naviserver-devel — Developer discussion

[naviserver-devel] Is this list still alive?

[Ian H. <har...@gm...>]

Haven't seen any messages for a while....

Hope everyone is well!

- Ian

[naviserver-devel] Even potential GSoC mentors/reviewers should sign into Melange

[Tomasz K. <tk...@da...>]

Hi!

If you've been contacted by a student in private about Google Summer
of Code or you have contributed to our GSoC 2009 idea list *please*
sign in to Melange (web app prepared by Google to handle GSoC 2009).
There we mentors can put private comments on student applications,
score them, etc. It is desirable that TCT members or Tcl/Tk
maintainers sign-in into Melange to be able to influence which
projects will be funded.

How to do it:

1. Make sure you have an GMail account (you can forward it easily to
you normal e-mail account in Settings so you don't have to use it to
read e-mail).
2. Sign-in with you GMail account to http://socghop.appspot.com
3. Choose LinkID (your screen name in Melange) to be somehow related
to your first+last name so it will be easy to identify you
4. Apply to be a mentor (last item in the left menu) for Tcl Community
Association - it is under this link
http://socghop.appspot.com/org/apply_mentor/google/gsoc2009

Do this even if you do not plan to mentor but you are willing to help
us refine the scope of projects from your field of expertise.

For me this is a declaration that we should add you to private e-mail
discussion about students applications.

--tkosiak

[naviserver-devel] [Tcl/Tk Community] We need *urgent* advertisement among CS students about Tcl in GSoC 2009

[Tomasz K. <tk...@gm...>]

Hi! I'm Tomasz Kosiak.

This year I have an honor to officially serve as an administrator for
Tcl/Tk Community involvement in Google Summer of Code 2009 program. I
would like to ask all broad Tcl community members to help us reach
students worldwide and encourage them to apply for a summer
scholarship.

If you have never heard of Google Summer of Code (aka GSoC) it is a
Google, Inc. sponsored initiative to fund aproximately 1000 student
scholarship $4500 each for coding during a summer for ideas freely
chosen by approved open-source organization. Tcl/Tk community is
approved again by Google this year (second time).

We seek opportunities to advertise among CS students our GSoC project
ideas. That is why I have prepared one slide info about Tcl/Tk and
GSoC. This e-mail can be freely forwarded directly to prospective
students or to your fellow university teachers so they can show this
slide for their students. The goal is to reach students worldwide
willing to apply to Tcl/Tk as mentoring organization.

See Tcl/Tk GSoC 2009 in a pill at http://tinyurl.com/tcl-gsoc2009 and
this is the only URL students shall write down to get started.

Official news about Tcl/Tk GSoC 2009 is at http://www.tcl.tk/gsoc/ and
http://wiki.tcl.tk/22182.

We, Tcl mentors and admins, are really willing to help students
prepare their applications. This is what our community is fond of.

As last year we *do* plan to act as umbrella organization for other
Tcl-related projects like Tk, XOTcl, Expect, AOLserver, Naviserver,
OpenACS, aMSN, Coccinella and others. We will try to lobby Google to
assure one student slot to those interested sub-communities. But we
need your ideas, mentors and promotion of our effort among students as
number of slots is mostly proportional to number of students
application received by an organization.

This is urgent as it would be best to reach students before the
weekend as student application deadline is set to April 3: 12 noon PDT
/ 19:00 UTC

Don't hesitate to contact me and Matthew Burke directly. Jeff Hobbs
(Tcl Guy  you all know) is also actively involved.

Regards and Happy Tcl'ing,
Tomasz Kosiak

PS. You can point students to other resources prepared by Google from
which I like the most the MS PowerPoint Slides from Philip Johnson's
presentation on Google Summer of Code at
http://csdl.ics.hawaii.edu/~johnson/screencasts/GSoC2009.ppt (see also
the video at http://csdl.ics.hawaii.edu/~johnson/screencasts/GSoC2009.mov).
These resource are both mentioned at official Google Blog in a post at
http://googleblog.blogspot.com/2009/03/supporting-students-in-open-source.html

Re: [naviserver-devel] DB Connection Handles nsdbipg

[Stephen D. <sd...@gm...>]

On Thu, Mar 26, 2009 at 3:09 PM, Ian Harding <har...@gm...> wrote:
> Please ignore, it is a permission issue.  The user had not rights in
> the schema, and since I used search_path, it didn't generate a
> permission denied error, it just didn't look there and generated a
> does not exist error.


Great.


FYI, there is some extra debug output available if you enable
server-wide debug logging. Each handle has an id and much of the log
output shows the id and the number of queries performed.  In this
case, you'd want to make sure that the same handle is being used for
both the SET command and the following query.

Re: [naviserver-devel] DB Connection Handles nsdbipg

[Ian H. <har...@gm...>]

Please ignore, it is a permission issue.  The user had not rights in
the schema, and since I used search_path, it didn't generate a
permission denied error, it just didn't look there and generated a
does not exist error.

Sorry about the noise.  Things seem to work as expected.

- Ian

On 3/26/09, Ian Harding <har...@gm...> wrote:
> On Wed, Mar 25, 2009 at 5:10 PM, Stephen Deasey <sd...@gm...> wrote:
>> On Wed, Mar 25, 2009 at 10:56 PM, Ian Harding <har...@gm...>
>> wrote:
>>> I am in a situation where I'd like to be able to set the search_path
>>> on each page request.  In AOLServer I would have just made a db handle
>>> getter function that would call ns_db gethandle, issue a quick SET
>>> command, and pass the handle back to the caller.
>
> <snip>
>
>> Alternatively, you could set the max number of handles to 0, in which
>> case the driver switches to handle-per-thread mode (See docs).  In
>> this case, when each command 'gets' a handle it will always get the
>> same one, as it is never actually returned to the pool but cached for
>> the thread.
>
> OK, so my maxhandles is 0, per the default, so I should always get the
> same handle in repeated calls.  I'm apparently not because when I
> issue
>
> dbi_dml {set search_path to test,public}
> dbi_rows {select * from tableintestbutnotpublic}
>
> I get an error stating that the table does not exist.
>
> Here's my postgres log
>
> LOG:  statement: set client_encoding to 'UTF8'
> LOG:  execute dbipg_0: set session client_encoding = 'UTF8'
> LOG:  execute dbipg_1: set session timezone = 'UTC'
> LOG:  execute dbipg_2: set session datestyle = 'ISO'
> LOG:  execute dbipg_3: set search_path to test,public
> ERROR:  relation "classification" does not exist
> STATEMENT:  SELECT classification
>                         FROM classification
>                         ORDER BY classification
>
> but it does exist...
>
> test=# \dt test.classification
>             List of relations
>  Schema |      Name      | Type  | Owner
> --------+----------------+-------+-------
>  test   | classification | table | user
> (1 row)
>
> So I wrote a test that goes like this
>
> if {[catch {
>
>         catch {dbi_dml {drop table test}}
>         catch {dbi_dml {drop schema test cascade}}
>
>     dbi_dml {
>         create schema test
>     }
>
>     dbi_dml {
>         create table test (
>             a integer not null,
>             b varchar not null,
>             c bytea,
>             test varchar
>         )
>     }
>     dbi_dml {insert into test (a, b, test) values (1, 'x', 'test1')}
>     dbi_dml {insert into test (a, b, test) values (2, 'y', 'test2')}
>
> } err]} {
>     ns_log error $err
> } else {
>     testConstraint table true
> }
>
>
> test schema-1 {bad search path} -constraints table -body {
>     dbi_dml {set search_path to test}
>     dbi_rows {select * from test}
> } -returnCodes error -result {relation "test" does not exist}
>
> test schema-2 {bad search path same thread} -constraints table -body {
>     dbi_rows {select * from test}
> } -returnCodes error -result {relation "test" does not exist}
>
> test schema-3 {good search path} -constraints table -body {
>     dbi_dml {set search_path to public}
>     dbi_rows {select * from test where a = 99}
> } -result {}
>
> Which passes when run with make test.
>
> So, what could be my problem here?  My config file is being used, and
> it does say maxhandles=0...
>
> ns_section      "ns/server/${server}/module/dbpg"
> ns_param   default       true
> ns_param   maxhandles    0
> ns_param   timeout       10
>
> - Ian
>

Re: [naviserver-devel] DB Connection Handles nsdbipg

[Ian H. <har...@gm...>]

On Wed, Mar 25, 2009 at 5:10 PM, Stephen Deasey <sd...@gm...> wrote:
> On Wed, Mar 25, 2009 at 10:56 PM, Ian Harding <har...@gm...> wrote:
>> I am in a situation where I'd like to be able to set the search_path
>> on each page request.  In AOLServer I would have just made a db handle
>> getter function that would call ns_db gethandle, issue a quick SET
>> command, and pass the handle back to the caller.

<snip>

> Alternatively, you could set the max number of handles to 0, in which
> case the driver switches to handle-per-thread mode (See docs).  In
> this case, when each command 'gets' a handle it will always get the
> same one, as it is never actually returned to the pool but cached for
> the thread.

OK, so my maxhandles is 0, per the default, so I should always get the
same handle in repeated calls.  I'm apparently not because when I
issue

dbi_dml {set search_path to test,public}
dbi_rows {select * from tableintestbutnotpublic}

I get an error stating that the table does not exist.

Here's my postgres log

LOG:  statement: set client_encoding to 'UTF8'
LOG:  execute dbipg_0: set session client_encoding = 'UTF8'
LOG:  execute dbipg_1: set session timezone = 'UTC'
LOG:  execute dbipg_2: set session datestyle = 'ISO'
LOG:  execute dbipg_3: set search_path to test,public
ERROR:  relation "classification" does not exist
STATEMENT:  SELECT classification
                        FROM classification
                        ORDER BY classification

but it does exist...

test=# \dt test.classification
            List of relations
 Schema |      Name      | Type  | Owner
--------+----------------+-------+-------
 test   | classification | table | user
(1 row)

So I wrote a test that goes like this

if {[catch {

        catch {dbi_dml {drop table test}}
        catch {dbi_dml {drop schema test cascade}}

    dbi_dml {
        create schema test
    }

    dbi_dml {
        create table test (
            a integer not null,
            b varchar not null,
            c bytea,
            test varchar
        )
    }
    dbi_dml {insert into test (a, b, test) values (1, 'x', 'test1')}
    dbi_dml {insert into test (a, b, test) values (2, 'y', 'test2')}

} err]} {
    ns_log error $err
} else {
    testConstraint table true
}


test schema-1 {bad search path} -constraints table -body {
    dbi_dml {set search_path to test}
    dbi_rows {select * from test}
} -returnCodes error -result {relation "test" does not exist}

test schema-2 {bad search path same thread} -constraints table -body {
    dbi_rows {select * from test}
} -returnCodes error -result {relation "test" does not exist}

test schema-3 {good search path} -constraints table -body {
    dbi_dml {set search_path to public}
    dbi_rows {select * from test where a = 99}
} -result {}

Which passes when run with make test.

So, what could be my problem here?  My config file is being used, and
it does say maxhandles=0...

ns_section      "ns/server/${server}/module/dbpg"
ns_param   default       true
ns_param   maxhandles    0
ns_param   timeout       10

- Ian

Re: [naviserver-devel] DB Connection Handles nsdbipg

[Ian H. <har...@gm...>]

On Wed, Mar 25, 2009 at 5:10 PM, Stephen Deasey <sd...@gm...> wrote:
> On Wed, Mar 25, 2009 at 10:56 PM, Ian Harding <har...@gm...> wrote:
>> I am in a situation where I'd like to be able to set the search_path
>> on each page request.  In AOLServer I would have just made a db handle
>> getter function that would call ns_db gethandle, issue a quick SET
>> command, and pass the handle back to the caller.
>>
>> With nsdbipg I don't know how to do this.  There doesn't seem to be an
>> explicit getting of a handle.  It seems that they are gotten
>> automatically on each call and released on each call to a nsdbi
>> function.
>
>
> Handles are managed automatically by default, but you can extend the
> lifetime with the eval command:
>
>    dbi_eval {
>        dbi_dml {set search_path to foo, public}
>        set users [dbi_rows {select * from users}]
>    }
>
> When you don't pass the -transaction switch, you are simply reusing
> the same handle.
>

Ah.  That works.

>
>> Maybe the datasource param can contain a schema setting, but even that
>> would not be what I want.  I want to be able to set the path ideally
>> on each connection, or on each call to the nsdbi functions.
>
>
> Do you want to set the path for the lifetime of the *database*
> connection, or are you trying to create 'virtual' users with one
> database pool, switching the schema search path for duration of the
> *http* connection?
>
> If it's the first, you could set the default path in the postgresql.conf file:
>
>    search_path = '$user, public'
>
> If it's the second, then maybe something like the above db_eval will
> work for you.
>
> You could wrap it up in a command like:
>
>    with_schema foo {
>        dbi_rows { ... }
>    }
>

Even better..

> Alternatively, you could set the max number of handles to 0, in which
> case the driver switches to handle-per-thread mode (See docs).  In
> this case, when each command 'gets' a handle it will always get the
> same one, as it is never actually returned to the pool but cached for
> the thread.  The idea behind this was as a performance optimisation
> for the case where pretty much all your conn threads perform queries
> and getting and putting is just overhead. But it would allow you in
> this case to set the schema path at the beginning of the http request
> and have it persist.
>

I think that's what I want.  All of my connections perform queries, so
if I set the search path first, it will stay in effect for the
lifetime of the connection.  I'll give that a shot.

Thanks!

- Ian

Re: [naviserver-devel] DB Connection Handles nsdbipg

[Stephen D. <sd...@gm...>]

On Wed, Mar 25, 2009 at 10:56 PM, Ian Harding <har...@gm...> wrote:
> I am in a situation where I'd like to be able to set the search_path
> on each page request.  In AOLServer I would have just made a db handle
> getter function that would call ns_db gethandle, issue a quick SET
> command, and pass the handle back to the caller.
>
> With nsdbipg I don't know how to do this.  There doesn't seem to be an
> explicit getting of a handle.  It seems that they are gotten
> automatically on each call and released on each call to a nsdbi
> function.


Handles are managed automatically by default, but you can extend the
lifetime with the eval command:

    dbi_eval {
        dbi_dml {set search_path to foo, public}
        set users [dbi_rows {select * from users}]
    }

When you don't pass the -transaction switch, you are simply reusing
the same handle.


> Maybe the datasource param can contain a schema setting, but even that
> would not be what I want.  I want to be able to set the path ideally
> on each connection, or on each call to the nsdbi functions.


Do you want to set the path for the lifetime of the *database*
connection, or are you trying to create 'virtual' users with one
database pool, switching the schema search path for duration of the
*http* connection?

If it's the first, you could set the default path in the postgresql.conf file:

    search_path = '$user, public'

If it's the second, then maybe something like the above db_eval will
work for you.

You could wrap it up in a command like:

    with_schema foo {
        dbi_rows { ... }
    }

Alternatively, you could set the max number of handles to 0, in which
case the driver switches to handle-per-thread mode (See docs).  In
this case, when each command 'gets' a handle it will always get the
same one, as it is never actually returned to the pool but cached for
the thread.  The idea behind this was as a performance optimisation
for the case where pretty much all your conn threads perform queries
and getting and putting is just overhead. But it would allow you in
this case to set the schema path at the beginning of the http request
and have it persist.

[naviserver-devel] DB Connection Handles nsdbipg

[Ian H. <har...@gm...>]

I am in a situation where I'd like to be able to set the search_path
on each page request.  In AOLServer I would have just made a db handle
getter function that would call ns_db gethandle, issue a quick SET
command, and pass the handle back to the caller.

With nsdbipg I don't know how to do this.  There doesn't seem to be an
explicit getting of a handle.  It seems that they are gotten
automatically on each call and released on each call to a nsdbi
function.

Maybe the datasource param can contain a schema setting, but even that
would not be what I want.  I want to be able to set the path ideally
on each connection, or on each call to the nsdbi functions.

Is there a way to do this?

Thanks!

- Ian

Re: [naviserver-devel] Buglets

[Ian H. <har...@gm...>]

On Mon, Feb 23, 2009 at 6:29 PM, Stephen Deasey <sd...@gm...> wrote:
> On Mon, Feb 23, 2009 at 10:12 PM, Ian Harding <har...@gm...> wrote:
>> Here are a couple things I've found during my migration from AOLServer
>> to Naviserver.
<snip>

>> 2.  dbipg does something strange when used with a template, and a
>> column the same name as the table.  This works as expected:
>>
>> default:nscp 1> dbi_rows {SELECT classification FROM classification}
>> UNCLASSIFIED
>>
>> This does not:
>>
>> default:nscp 2> dbi_rows {select classificationid, classification from
>> classification} {
>> default:nscp 2>>> $classificationid -> $classification
>> default:nscp 2>>> }
>>
>> 1 -> 1
>
>
> This works for me, if I'm reading your example right. I've added a
> test; can you try it out on your machine?
>
>    http://bitbucket.org/naviserver/nsdbipg/changeset/0fae5f9a2fb9/
>

OK, Nothing to see here.  My PostgreSQL  install was slightly broken.
/usr/lib/libpq.so was pointing to an older version.  All tests pass.
Funny that was the only bug that created...

Thanks!

- Ian

Re: [naviserver-devel] AOLserver/Naviserver in Tcl/Tk Google Summer of Code 2009 effort

[Tomasz K. <tk...@da...>]

The AOLserver project was:

AOLserver-GD integration: Further Rationalization of API, Sharing of
Graphics Among Server Threads, Building Graphing and Sparkline
Packages and Improving the Use of Autoconfig and SWIG
by Matthew Gagen, mentored by Matthew Burke

As far as I know the student was a univeristy fresman so most of the
efforts he put in understanding things in AOLserver & GD. Code
produced by him is available at
http://code.google.com/p/google-summer-of-code-2008-tcl/downloads/list,
but the problem was that there was not much accomplished to be
published (but Google required to publish something). During the
process it was decided that we won't do the work for a student. Anyway
the student did what he can and maybe there were some fault at the
mentoring organizatation side that we didn't manage to handle his lack
of skills (it was our first year). So in the end it was decided to
give him positive evaluation for the effort he put into the project
though result was quite dissatisfying.

That is how I understand things from my conversation with Metthew
during Tcl 2008 Conference. Matthew probably can give you more
details, but it seems he is extremely busy now with his paywork at the
university.

More about a GSoC impressions can be found at http://wiki.tcl.tk/21777
or in the presentations givein during the European and US Tcl/Tk
conferences.

Success ingerdients seems to be: advertising possibility to apply to
GSoC among students, proper student screening to choose promissing
students, project subject tailored for the promissing student and a
lot of communiction between mentor and a student. So we are looking
forward that maybe you can provide some of these to be more successful
this year..

Regards,
Tomasz Kosiak



On Fri, Mar 13, 2009 at 8:34 AM, Stephen Deasey <sd...@gm...> wrote:
> On Thu, Mar 12, 2009 at 11:06 PM, Tomasz Kosiak <tk...@gm...> wrote:
>> Hi!
>>
>> Is there someone interested to be a mentor from AOLserver/Naviserver
>> community within Tcl GSoC 2009 effort? Do you have any well defined
>> project ideas which could be written down at http://wiki.tcl.tk/22182.
>>
>> As Matthew mentioned last year Tcl acted as umbrella organization for
>> Tcl/Tk/AOLserver/OpenACS/XOTcl and we were given 9 slots which IMHO
>> shows that this is the right direction.
>>
>> We did one AOLserver and one XOTcl project which were somehow OpenACS related.
>
>
> What was the result of the AOLserver project?
>

Re: [naviserver-devel] AOLserver/Naviserver in Tcl/Tk Google Summer of Code 2009 effort

[Stephen D. <sd...@gm...>]

On Thu, Mar 12, 2009 at 11:06 PM, Tomasz Kosiak <tk...@gm...> wrote:
> Hi!
>
> Is there someone interested to be a mentor from AOLserver/Naviserver
> community within Tcl GSoC 2009 effort? Do you have any well defined
> project ideas which could be written down at http://wiki.tcl.tk/22182.
>
> As Matthew mentioned last year Tcl acted as umbrella organization for
> Tcl/Tk/AOLserver/OpenACS/XOTcl and we were given 9 slots which IMHO
> shows that this is the right direction.
>
> We did one AOLserver and one XOTcl project which were somehow OpenACS related.


What was the result of the AOLserver project?

[naviserver-devel] AOLserver/Naviserver in Tcl/Tk Google Summer of Code 2009 effort

[Tomasz K. <tk...@gm...>]

Hi!

Is there someone interested to be a mentor from AOLserver/Naviserver
community within Tcl GSoC 2009 effort? Do you have any well defined
project ideas which could be written down at http://wiki.tcl.tk/22182.

As Matthew mentioned last year Tcl acted as umbrella organization for
Tcl/Tk/AOLserver/OpenACS/XOTcl and we were given 9 slots which IMHO
shows that this is the right direction.

We did one AOLserver and one XOTcl project which were somehow OpenACS related.

Regards,
Tomasz Kosiak

Re: [naviserver-devel] Mercurial help

[Vlad S. <vl...@cr...>]

Uh, it is a little bit complicated, especially with history :-((

Stephen Deasey wrote:
> On Wed, Mar 11, 2009 at 5:09 PM, Vlad Seryakov <vl...@cr...> wrote:
>> yes, no history yet, i was doing baby steps
> 
> 
> Oh OK.  I'll let you play with it.
> 
> Note though that mercurial history is immutable. You will have to
> delete the repos and convert again.
> 
> 
>> I used my bitbucket name, it is vseryakov, on SF it was taken, so SF name is just seryakov :-((
> 
> 
> There's two names here: your bitbucket username, which mercurial
> doesn't know or care anything about, and the name you add to the
> commit log.
> 
> This last one can be absolutely anything you like: seryakov,
> vseryakov, sdeasey... mercurial enforces nothing. But by convention
> it's your name and email address, and you keep it consistent so it
> doesn't look like there's two of you :-)
> 
> Just add:
> 
>     [ui]
>     username=Vlad Seryakov <vlad@some.where>
> 
> to ~/.hgrc and forget about it.
> 
> 
> When converting repos from CVS you need to map the old-style SF
> username into a new-style mercurial name/email.
> 
> I did it like this:
> 
>     hg convert -A ~/tmp/sf.authors \
>        --config convert.hg.usebranchnames=0 \
>        --config convert.hg.clonebranches=1 \
>        ~/in/naviserver-modules-HEAD/nsudp/ nsudp-CVSROOT-hg
> 
> and in sf.authors are lines like:
> 
>     seryakov=Vlad Seryakov <vl...@cr...>
> 
> 
> 
>> Stephen Deasey wrote:
>>> On Wed, Mar 11, 2009 at 3:10 AM, Vlad Seryakov <vl...@cr...> wrote:
>>>> imported all
>>>>
>>>> i skipped nsffmpeg, nsotcl, nstcp, i will not going to support themand
>>>> they will not compile, so no point keeping broken things.
>>>>
>>>> nsocaml still may be of interest even it is not compiling as it is now.
>>>
>>> Hmm, they all just say 'Initial import', you didn't import the
>>> history. And you changed your name again, so there's about 3 'yous' on
>>> there now.
>>>
>>> I'll convert them -- there's only 9.
>>>
>>> ------------------------------------------------------------------------------
>>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>>> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
>>> software that enables intelligent coding and step-through debugging.
>>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>>> _______________________________________________
>>> naviserver-devel mailing list
>>> nav...@li...
>>> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>>>
>> ------------------------------------------------------------------------------
>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
>> software that enables intelligent coding and step-through debugging.
>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>> _______________________________________________
>> naviserver-devel mailing list
>> nav...@li...
>> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>>
> 
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Mercurial help

[Stephen D. <sd...@gm...>]

On Wed, Mar 11, 2009 at 5:09 PM, Vlad Seryakov <vl...@cr...> wrote:
> yes, no history yet, i was doing baby steps


Oh OK.  I'll let you play with it.

Note though that mercurial history is immutable. You will have to
delete the repos and convert again.


> I used my bitbucket name, it is vseryakov, on SF it was taken, so SF name is just seryakov :-((


There's two names here: your bitbucket username, which mercurial
doesn't know or care anything about, and the name you add to the
commit log.

This last one can be absolutely anything you like: seryakov,
vseryakov, sdeasey... mercurial enforces nothing. But by convention
it's your name and email address, and you keep it consistent so it
doesn't look like there's two of you :-)

Just add:

    [ui]
    username=Vlad Seryakov <vlad@some.where>

to ~/.hgrc and forget about it.


When converting repos from CVS you need to map the old-style SF
username into a new-style mercurial name/email.

I did it like this:

    hg convert -A ~/tmp/sf.authors \
       --config convert.hg.usebranchnames=0 \
       --config convert.hg.clonebranches=1 \
       ~/in/naviserver-modules-HEAD/nsudp/ nsudp-CVSROOT-hg

and in sf.authors are lines like:

    seryakov=Vlad Seryakov <vl...@cr...>



> Stephen Deasey wrote:
>> On Wed, Mar 11, 2009 at 3:10 AM, Vlad Seryakov <vl...@cr...> wrote:
>>> imported all
>>>
>>> i skipped nsffmpeg, nsotcl, nstcp, i will not going to support themand
>>> they will not compile, so no point keeping broken things.
>>>
>>> nsocaml still may be of interest even it is not compiling as it is now.
>>
>>
>> Hmm, they all just say 'Initial import', you didn't import the
>> history. And you changed your name again, so there's about 3 'yous' on
>> there now.
>>
>> I'll convert them -- there's only 9.
>>
>> ------------------------------------------------------------------------------
>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
>> software that enables intelligent coding and step-through debugging.
>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>> _______________________________________________
>> naviserver-devel mailing list
>> nav...@li...
>> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>>
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] Mercurial help

[Vlad S. <vl...@cr...>]

yes, no history yet, i was doing baby steps

I used my bitbucket name, it is vseryakov, on SF it was taken, so SF name is just seryakov :-((

Stephen Deasey wrote:
> On Wed, Mar 11, 2009 at 3:10 AM, Vlad Seryakov <vl...@cr...> wrote:
>> imported all
>>
>> i skipped nsffmpeg, nsotcl, nstcp, i will not going to support themand
>> they will not compile, so no point keeping broken things.
>>
>> nsocaml still may be of interest even it is not compiling as it is now.
> 
> 
> Hmm, they all just say 'Initial import', you didn't import the
> history. And you changed your name again, so there's about 3 'yous' on
> there now.
> 
> I'll convert them -- there's only 9.
> 
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Mercurial help

[Stephen D. <sd...@gm...>]

On Wed, Mar 11, 2009 at 3:10 AM, Vlad Seryakov <vl...@cr...> wrote:
> imported all
>
> i skipped nsffmpeg, nsotcl, nstcp, i will not going to support themand
> they will not compile, so no point keeping broken things.
>
> nsocaml still may be of interest even it is not compiling as it is now.


Hmm, they all just say 'Initial import', you didn't import the
history. And you changed your name again, so there's about 3 'yous' on
there now.

I'll convert them -- there's only 9.

Re: [naviserver-devel] Mercurial help

[Vlad S. <vl...@cr...>]

imported all

i skipped nsffmpeg, nsotcl, nstcp, i will not going to support themand 
they will not compile, so no point keeping broken things.

nsocaml still may be of interest even it is not compiling as it is now.


Stephen Deasey wrote:
> On Tue, Mar 10, 2009 at 4:14 PM, Vlad Seryakov <vl...@cr...> wrote:
>> Stephen,
>>
>> Can you describe how to import module into hg on bitbucket, it keeps saying i am not authorized.
> 
> 
> Log in as 'naviserver' at bitbucket.org.
> 
> Got to:
> 
>     http://bitbucket.org/repo/create
> 
> Fill out the info.  Once created, in the admin tab, on the right hand
> side under Permissions -> Administrators  add verseryakov and whoever
> else you'd like.
> 
> Back on your machine...
> 
> You can now clone the repo from bitbucket to your machine if you like,
> and commit straight into it -- the repo url will be set correctly for
> you.
> 
> As a shortcut, you can add something like this to your ~/.hgrc file:
> 
>     [paths]
>     nsfoo = https://vse...@bi.../naviserver/nsfoo/
> 
> this allows you to refer to the repo on bitbucket as 'nsfoo' on the
> command line. So, if you've converted one of the remaining repos and
> you want to push it, without first cloning from bitbucket, you just:
> 
>     cd ~/nsfoo-hg
>     hg outgoing nsfoo
>     hg push nsfoo
> 
> The 'outgoing' tells you what it would push without actually pushing.
> 
> If you edit the ~/nsfo-hg/.hg/hgrc to add:
> 
>     [paths]
>     default = http://bitbucket.org/naviserver/nsfoo
> 
> you don't have to tell it where you want to push to. Just:
> 
>     cd ~/nsfoo-hg
>     hg push
> 
> 
> Once it's up there, *then* you can customise the rest of the Admin
> options on bitbucket -- set up the commit emails etc. (look at
> naviserver for an example). Do this second so you don't spam the list
> with 5 years of history :-)
> 
> 
> I've converted some more modules.  Here's what's left to do:
> 
> nsconf
> nsexpat
> #nsstats
> nssys
> nstk
> nszlib
> 
> nsffmpeg
> nsfortune
> nsocaml
> nsotcl
> nssavi
> 
> nsdbext
> nsdbpd
> 
> 
> I know you said some where perhaps obsolete now, but what they hey,
> better not to loose them.
> 
> You could leave nsdbext and nsdbpd for me if you like. I was going to
> try merging in the whole aolserver history, which is a little more
> work.
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Mercurial help

[Vlad S. <vl...@cr...>]

Oh, login as naviserver was the trick,

thank you i will import all remaining modules

Stephen Deasey wrote:
> On Tue, Mar 10, 2009 at 4:14 PM, Vlad Seryakov <vl...@cr...> wrote:
>> Stephen,
>>
>> Can you describe how to import module into hg on bitbucket, it keeps saying i am not authorized.
> 
> 
> Log in as 'naviserver' at bitbucket.org.
> 
> Got to:
> 
>     http://bitbucket.org/repo/create
> 
> Fill out the info.  Once created, in the admin tab, on the right hand
> side under Permissions -> Administrators  add verseryakov and whoever
> else you'd like.
> 
> Back on your machine...
> 
> You can now clone the repo from bitbucket to your machine if you like,
> and commit straight into it -- the repo url will be set correctly for
> you.
> 
> As a shortcut, you can add something like this to your ~/.hgrc file:
> 
>     [paths]
>     nsfoo = https://vse...@bi.../naviserver/nsfoo/
> 
> this allows you to refer to the repo on bitbucket as 'nsfoo' on the
> command line. So, if you've converted one of the remaining repos and
> you want to push it, without first cloning from bitbucket, you just:
> 
>     cd ~/nsfoo-hg
>     hg outgoing nsfoo
>     hg push nsfoo
> 
> The 'outgoing' tells you what it would push without actually pushing.
> 
> If you edit the ~/nsfo-hg/.hg/hgrc to add:
> 
>     [paths]
>     default = http://bitbucket.org/naviserver/nsfoo
> 
> you don't have to tell it where you want to push to. Just:
> 
>     cd ~/nsfoo-hg
>     hg push
> 
> 
> Once it's up there, *then* you can customise the rest of the Admin
> options on bitbucket -- set up the commit emails etc. (look at
> naviserver for an example). Do this second so you don't spam the list
> with 5 years of history :-)
> 
> 
> I've converted some more modules.  Here's what's left to do:
> 
> nsconf
> nsexpat
> #nsstats
> nssys
> nstk
> nszlib
> 
> nsffmpeg
> nsfortune
> nsocaml
> nsotcl
> nssavi
> 
> nsdbext
> nsdbpd
> 
> 
> I know you said some where perhaps obsolete now, but what they hey,
> better not to loose them.
> 
> You could leave nsdbext and nsdbpd for me if you like. I was going to
> try merging in the whole aolserver history, which is a little more
> work.
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Mercurial help

[Stephen D. <sd...@gm...>]

On Tue, Mar 10, 2009 at 4:41 PM, Stephen Deasey <sd...@gm...> wrote:
> On Tue, Mar 10, 2009 at 4:14 PM, Vlad Seryakov <vl...@cr...> wrote:
>> Stephen,
>>
>> Can you describe how to import module into hg on bitbucket, it keeps saying i am not authorized.
>
>
> Log in as 'naviserver' at bitbucket.org.
>
> Got to:
>
>    http://bitbucket.org/repo/create
>
> Fill out the info.  Once created, in the admin tab, on the right hand
> side under Permissions -> Administrators  add verseryakov and whoever
> else you'd like.
>
> Back on your machine...
>
> You can now clone the repo from bitbucket to your machine if you like,
> and commit straight into it -- the repo url will be set correctly for
> you.
>
> As a shortcut, you can add something like this to your ~/.hgrc file:
>
>    [paths]
>    nsfoo = https://vse...@bi.../naviserver/nsfoo/
>
> this allows you to refer to the repo on bitbucket as 'nsfoo' on the
> command line. So, if you've converted one of the remaining repos and
> you want to push it, without first cloning from bitbucket, you just:
>
>    cd ~/nsfoo-hg
>    hg outgoing nsfoo
>    hg push nsfoo
>
> The 'outgoing' tells you what it would push without actually pushing.
>
> If you edit the ~/nsfo-hg/.hg/hgrc to add:
>
>    [paths]
>    default = http://bitbucket.org/naviserver/nsfoo
>
> you don't have to tell it where you want to push to. Just:
>
>    cd ~/nsfoo-hg
>    hg push
>
>
> Once it's up there, *then* you can customise the rest of the Admin
> options on bitbucket -- set up the commit emails etc. (look at
> naviserver for an example). Do this second so you don't spam the list
> with 5 years of history :-)
>
>
> I've converted some more modules.  Here's what's left to do:
>
> nsconf
> nsexpat
> #nsstats
> nssys
> nstk
> nszlib
>
> nsffmpeg
> nsfortune
> nsocaml
> nsotcl
> nssavi
>
> nsdbext
> nsdbpd
>
>
> I know you said some where perhaps obsolete now, but what they hey,
> better not to loose them.
>
> You could leave nsdbext and nsdbpd for me if you like. I was going to
> try merging in the whole aolserver history, which is a little more
> work.
>


Oh yeah, if anyone has a few spare minutes you could add the missing
modules to ohloh:

   https://www.ohloh.net/tags/naviserver

Then we can link them up in the Tcl wiki and embed the ohloh graphs 'n stuff.

Re: [naviserver-devel] Mercurial help

[Stephen D. <sd...@gm...>]

On Tue, Mar 10, 2009 at 4:14 PM, Vlad Seryakov <vl...@cr...> wrote:
> Stephen,
>
> Can you describe how to import module into hg on bitbucket, it keeps saying i am not authorized.


Log in as 'naviserver' at bitbucket.org.

Got to:

    http://bitbucket.org/repo/create

Fill out the info.  Once created, in the admin tab, on the right hand
side under Permissions -> Administrators  add verseryakov and whoever
else you'd like.

Back on your machine...

You can now clone the repo from bitbucket to your machine if you like,
and commit straight into it -- the repo url will be set correctly for
you.

As a shortcut, you can add something like this to your ~/.hgrc file:

    [paths]
    nsfoo = https://vse...@bi.../naviserver/nsfoo/

this allows you to refer to the repo on bitbucket as 'nsfoo' on the
command line. So, if you've converted one of the remaining repos and
you want to push it, without first cloning from bitbucket, you just:

    cd ~/nsfoo-hg
    hg outgoing nsfoo
    hg push nsfoo

The 'outgoing' tells you what it would push without actually pushing.

If you edit the ~/nsfo-hg/.hg/hgrc to add:

    [paths]
    default = http://bitbucket.org/naviserver/nsfoo

you don't have to tell it where you want to push to. Just:

    cd ~/nsfoo-hg
    hg push


Once it's up there, *then* you can customise the rest of the Admin
options on bitbucket -- set up the commit emails etc. (look at
naviserver for an example). Do this second so you don't spam the list
with 5 years of history :-)


I've converted some more modules.  Here's what's left to do:

nsconf
nsexpat
#nsstats
nssys
nstk
nszlib

nsffmpeg
nsfortune
nsocaml
nsotcl
nssavi

nsdbext
nsdbpd


I know you said some where perhaps obsolete now, but what they hey,
better not to loose them.

You could leave nsdbext and nsdbpd for me if you like. I was going to
try merging in the whole aolserver history, which is a little more
work.

[naviserver-devel] Mercurial help

[Vlad S. <vl...@cr...>]

Stephen,

Can you describe how to import module into hg on bitbucket, it keeps saying i am not authorized.

Thanks

Re: [naviserver-devel] commit/naviserver: Vlad Seryakov : Added call to nsproxy Tcl_FindExecutable(argv[0]); to keep Tcl 8.5 from crashing in thread related initialization.

[Vlad S. <vl...@cr...>]

Oops, missed it

Stephen Deasey wrote:
> On Sun, Mar 8, 2009 at 4:13 PM,  <com...@bi...> wrote:
>> 1 new changeset in naviserver:
>>
>> http://www.bitbucket.org/naviserver/naviserver/changeset/4b0584c5e64b/
>> changeset:   r2184:4b0584c5e64b
>> user:        Vlad Seryakov
>> date:        2009-03-08 17:12:53
>> summary:     Added call to nsproxy Tcl_FindExecutable(argv[0]); to keep Tcl 8.5 from crashing in thread related initialization.
> 
> 
> Already present...
> 
>   http://bitbucket.org/naviserver/naviserver/src/4b0584c5e64b/nsproxy/nsproxylib.c#cl-432
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] commit/naviserver: Vlad Seryakov : Added call to nsproxy Tcl_FindExecutable(argv[0]); to keep Tcl 8.5 from crashing in thread related initialization.

[Stephen D. <sd...@gm...>]

On Sun, Mar 8, 2009 at 4:13 PM,  <com...@bi...> wrote:
> 1 new changeset in naviserver:
>
> http://www.bitbucket.org/naviserver/naviserver/changeset/4b0584c5e64b/
> changeset:   r2184:4b0584c5e64b
> user:        Vlad Seryakov
> date:        2009-03-08 17:12:53
> summary:     Added call to nsproxy Tcl_FindExecutable(argv[0]); to keep Tcl 8.5 from crashing in thread related initialization.


Already present...

  http://bitbucket.org/naviserver/naviserver/src/4b0584c5e64b/nsproxy/nsproxylib.c#cl-432

[naviserver-devel] Wiki and repository

[Vlad S. <vl...@cr...>]

I think i converted all SF wiki pages into Tcl wiki, so we can close SF 
wiki or make it redirect to Tcl wiki.

Second, i tried to import remaining modules from CVS, not sure or i 
forgot, i cannot create repositories in the bitbucket naviserver, i use 
hg init but it does not allow via ssh, i still cannot import my public 
key, it keeps saying it is invalid and i use openssh.
-- 
Vlad Seryakov
vl...@cr...
http://www.crystalballinc.com/vlad/