
handle POSTed data the same way, _theoretically_. There is a bit of
proof-of-concept for that in the current timeline output:

...
"files":[{
"name":"src/json_user.c",
"uuid":"d860ffd974654d397cf3be9a39ec1eb99ea472b2",
"prevUuid":"4b9322d63bad1b18bb3fd24e435277fc852db7d6",
"state":"modified",

"downloadPath":"/raw/src/json_user.c?name=d860ffd974654d397cf3be9a39ec1eb99ea472b2"
  }]


-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Alek Paunov]

Hi Nolan, Stephan,

I think that this is doable for a restricted subclass of usage scenarios 
(these conforming workflow style cited by Stephan above +  bunch of 
other restrictions).


We can benefit the fantastic opportunity of sqlite as fossil artifacts 
container - I think that it is possible to be implemented git protocol 
enabled python server with dulwich [1] on top of the fossil db, once we 
are able to manage fossil/git commit hashes dictionary in additional db 
table. Hg-git [2] also uses similar approach (but client side, It seems 
to me that server side is the easier one).


As Fossil API, for the proof of the concept, the server can use the JSON 
API, Stephan ?


Kind Regards,
Alek

[1] http://git.samba.org/?p=jelmer/dulwich.git
[2] http://hg-git.github.com/

P.S. github != git, maybe tickets and wikis can be synchronized too (to 
some level) using the gihub API.



On 04.11.2011 19:54, Nolan Darilek wrote:

Thanks, but that's not really what I asked.

I totally get Fossil's development model, have used it for over a year
and think that it'd be a great fit for this particular community. I also
read this message when it was originally posted. But I may be working
with people who would rather submit a quick change via Github rather
than download and install a new piece of software. Yes, I get that it's
easy, I'm just thinking that it might be a barrier here. So let me make
the question more explicit:

1. Can I export my project to a Git repository, push that to Github,
make a few commits, export the changes to Git and push the repository
again? If so, will it look identical to the repository after the first
step with a few extra commits such that someone who pulls doesn't get
told that the repository after the changes is different?

2. If my canonical Fossil repository advances and someone makes changes
on Github, can I do an incremental import of the Git repository and only
get their changes without creating an entirely new Fossil repository?

3. Has anyone else done this, and how does it work? I'd really rather
use Fossil, but am worried about losing contributors who don't want to
learn a new and simpler system. Since we're developing applications to
meet immediate needs (software for the various occupations), the
response I may get from people might be to use the tools that everyone
knows to maximize the community's ability/willingness to chip in. If
that response comes, hopefully I can say that Fossil interacts with Git
in the manner I've described here and can meet the need.

Thanks.


On 11/04/2011 12:06 PM, Stephan Beal wrote:

On Fri, Nov 4, 2011 at 5:43 PM, Nolan Darilek mailto:no...@thewordnerd.info>> wrote:

some pushback from Git users. Is it possible to use Fossil in a
workflow with people who would rather use Git/Github?


Richard wrote a nice summary of that on Oct 16th which i'll paste in
here:

---
Fossil does not currently support a hierarchical development model
very well. It wants everybody to be a peer. It wants all developers to
see everything all the time. Fossil strives to avoid a "peeking order"
in which some developers are hidden from view behind "lieutenants".
This is a more egalitarian model, but also one that does not scale as
well.

To better support a hierarchy, Fossil would need the ability to sync
individual branches in addition to its current behavior of always
syncing everything on every sync request. (Recall that I asked for
volunteers to implement such a thing a while back.) But adding that
feature quickly gets complicated when you then try to figure out how
to deal with auto-sync. You could, I suppose, put your local Fossil
into a mode where it only syncs the branch you are currently working
on or switching to. But what about Wiki and Tickets and Events? Do
they get synced or not? Once you leave the comfort of Fossils original
model of "everybody sees all the code all the time" then various
operational questions of this kind start to come up.
---

--
- stephan beal
http://wanderinghorse.net/home/stephan/


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users





___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Nolan Darilek]

Thanks, but that's not really what I asked.

I totally get Fossil's development model, have used it for over a year 
and think that it'd be a great fit for this particular community. I also 
read this message when it was originally posted. But I may be working 
with people who would rather submit a quick change via Github rather 
than download and install a new piece of software. Yes, I get that it's 
easy, I'm just thinking that it might be a barrier here. So let me make 
the question more explicit:


1. Can I export my project to a Git repository, push that to Github, 
make a few commits, export the changes to Git and push the repository 
again? If so, will it look identical to the repository after the first 
step with a few extra commits such that someone who pulls doesn't get 
told that the repository after the changes is different?


2. If my canonical Fossil repository advances and someone makes changes 
on Github, can I do an incremental import of the Git repository and only 
get their changes without creating an entirely new Fossil repository?


3. Has anyone else done this, and how does it work? I'd really rather 
use Fossil, but am worried about losing contributors who don't want to 
learn a new and simpler system. Since we're developing applications to 
meet immediate needs (software for the various occupations), the 
response I may get from people might be to use the tools that everyone 
knows to maximize the community's ability/willingness to chip in. If 
that response comes, hopefully I can say that Fossil interacts with Git 
in the manner I've described here and can meet the need.


Thanks.


On 11/04/2011 12:06 PM, Stephan Beal wrote:
On Fri, Nov 4, 2011 at 5:43 PM, Nolan Darilek > wrote:


some pushback from Git users. Is it possible to use Fossil in a
workflow with people who would rather use Git/Github?


Richard wrote a nice summary of that on Oct 16th which i'll paste in here:

---
Fossil does not currently support a hierarchical development model 
very well.  It wants everybody to be a peer.  It wants all developers 
to see everything all the time.  Fossil strives to avoid a "peeking 
order" in which some developers are hidden from view behind 
"lieutenants".  This is a more egalitarian model, but also one that 
does not scale as well.


To better support a hierarchy, Fossil would need the ability to sync 
individual branches in addition to its current behavior of always 
syncing everything on every sync request. (Recall that I asked for 
volunteers to implement such a thing a while back.)  But adding that 
feature quickly gets complicated when you then try to figure out how 
to deal with auto-sync.  You could, I suppose, put your local Fossil 
into a mode where it only syncs the branch you are currently working 
on or switching to.  But what about Wiki and Tickets and Events?  Do 
they get synced or not?  Once you leave the comfort of Fossils 
original model of "everybody sees all the code all the time" then 
various operational questions of this kind start to come up.

---

--
- stephan beal
http://wanderinghorse.net/home/stephan/


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] Fossil and Git joint projects?

[Stephan Beal]

On Fri, Nov 4, 2011 at 5:43 PM, Nolan Darilek wrote:

> some pushback from Git users. Is it possible to use Fossil in a workflow
> with people who would rather use Git/Github?


Richard wrote a nice summary of that on Oct 16th which i'll paste in here:

---
Fossil does not currently support a hierarchical development model very
well.  It wants everybody to be a peer.  It wants all developers to see
everything all the time.  Fossil strives to avoid a "peeking order" in
which some developers are hidden from view behind "lieutenants".  This is a
more egalitarian model, but also one that does not scale as well.

To better support a hierarchy, Fossil would need the ability to sync
individual branches in addition to its current behavior of always syncing
everything on every sync request. (Recall that I asked for volunteers to
implement such a thing a while back.)  But adding that feature quickly gets
complicated when you then try to figure out how to deal with auto-sync.
 You could, I suppose, put your local Fossil into a mode where it only
syncs the branch you are currently working on or switching to.  But what
about Wiki and Tickets and Events?  Do they get synced or not?  Once you
leave the comfort of Fossils original model of "everybody sees all the code
all the time" then various operational questions of this kind start to come
up.
---

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] Fossil and Git joint projects?

[Nolan Darilek]

Hey folks.

So far, most of my Fossil use has been on projects I own almost 
exclusively. However, I'm about to start working on projects with a 
community, and while I'm going to push Fossil adoption hard, I may get 
some pushback from Git users. Is it possible to use Fossil in a workflow 
with people who would rather use Git/Github? IOW, can I make Fossil my 
official project source but periodically push updates to Github, then 
import the results of people's contributions into the official repo?


I see an incremental option for import. I'm guessing this helps with 
that? Can I just incrementally import the latest changes from Git 
repositories? I'm guessing this just adds in changesets from Git that 
aren't present in Fossil?


I don't see a symmetric option in the export command. Do I just have to 
keep exporting the full repository each time? I'm guessing that the 
exports should be identical such that when I go to push the new export 
to Github, that it  won't look like an entirely new repo.


Hoping that I can push Fossil adoption for these projects, but I'm 
worried that I'll have to use the lowest common denominators in these 
communities, and in this case that will be Git. In that case, I'd really 
rather build the project/site in Fossil, then win people over when they 
realize that their Fossil clone of the project includes the wiki, 
ticketing system, etc. and their Git pulls don't. :)

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

fossil-users@lists.fossil-scm.org

[Tomek Kott]

Hi folks,

I just ran into the following problem. I have a wiki page that is titled
"DC Contact & Sample". When I go through and browse from the home page, I
get the link "http://127.0.0.1:8080/wiki?name=DC+-+Contact+%26amp%3B+Sample";
(i.e., I wrote [DC Contact & Sample] in the wiki editor).

However, on the timeline, the link is "
http://127.0.0.1:8080/wiki?name=DC+-+Contact+%26amp%3Bamp%3B+Sample";. If
you notice, there is an extra ";amp" that sneaks into the url.

Additionally, the $title variable for displaying the wiki page also shows:
"DC - Contact & Sample"(i.e., the HTML code is: DC - Contact
& Sample)

This is using fossil 1.20 (a75e2d2504).

Thanks!

Tomek
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Richard Hipp]

On Fri, Nov 4, 2011 at 5:31 AM, Zeev Pekar  wrote:

> 2) is it possible to deny access to folders of the source tree to
> certain users and grant access to the rest of the folders? (like in
> gitolite)
>
>
With Fossil (or git or hg or mtn or bzr) each user has their own complete
copy of the repository on their local machine.  And so each user can do
anything they want to their local copy of the repo and there is nothing you
the administrator can do to prevent it.  That's the nature of DVCS.

Gitolite says it provides per-directory "write" permissions on the central
repo.  "read" permission is always per-repo.  I presume they do this by
denying push to repos that have modified directories that they are
forbidden to modify.  Or perhaps they are playing some really nasty games
with rebase.  But probably the former.

Fossil does not do that and I have precious little motivation to make it do
that.

Fossil strives to be low-ceremony.  By that I mean that you don't need a
lot of central administrator setup and permission in order to participate
in a Fossil-hosted project.  The idea is that you trust your developers to
do the right thing, to follow the rules you have created for your project,
and to not modify files which they are not permitted to modify.  This is
not an unreasonable idea, since if you cannot trust your developers then
you have way more serious problems.

To complement its low-ceremony operating principle, Fossil also maintains a
detailed and immutable audit trail and tools to detect and reverse
undesirable changes and promote situational awareness.  So if a developer
does break the rules by checking into the wrong directory, he will be found
out, the changes can be backed out, and the rogue developer can be dealt
with administratively.

I have used both low-ceremony and high-ceremony systems and I have found
that the low-ceremony designs yield less frustration and greater
productivity, which is why Fossil is designed to be low-ceremony and why I
am so unmotivated to give it high-ceremony features.

-- 
D. Richard Hipp
d...@sqlite.org
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Konstantin Khomoutov]

On Fri, 04 Nov 2011 13:52:04 +0200
Zeev Pekar  wrote:

[...]
> > It would be impossible to implement within fossil's world view.
> > Once i clone a repo i have the whole thing, which i can then
> > manipulate (with admin-level rights) on my machine - you cannot
> > stop me from checking out a given path once i have access to a
> > clone of the repo. There is no mechanism with which fossil can
> > completely protect subdirectories unless it explicitly supports
> > "partial clones" (pulling only part of the repo).
> 
> Ok, maybe implement "partial clones"?
I think the way to go would be to have something resembling Git's
submodules.  That is, keep each "subproject" in a separate repo and
have a way to the commits in the "master" ("intergating" or whatever)
repo to reference specific commits in those satellite repos
(submodules).
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Stephan Beal]

On Fri, Nov 4, 2011 at 12:52 PM, Zeev Pekar  wrote:

> **
>
> Ok, maybe implement "partial clones"?
>
>  If you have a partial clone, fossil could not guaranty semantic
consistency, could not accurately follow timeline information, etc. Any
file in subdir a/b/c could refer to another file in subtree d/e/f, If you
then restrict access to d/e/f, then any partial clone is useless because
the parts we have access to (d/e/f) require parts which we do not have
(a/b/c). A partial history is useless - i cannot imagine how fossil could
possibly function with a partial history, If it was asked to show the
history for artifact abcdabcd and it cannot find it in a partial clone then
fossil cannot know if the artifact does not exist at all or if it exists in
a restricted subdir. And those are just the points which immediately come
to mind. Those more familiar with the internal structures will certainly
also see cases which such a feature would break.

The size of the can of worms such a feature would open is... well, it's
really darned big.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Zeev Pekar]

- Original message -
> On Fri, Nov 4, 2011 at 12:25 PM, Zeev Pekar  wrote:
> 
> > **
> > 
> > I do miss this feature... :( What is the best workaround for this
> > today?
> > 
> 
> To use a DVCS which supports this. i'm not personally aware of any.
> 
gitolite supports ACL on paths if I'm not mistaken...

> 
> > To developers: do you think this is important enough for you to
> > implement it one day?
> > 
> 
> It would be impossible to implement within fossil's world view. Once i
> clone a repo i have the whole thing, which i can then manipulate (with
> admin-level rights) on my machine - you cannot stop me from checking out
> a given path once i have access to a clone of the repo. There is no
> mechanism with which fossil can completely protect subdirectories unless
> it explicitly supports "partial clones" (pulling only part of the repo).

Ok, maybe implement "partial clones"?
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Stephan Beal]

On Fri, Nov 4, 2011 at 12:25 PM, Zeev Pekar  wrote:

> **
>
> I do miss this feature... :( What is the best workaround for this today?
>

To use a DVCS which supports this. i'm not personally aware of any.


> To developers: do you think this is important enough for you to implement
> it one day?
>

It would be impossible to implement within fossil's world view. Once i
clone a repo i have the whole thing, which i can then manipulate (with
admin-level rights) on my machine - you cannot stop me from checking out a
given path once i have access to a clone of the repo. There is no mechanism
with which fossil can completely protect subdirectories unless it
explicitly supports "partial clones" (pulling only part of the repo).

If you mean to restrict _checkin_ access to a given subdir, i don't know if
fossil _could_ do that, but it doesn't currently (and, IMO, shouldn't).

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Zeev Pekar]

> > 2) is it possible to deny access to folders of the source tree to
> > certain users and grant access to the rest of the folders? (like in
> > gitolite)
> 
> There is no ACL on path names.

I do miss this feature... :( What is the best workaround for this today?

To developers: do you think this is important enough for you to implement it 
one day?

> 
> > ps: I've asked Anjuta developers to create a fossil plug-in and they
> > agreed :)
> > https://bugzilla.gnome.org/show_bug.cgi?id=663313
> 
> Nice!

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Alek Paunov]

On 04.11.2011 11:45, Stephan Beal wrote:

@Everyone else: please correct that if it's wrong.


No need of correction, I use fossil this way too:

cat /etc/systemd/system/fossil.service

[Unit]
Description=Fossil Repositories
After=network.target

[Service]
Type=simple
User=fossil
Group=src
ExecStart=/usr/bin/fossil server --port 39127 /srv/source/fossil

[Install]
WantedBy=multi-user.target

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Joerg Sonnenberger]

On Fri, Nov 04, 2011 at 11:31:26AM +0200, Zeev Pekar wrote:
> 1) if I have 2 projects do I have to a) create 2 separate repositories
> or do I b) put both source trees in one repository? if a), how do I run
> fossil server to make both repositories accessible at the same time from
> outside?

Create two separate repositories. If you put them in a directory and end
the name in .fossil, you can point the server to the directory and serve
both.

> 2) is it possible to deny access to folders of the source tree to
> certain users and grant access to the rest of the folders? (like in
> gitolite)

There is no ACL on path names.

> ps: I've asked Anjuta developers to create a fossil plug-in and they
> agreed :)
> https://bugzilla.gnome.org/show_bug.cgi?id=663313

Nice!

Joerg
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Re: [fossil-users] 2 questions

[Stephan Beal]

On Fri, Nov 4, 2011 at 10:31 AM, Zeev Pekar  wrote:

> 1) if I have 2 projects do I have to a) create 2 separate repositories
> or do I b) put both source trees in one repository?


It's personal personal preference. i keep everything separate (which means
have 50 little fossil repos) while some prefer one mega-repo for everything.


> if a), how do I run
> fossil server to make both repositories accessible at the same time from
> outside?
>

If i'm not mistaken (but i have never tried this):

- name all of your fossil files *.fossil and put them in one directory
- run 'fossil server .' from that directory

@Everyone else: please correct that if it's wrong.


> 2) is it possible to deny access to folders of the source tree to
> certain users and grant access to the rest of the folders? (like in
> gitolite)
>

Not within a single repo. Access rights are per-repo, so you can
selectively give users access only to a whole repo at once, not specific
directories of a project.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] 2 questions

[Zeev Pekar]

Hi,

I have 2 questions regarding fossil:

1) if I have 2 projects do I have to a) create 2 separate repositories
or do I b) put both source trees in one repository? if a), how do I run
fossil server to make both repositories accessible at the same time from
outside?

2) is it possible to deny access to folders of the source tree to
certain users and grant access to the rest of the folders? (like in
gitolite)

thank you
Zeev

ps: I've asked Anjuta developers to create a fossil plug-in and they
agreed :)
https://bugzilla.gnome.org/show_bug.cgi?id=663313




___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users