naviserver-devel — Developer discussion

Re: [naviserver-devel] Does Ns_ParseObjv() look ok?

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

> Oh yes, Stephen, I love the arg parsing code This is really
> a relief. After years any years of parsing... heh..


Well, it would be great if it actually worked...  :-/

I don't think the return value of the parse-proc can be overloaded to
mean both the number of args left and 'break' (by saying zero args
left).

The way it is now, a command like [ns_foo -blah whatsit] will return 0
args left from the proc which handles the -foo switch because there
really are no args left.  But this is used as a signal to end option
parsing so objc does't get updated to reflect that fact that 'whatsit'
was consumed.  The result is that 'whatsit' gets treated as an arg
even though it was already handled.

I think the signature of the parse procs will have to be changed to
return TCL_OK, TCL_ERROR, TCL_BREAK, and pass a pointer to objc which
they will be expected to update.  What do you think?



On Mon, 28 Feb 2005 10:05:38 +0100, Zoran Vasiljevic <zv...@ar...> wrote:
> On Monday 28 February 2005 08:25, Stephen Deasey wrote:
> > I was thinking about making the Ns_ObjvBool proc look at it's arg, and
> > if present, place that in the destination pointer and not consume an
> > object from the objv array.  The presence of the option (without
> > argument) alone would signify true, it's absence false.
> >
> > Then I realised I left out the short-circuit logic...  When a --
> > (break) arg is found the options are assumed to have ended and we jump
> > to arg processing.  The signal for this is that the Ns_ObjvBreak proc
> > returns exactly objc.  But that's really what the Bool proc should
> > return in the case described above when it doesn't consume an arg.
> >
> > Unfortunately it's not as simple as that.  The "wrong num args"
> > description needs to know that the option takes no argument too.
> >
> >
> > Any thoughts?
> 
> And if Ns_ObjvBreak returns zero as a signal that none
> of the options are expected any more:
> 
>     remain = specPtr->proc(specPtr->dest, interp, objc, objv + objvIndex,
>                            specPtr->arg);
>     if (remain == 0 || remain == objc) {
>         remain = objc;
>         break;
>     }
> 
> Hm?
> Zoran

Re: [naviserver-devel] nsdsh - NaviServer shell for testing

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

There's the command ns_geturl which if called without specifying a
host dummies up a request and fetches the result from the local
machine.

If there was an nsdsh which loaded with some kind of minimal fake
server config so that the server commands were available, connection
testing stuff would Just Work with the above command.  It's kind of
limited, so it may need extended with extra headers etc...




On Tue, 8 Mar 2005 20:00:55 +0100, Zoran Vasiljevic <zv...@ar...> wrote:
> On Tuesday 08 March 2005 15:30, Vlad Seryakov wrote:
> > i have bash alias ncp="socat readline tcp4:localhost:2080" and then i
> > can type/execute anything
> 
> Ah... GPL! I knew there is a catch! My first impression was:
>   DAMN! So many development for nothing!
>     (after I saw that this thingy would suit us perfectly
>      and we have invested so much in our nsdchat thing...)
> 
> and then the cold shower after reading the license!
> OK. For non-commercial inhouse use, this is pretty nice tool.
> But as we have a commercial package we're selling we have to
> be very careful with those licensing issues...
> 
> Nevertheless, it is a good idea to use this to write tests
> and run them against the running server instance. Can this
> thing read files and spit their content to socket?
> You know what I mean? Have a Tcl script in a file and you
> kinda redirect it to nscp?
> 
> Cheers
> Zoran
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] Roadmap

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

On Tue, 08 Mar 2005 09:53:21 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> > /usr/local/ns/ ...
> >   bin
> >   include
> >   lib
> >   log
> >   man
> >   modules
> >   sample-config.tcl
> >   servers
> >
> 
> Can we move "tcl" from modules/ to the top, ussually modules are
> empty directory with tcl entry only, do you use modules/ at all?

I didn't realise tcl was in that directory, I was going to suggest
moving it in there and renaming it nstcl  :-)

We have no standard way for people to create Tcl only modules and
distribute them as they do binary modules.  You often see the
instruction "just copy this into your private tcl directory...".  I
guess we should fix that.

My modules directory is full of binary modules.  The libnsd.so and
libnsdb.so go in the libs directory.

A nice Tcl module to ship as standard would be nstelemetry, which is
currently a single adp file.  With extra hooks into [ns_info filters]
etc. it could push back the need for a lot of documentation.

Re: [naviserver-devel] ns_syslog?

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

Sorry, no Windows programming experience.

A module sounds good.  For it to make sense in the core it would have
to be integrated with the standard logging routines, otherwise it
would just be this thing stuck on the side.

The static SysLog function you added for the watchdog is pretty small,
so no great duplication of code.



On Tue, 15 Mar 2005 20:38:40 +0100, Zoran Vasiljevic <zv...@ar...> wrote:
> Hi!
> 
> Vlad suggested in the RFE #1156875 (Watchdog process restarts failed server)
> to include the interface to Unix syslog facility.
> 
> I'm actually very interested to get a kind-of abstraction
> (if possible) which would allow us to be more platform
> independent in terms of being able to send logs to the
> windows server log as well. I'm not that familiar with
> the windows platform but I have a collegue who is.
> 
> The task would be to abstract both and wrap them
> in a module, not necessarily the server itself.
> The same way the nslog is done. In this case it
> would be kindof nssyslog or similar.
> 
> I will try to see with my collegue how those two
> different logging systems could be abstracted
> and wether this is possible at all.
> 
> Do you guys have any experience on that?
> 
> Cheers
> Zoran
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] ns_syslog?

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

And actually there is already AS module that provides syslog with other 
system related functions

ftp://ftp.crystalballinc.com/pub/vlad/nssys.tar.gz

----------------------------------------------------
  ns_syslog

  Usage:
    ns_syslog ?-facility f -options o -ident i? priority message

   facility    - kernel, cron, authpriv, mail, local0, local1, daemon, 
local2,
                 news, local3, local4, local5, local6, syslog, local7, 
auth, uucp, lpr, user
   options     - list with any of { CONS NDELAY PERROR PID ODELAY NOWAIT }
   ident       - ident is prepended to every message, and is typically 
the program name
   priority    - info, alert, emerg, err, notice, warning, error, crit, 
debug

  Returns:
    nothing


Zoran Vasiljevic wrote:
> Hi!
> 
> Vlad suggested in the RFE #1156875 (Watchdog process restarts failed server)
> to include the interface to Unix syslog facility.
> 
> I'm actually very interested to get a kind-of abstraction
> (if possible) which would allow us to be more platform 
> independent in terms of being able to send logs to the
> windows server log as well. I'm not that familiar with
> the windows platform but I have a collegue who is.
> 
> The task would be to abstract both and wrap them 
> in a module, not necessarily the server itself.
> The same way the nslog is done. In this case it
> would be kindof nssyslog or similar.
> 
> I will try to see with my collegue how those two
> different logging systems could be abstracted
> and wether this is possible at all.
> 
> Do you guys have any experience on that?
> 
> Cheers
> Zoran
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] ns_syslog?

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

It looks pretty simple, but having syslog as a module will result that 
for watcher and other system things there will be
static version in the core. And usually, this is just one function, 
either in Windows or Linux.

http://www.byte.com/art/9604/sec15/art3.htm

Zoran Vasiljevic wrote:
> Hi!
> 
> Vlad suggested in the RFE #1156875 (Watchdog process restarts failed server)
> to include the interface to Unix syslog facility.
> 
> I'm actually very interested to get a kind-of abstraction
> (if possible) which would allow us to be more platform 
> independent in terms of being able to send logs to the
> windows server log as well. I'm not that familiar with
> the windows platform but I have a collegue who is.
> 
> The task would be to abstract both and wrap them 
> in a module, not necessarily the server itself.
> The same way the nslog is done. In this case it
> would be kindof nssyslog or similar.
> 
> I will try to see with my collegue how those two
> different logging systems could be abstracted
> and wether this is possible at all.
> 
> Do you guys have any experience on that?
> 
> Cheers
> Zoran
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

[naviserver-devel] ns_syslog?

[Zoran V. <zv...@ar...>]

Hi!

Vlad suggested in the RFE #1156875 (Watchdog process restarts failed server)
to include the interface to Unix syslog facility.

I'm actually very interested to get a kind-of abstraction
(if possible) which would allow us to be more platform 
independent in terms of being able to send logs to the
windows server log as well. I'm not that familiar with
the windows platform but I have a collegue who is.

The task would be to abstract both and wrap them 
in a module, not necessarily the server itself.
The same way the nslog is done. In this case it
would be kindof nssyslog or similar.

I will try to see with my collegue how those two
different logging systems could be abstracted
and wether this is possible at all.

Do you guys have any experience on that?

Cheers
Zoran

Re: [naviserver-devel] nsdsh - NaviServer shell for testing

[Zoran V. <zv...@ar...>]

On Tuesday 08 March 2005 15:30, Vlad Seryakov wrote:
> i have bash alias ncp="socat readline tcp4:localhost:2080" and then i 
> can type/execute anything


Ah... GPL! I knew there is a catch! My first impression was:
  DAMN! So many development for nothing!
    (after I saw that this thingy would suit us perfectly
     and we have invested so much in our nsdchat thing...)

and then the cold shower after reading the license!
OK. For non-commercial inhouse use, this is pretty nice tool.
But as we have a commercial package we're selling we have to
be very careful with those licensing issues...

Nevertheless, it is a good idea to use this to write tests
and run them against the running server instance. Can this
thing read files and spit their content to socket?
You know what I mean? Have a Tcl script in a file and you 
kinda redirect it to nscp?

Cheers
Zoran

Re: [naviserver-devel] Roadmap

[Bernd E. <eid...@we...>]

> Then nscgi - usefull, nsco - usefull, nsperm - ???, nsext - ???

seems there is a constant interest in nscgi, at least it was talked about a 
lot of times in the aol-mailinglist.

> Not much to have working server, some DB modules should be there, at
> least PG, MySQL, SQLite,
> thi is the main selling pitch currentlu for AS, native DB support and DB
> pooling but none of the DB modules
> included.

yes, the db modules are important. e.g. openacs community uses postgres and 
oracle.
and maybe smth. for "contrib": there is a sapi/aolserver directory in the 
PHP-distribution and a patch for that aolserver.c file on the wiki. the 
original developer seems to be gone (not of the patch, of the file itself).

I will try that with 4.3.x and 5.0.x versions of PHP and write a small README 
about it.

Re: [naviserver-devel] How to set output headers w/o sending content?

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

New driver looks interesting but nothing major different from 4.0 except 
limit pools, and the
driver code/logic looks even more complex. Connection and Sock 
separation is good but still
except dynamic limits i do not see advantages of the new driver, of 
course it is not finished yet but i heard that AOL
is going to start testing 4.1 already, so it looks like 4.1 will have 
what we can see in CVS and something more
flexible for new protocols might be there by the next year only.

I wouldn't wait for them.


Stephen Deasey wrote:
> It looks like the change was added to cvs-HEAD ~6 weeks ago, so this
> probably wont find it's way to 4.0.10 at all...
> 
> I was just thinking we should keep parity with AOLserver here.  The
> new driver code promised for AOLserver might depend upon it and it
> might start to get messy when we merge with 4.1 if we have to rename
> everything.
> 
> 
-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] Roadmap

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

> 
> I think it would be very difficult to come up with a standard set of
> these higher-level features that would please everyone.  I'd like to
> see lot of extra modules in cvs though.
> 

Yes, in CVS, packaging is another issue but having module sin CVS and be 
able to download them from one
place is what i mean.

> 
> Very few.  Less than we have now?  Does anybody use that external db stuff?

Currently we have just nssock which is not module anymore per se, just 
an attachement to core http driver
for loading more than once instance which is not always clear. May be we 
can teach http driver to bind to as many
addresses without loading different modules.
Then nscgi - usefull, nsco - usefull, nsperm - ???, nsext - ???

Not much to have working server, some DB modules should be there, at 
least PG, MySQL, SQLite,
thi is the main selling pitch currentlu for AS, native DB support and DB 
pooling but none of the DB modules
included.

> I'd like to change the installation layout.  Currently it's something like this:
> 
> /usr/local/ns/ ...
>   bin
>   include
>   lib
>   log
>   man
>   modules
>   sample-config.tcl
>   servers
> 

Can we move "tcl" from modules/ to the top, ussually modules are
empty directory with tcl entry only, do you use modules/ at all?

Also in my installations i use etc/ for all config files, at least i 
have nsd.tcl and mime.tcl(with huge list of all mime types)
there plus some other config specific files. Do you think this is usefull?


Here is my structure ( i am not sugesting it, just show what i use):

/usr/local/ns
    bin/
    lib/
    html/
    include/
    etc/
    logs/
    tcl/

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] nsdsh - NaviServer shell for testing

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

I am using ncp module with socat which supports readline:

i have bash alias ncp="socat readline tcp4:localhost:2080" and then i 
can type/execute anything

there. Never thought about anything else since. The only thing that 
needs to be tested with browser
is connection stuff, like cookies/headers, but that wget is for, prepare 
file with everything and then just
call it, a lot of manual work but actually i donot use it very often, 
browser is s till real testing environment.

Zoran Vasiljevic wrote:
> On Tuesday 08 March 2005 09:09, Stephen Deasey wrote:
> 
> 
>>I'm not sure what the right way to go here is.  Would such an nsdsh be
>>useful, or should we just add more to the server binary, or should I
>>just be a bit more creative when writing tests..?
>>
> 
> 
> I do not know if this is what you'd need...
> 
> What we've done is to spawn a thread on server start which opens a
> named pipe on the filesystem. The location and naming convention of
> the named pipe is well known. The thread reads incoming data more/less
> the same as nscp and executes the command. We also wrote a very tiny
> utility in Tcl (and wrapped it into a single exectuable with mktclapp)
> which also connects to the same named pipe, reads stdin for commands
> and passes the input to the pipe.
> 
> This looks like:
> 
> zvlinux:/usr/local/homes/zv # cd /usr/local/aw
> zvlinux:/usr/local/aw # bin/nsdchat
> % srvinfo version
> 4.0
> 
> The "nsdchat" is this small utility. This way we can execute any
> server command on a running server instance. The thread which reads
> the pipe executes command in a secure Tcl interp and we have opened
> a special API for that. This need not be, of course.
> 
> So, this way we are actually using the server itself to run arbitrary 
> scripts. The nsdchat utility can also read files on the filesystem 
> and pass them to server for execution:
> 
> zvlinux:/usr/local/aw # echo "srvinfo version" > my.tcl
> zvlinux:/usr/local/aw # bin/nsdchat < my.tcl
> 4.0
> 
> Is this the functionality you're looking for?
> 
> Cheers
> Zoran
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] nsdsh - NaviServer shell for testing

[Zoran V. <zv...@ar...>]

On Tuesday 08 March 2005 09:09, Stephen Deasey wrote:

> I'm not sure what the right way to go here is.  Would such an nsdsh be
> useful, or should we just add more to the server binary, or should I
> just be a bit more creative when writing tests..?
> 

I do not know if this is what you'd need...

What we've done is to spawn a thread on server start which opens a
named pipe on the filesystem. The location and naming convention of
the named pipe is well known. The thread reads incoming data more/less
the same as nscp and executes the command. We also wrote a very tiny
utility in Tcl (and wrapped it into a single exectuable with mktclapp)
which also connects to the same named pipe, reads stdin for commands
and passes the input to the pipe.

This looks like:

zvlinux:/usr/local/homes/zv # cd /usr/local/aw
zvlinux:/usr/local/aw # bin/nsdchat
% srvinfo version
4.0

The "nsdchat" is this small utility. This way we can execute any
server command on a running server instance. The thread which reads
the pipe executes command in a secure Tcl interp and we have opened
a special API for that. This need not be, of course.

So, this way we are actually using the server itself to run arbitrary 
scripts. The nsdchat utility can also read files on the filesystem 
and pass them to server for execution:

zvlinux:/usr/local/aw # echo "srvinfo version" > my.tcl
zvlinux:/usr/local/aw # bin/nsdchat < my.tcl
4.0

Is this the functionality you're looking for?

Cheers
Zoran

Re: [naviserver-devel] How to set output headers w/o sending content?

[Zoran V. <zv...@ar...>]

On Tuesday 08 March 2005 06:02, Stephen Deasey wrote:
> 2005-03-08 06:02
>=20
> I don't think you need this. =A0In 4.x all content is read fromt the
> network and buffered before your code ever runs. =A0You're just copying
> it from one buffer to another.


OK. I did not know that. This command is still there from
the 3.0 in one of my modules. I can just scrap it then.

Zoran

[naviserver-devel] nsdsh - NaviServer shell for testing

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

Neither RedHat or Fedora ship with a threaded build of Tcl, so I build
my own package and install in parallel with the original.  I've been
naming the tclsh84 executable nsdsh.

Our tests, such as they are, also need to run in a threaded Tcl shell.
 Which one? How to find it?  Maybe we should build an nsdsh which
would load libnsd.so by default and link to the correct Tcl libs.

It's also currently quite difficult to test certain aspects of the
server.  e.g. for cookie handling you really want a connection
available, but making actual network connections in a test environment
seems a little fragile.  So I thought maybe the nsdsh could provide
extra testing hooks.

It could take a '-t' option on the command line and load a standard
server config file, using it to more fully emulate a running server
than the current libnsd does.  With this option present, stdin could
be the request stream, stdout the response, and stderr the log file? 
It might register some extra _ns_debug_* commands?

I'm not sure what the right way to go here is.  Would such an nsdsh be
useful, or should we just add more to the server binary, or should I
just be a bit more creative when writing tests..?

Re: [naviserver-devel] Roadmap

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

On Sun, 06 Mar 2005 18:17:06 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> Hi guys,
> 
> Do you have any roadmap-kind of plans or thoughts about naviserver?
> Slowly it is getting to the point when it can replace all my patched
> aolservers, i already use it at home instead of AS. As i remember, the
> initial goal was to patch AS with all out modifications so we start
> using it instead of AS and then decide what to do next, right?
> 
> So,
> 
> 1. Do you have any thoughts about Web tools/frameworks we can bundle
> with it?


I think it would be very difficult to come up with a standard set of
these higher-level features that would please everyone.  I'd like to
see lot of extra modules in cvs though.


> 2. Which modules should be in the distrib?


Very few.  Less than we have now?  Does anybody use that external db stuff?


> 3. Docs are still needed to be addressed, when i do not have strength to
> develop late evenings i could write/add docs but which format we gonna use?
> 4. Do we need to refine installation procedure?
> After installation, going to port 8080 (this way no need root to install
> and test it) the user should see some kind of web page with at least
> stats and docs (similar to apache, that is the big hole in AS right now)


I'd like to change the installation layout.  Currently it's something like this:

/usr/local/ns/ ...
  bin
  include
  lib
  log
  man
  modules
  sample-config.tcl
  servers

I'd like to drop the enforced 'ns' so that --prefix-... works
properly.  I'd also like to teach the Makefile to obey the standard
configure options like --libdir=... etc.  This enables RPM and Deb
packages to install things under /usr/lib64 if the machine is an
Opteron, for example.

Also, placing the 'servers' directory under /usr/local doesn't make
much sense.  This is root owned so by default you can't edit your own
web content.  I think the standard place for Linux is now /srv, not
sure about the built-in configure default.

It also makes sense to me to ship some kind of server deployment
script.  So rather than install the default servers/server1 and try to
grab port 8080, the user would type 'nsdeploy myserver /srv/myserver'
or whatever, possible with extra options, and it would create the
hierarchy for them.  I always end up having multiple installations on
one machine.  The script could ask questions if you don't supply any
args, touch up the config file etc.

Re: [naviserver-devel] Roadmap

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

Sounds like a packaging problem to me.  Bundling modules into the core
will only turn it into a development/release problem.

Large projects actually seem to be going the other way:  Postgres are
moving things out of core and onto gborg;  X.org is busy splitting up
into server, fonts, utils, etc.

What about the many modules and features that don't make it into core,
do we just acknowledge that they will be a complete pain to build and
install?


I'm so lazy that I package all my software into RPMs.  I only have to
do it once, and from then on it's a doddle to install consistently on
multiple machines without referring to docs or making mistakes.

If y'all are dead set against whatever packaging system is native to
your OS, you still have options.  The Gnome and KDE projects for
example have created build scripts which automatically download the
source from cvs, build, and install, managing dependencies and build
order.


Can we put all non-core modules in a 'modules' directory in cvs?  I
thought it would be cool if each one had some meta data associated
with it, maybe use something like the DOAP format:
http://usefulinc.com/doap   There would be a little script which
automatically builds web pages for all modules each night.   Packages
would be auto-built and placed int yum/apt repositories.  There would
be links to CVS, the bug tracker for that module, the latest entries
in the ChangeLog...


That's my grand vision for modules and packaging -- a small shell
script will replace me!  How does this sound?  What OS are you
installing on?




On Mon, 7 Mar 2005 16:34:55 +0100, Bernd Eidenschink <eid...@we...> wrote:
> Hi Vlad,
> 
> > > I have no idea. We are using the server with our own templating
> > > framework based on tDOM but I do not think this is something
> > > of general interest.
> >
> > I have light-weight templating i can extract and put into modules/tcl,
> > that will be one small file and will support OACS-style templating
> > where .adp and .tcl are define template. But it requires some work to be
> > extracted
> > from my "big" framework system.
> 
> I would also like to extract some functionality from our toolkit. It would be
> nice if there is a small toolkit with basic stuff like database abstraction,
> page and session management, authentication and basic permissions.
> I like the idea to rewrite it in a way so that it utilizes new naviserver
> functions like the c-cookie-api or database stuff (I read from Stephen?).
> In our framework we do not map and use the standard ADP-stuff, we use
> registered filters and procs and ns_adp_parse files, but it should be
> managable to rewrite it.
> 
> > For public version  in the future i would include as manu maintained
> > modules as possible.
> 
> What do you think about a "contrib" directory (like in PostgreSQL or other
> packages):
> 
> <QUOTE>
> 
> The PostgreSQL contrib tree
> ---------------------------
> This subtree contains porting tools, analysis utilities, and plug-in
> features that are not part of the core PostgreSQL system, mainly because
> they address a limited audience or are too experimental to be part of
> the main source tree.  This does not preclude their usefulness.
> 
> Each subdirectory contains a README file with information about the
> module.
> 
> </QUOTE>
> 
> Something similar. This way it would be clearly stated that something is
> useful and/or experimental and/or just new (Message: "Give it a try").
> 
> Maybe this would be a place for a chroot-installation-script, like my first
> try here:
> http://www.kinetiqa.de/aolserver/
> (But rewritten, it's a little bit old and worries about too many packages)
> 
> Bernd.

Re: [naviserver-devel] How to set output headers w/o sending content?

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

It looks like the change was added to cvs-HEAD ~6 weeks ago, so this
probably wont find it's way to 4.0.10 at all...

I was just thinking we should keep parity with AOLserver here.  The
new driver code promised for AOLserver might depend upon it and it
might start to get messy when we merge with 4.1 if we have to rename
everything.



On Tue, 08 Mar 2005 00:06:19 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> Done except naming, it is different in AS 4.1 CVS and 4.0.x, does it
> really matter much? But i am not insisting.
> 
> Stephen Deasey wrote:
> > How about removing the comment, marking the function void, and
> > changing the naming to match the addtions to AOLserver?
> >
> >
> > On Mon, 07 Mar 2005 23:53:35 -0500, Vlad Seryakov
> > <vl...@cr...> wrote:
> >
> >>Oops, i was thinking to retern previous status but then change my mind,
> >>comment left, i will correct it.
> >>
> >>Stephen Deasey wrote:
> >>
> >>>The comment for Ns_ConnSetResponseStatus() says that it reurns the
> >>>previous status.  It actually returns the current, just-set status.
> >>>So does ns_conn status.
> >>>
> >>>I don't think it's neccessary.  The other simillar functions in that
> >>>file have no return value. (Ns_ConnSetUrlEncoding(),
> >>>Ns_ConnSetWriteEncodedFlag()...)
> >>>
> >>>
> >>>Added six weeks ago to AOLserver cvs: Ns_ConnSetStatus(), Ns_ConnGetStatus()
> >>>
> >>>http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nsd/conn.c?r1=1.1&r2=1.2
> >>>
> >>>
> >>>
> >>>
> >>>On Sat, 05 Mar 2005 13:22:31 -0500, Vlad Seryakov
> >>><vl...@cr...> wrote:
> >>>
> >>>
> >>>>I know, it is just the core of the update to be approved
> >>>>
> >>>>Zoran Vasiljevic wrote:
> >>>>
> >>>>
> >>>>>On Saturday 05 March 2005 18:36, Vlad Seryakov wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>>>Patch is attached, if everybody is Okay with it i will commit it
> >>>>>
> >>>>>
> >>>>>Do not forget to update include/ns.h
> >>>>>and put the prototype there!
> >>>
> >>>
> >>>
> >>>-------------------------------------------------------
> >>>SF email is sponsored by - The IT Product Guide
> >>>Read honest & candid reviews on hundreds of IT Products from real users.
> >>>Discover which products truly live up to the hype. Start reading now.
> >>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> >>>_______________________________________________
> >>>naviserver-devel mailing list
> >>>nav...@li...
> >>>https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> >>
> >>--
> >>Vlad Seryakov
> >>571 262-8608 office
> >>vl...@cr...
> >>http://www.crystalballinc.com/vlad/
> >>
> >>-------------------------------------------------------
> >>SF email is sponsored by - The IT Product Guide
> >>Read honest & candid reviews on hundreds of IT Products from real users.
> >>Discover which products truly live up to the hype. Start reading now.
> >>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> >>_______________________________________________
> >>naviserver-devel mailing list
> >>nav...@li...
> >>https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> >>
> >
> >
> >
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now.
> > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> > _______________________________________________
> > naviserver-devel mailing list
> > nav...@li...
> > https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 
> --
> Vlad Seryakov
> 571 262-8608 office
> vl...@cr...
> http://www.crystalballinc.com/vlad/
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] How to set output headers w/o sending content?

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

Done except naming, it is different in AS 4.1 CVS and 4.0.x, does it 
really matter much? But i am not insisting.

Stephen Deasey wrote:
> How about removing the comment, marking the function void, and
> changing the naming to match the addtions to AOLserver?
> 
> 
> On Mon, 07 Mar 2005 23:53:35 -0500, Vlad Seryakov
> <vl...@cr...> wrote:
> 
>>Oops, i was thinking to retern previous status but then change my mind,
>>comment left, i will correct it.
>>
>>Stephen Deasey wrote:
>>
>>>The comment for Ns_ConnSetResponseStatus() says that it reurns the
>>>previous status.  It actually returns the current, just-set status.
>>>So does ns_conn status.
>>>
>>>I don't think it's neccessary.  The other simillar functions in that
>>>file have no return value. (Ns_ConnSetUrlEncoding(),
>>>Ns_ConnSetWriteEncodedFlag()...)
>>>
>>>
>>>Added six weeks ago to AOLserver cvs: Ns_ConnSetStatus(), Ns_ConnGetStatus()
>>>
>>>http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nsd/conn.c?r1=1.1&r2=1.2
>>>
>>>
>>>
>>>
>>>On Sat, 05 Mar 2005 13:22:31 -0500, Vlad Seryakov
>>><vl...@cr...> wrote:
>>>
>>>
>>>>I know, it is just the core of the update to be approved
>>>>
>>>>Zoran Vasiljevic wrote:
>>>>
>>>>
>>>>>On Saturday 05 March 2005 18:36, Vlad Seryakov wrote:
>>>>>
>>>>>
>>>>>
>>>>>>Patch is attached, if everybody is Okay with it i will commit it
>>>>>
>>>>>
>>>>>Do not forget to update include/ns.h
>>>>>and put the prototype there!
>>>
>>>
>>>
>>>-------------------------------------------------------
>>>SF email is sponsored by - The IT Product Guide
>>>Read honest & candid reviews on hundreds of IT Products from real users.
>>>Discover which products truly live up to the hype. Start reading now.
>>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>>_______________________________________________
>>>naviserver-devel mailing list
>>>nav...@li...
>>>https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>>
>>--
>>Vlad Seryakov
>>571 262-8608 office
>>vl...@cr...
>>http://www.crystalballinc.com/vlad/
>>
>>-------------------------------------------------------
>>SF email is sponsored by - The IT Product Guide
>>Read honest & candid reviews on hundreds of IT Products from real users.
>>Discover which products truly live up to the hype. Start reading now.
>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>_______________________________________________
>>naviserver-devel mailing list
>>nav...@li...
>>https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>>
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] How to set output headers w/o sending content?

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

I don't think you need this.  In 4.x all content is read fromt the
network and buffered before your code ever runs.  You're just copying
it from one buffer to another.


On Sat, 5 Mar 2005 10:38:09 +0100, Zoran Vasiljevic <zv...@ar...> wrote:
> > What is [ns_conn discardcontent] used for?
> 
> This I have made some years ago and it may not be
> needed anymore today (not sure). What I ment with
> that is: by processing the incoming request, I
> look into http headers. If I decide to drop the
> request because there is something in the headers
> I don't like, I will handle all of the incoming
> data from the client and just discard everything.
> This is important for keepalive connections because
> you have to be careful and read exactly the
> contentLength bytes from the peer. I somehow could
> not do that using existing ns_ commands so I invented
> my own. Question: is there another way of doing this?
> Look at the implementation I have now:
> 
> static int
> NsxDiscardCmd (cl, interp, objc, objv)
>     ClientData      cl;
>     Tcl_Interp    * interp;
>     int             objc;
>     Tcl_Obj       * CONST objv[];
> {
>     size_t buflen;
>     int nb;
>     char *buf;
>     Ns_Conn *conn;
> 
>     /*
>      * Syntax: nsx discardcontent
>      */
> 
>     conn = Ns_TclGetConn(interp);
>     if (conn == NULL) {
>         Tcl_AppendResult(interp, "no connection", NULL);
>         return TCL_ERROR;
>     }
> 
>     buflen = 32*1024;
>     buf = (char*)Tcl_Alloc(buflen);
> 
>     /*
>      * Sink conn->contentLength bytes from connection
>      */
> 
>     while(conn->contentLength > 0) {
>         nb = (conn->contentLength < buflen) ? conn->contentLength : buflen;
>         nb = Ns_ConnRead(conn, buf, nb);
>         if (nb == NS_ERROR) {
>             break;
>         }
>         conn->contentLength -= nb;
>     }
> 
>     Tcl_Free((char*)buf);
> 
>     return TCL_OK;
> }

Re: [naviserver-devel] How to set output headers w/o sending content?

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

How about removing the comment, marking the function void, and
changing the naming to match the addtions to AOLserver?


On Mon, 07 Mar 2005 23:53:35 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> Oops, i was thinking to retern previous status but then change my mind,
> comment left, i will correct it.
> 
> Stephen Deasey wrote:
> > The comment for Ns_ConnSetResponseStatus() says that it reurns the
> > previous status.  It actually returns the current, just-set status.
> > So does ns_conn status.
> >
> > I don't think it's neccessary.  The other simillar functions in that
> > file have no return value. (Ns_ConnSetUrlEncoding(),
> > Ns_ConnSetWriteEncodedFlag()...)
> >
> >
> > Added six weeks ago to AOLserver cvs: Ns_ConnSetStatus(), Ns_ConnGetStatus()
> >
> > http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nsd/conn.c?r1=1.1&r2=1.2
> >
> >
> >
> >
> > On Sat, 05 Mar 2005 13:22:31 -0500, Vlad Seryakov
> > <vl...@cr...> wrote:
> >
> >>I know, it is just the core of the update to be approved
> >>
> >>Zoran Vasiljevic wrote:
> >>
> >>>On Saturday 05 March 2005 18:36, Vlad Seryakov wrote:
> >>>
> >>>
> >>>>Patch is attached, if everybody is Okay with it i will commit it
> >>>
> >>>
> >>>Do not forget to update include/ns.h
> >>>and put the prototype there!
> >
> >
> >
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now.
> > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> > _______________________________________________
> > naviserver-devel mailing list
> > nav...@li...
> > https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 
> --
> Vlad Seryakov
> 571 262-8608 office
> vl...@cr...
> http://www.crystalballinc.com/vlad/
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] How to set output headers w/o sending content?

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

Oops, i was thinking to retern previous status but then change my mind, 
comment left, i will correct it.

Stephen Deasey wrote:
> The comment for Ns_ConnSetResponseStatus() says that it reurns the
> previous status.  It actually returns the current, just-set status. 
> So does ns_conn status.
> 
> I don't think it's neccessary.  The other simillar functions in that
> file have no return value. (Ns_ConnSetUrlEncoding(),
> Ns_ConnSetWriteEncodedFlag()...)
> 
> 
> Added six weeks ago to AOLserver cvs: Ns_ConnSetStatus(), Ns_ConnGetStatus()
> 
> http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nsd/conn.c?r1=1.1&r2=1.2
> 
> 
> 
> 
> On Sat, 05 Mar 2005 13:22:31 -0500, Vlad Seryakov
> <vl...@cr...> wrote:
> 
>>I know, it is just the core of the update to be approved
>>
>>Zoran Vasiljevic wrote:
>>
>>>On Saturday 05 March 2005 18:36, Vlad Seryakov wrote:
>>>
>>>
>>>>Patch is attached, if everybody is Okay with it i will commit it
>>>
>>>
>>>Do not forget to update include/ns.h
>>>and put the prototype there!
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] How to set output headers w/o sending content?

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

The comment for Ns_ConnSetResponseStatus() says that it reurns the
previous status.  It actually returns the current, just-set status. 
So does ns_conn status.

I don't think it's neccessary.  The other simillar functions in that
file have no return value. (Ns_ConnSetUrlEncoding(),
Ns_ConnSetWriteEncodedFlag()...)


Added six weeks ago to AOLserver cvs: Ns_ConnSetStatus(), Ns_ConnGetStatus()

http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nsd/conn.c?r1=1.1&r2=1.2




On Sat, 05 Mar 2005 13:22:31 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> I know, it is just the core of the update to be approved
> 
> Zoran Vasiljevic wrote:
> > On Saturday 05 March 2005 18:36, Vlad Seryakov wrote:
> >
> >>Patch is attached, if everybody is Okay with it i will commit it
> >
> >
> > Do not forget to update include/ns.h
> > and put the prototype there!

Re: [naviserver-devel] Re: [naviserver-commits] naviserver/nslog nslog.c,1.1.1.1,1.2

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

http://cvs.sourceforge.net/viewcvs.py/naviserver/naviserver/nslog/nslog.c?rev=1.3&view=log

The cvs commit message for nslog/nslog.c now reads:

"Add Ns_ConnSetResponseStatus command to be able to se response status
without touching headers.  Also ns_conn status has been extended with
ability to set response status if third argument is provided. ns_conn
status ?newStatus?"

Which will thoroughly confuse someone in the future.  Please be
careful about combining changes in a single commit.



On Thu, 03 Mar 2005 10:08:33 -0500, Vlad Seryakov
<vl...@cr...> wrote:
> I've been using that new module for so long time that i did not realized
> that the code is so different than
> current style, i am sorry.
> 
> > The log message doesn't say what has changed, just that something has
> > changed, although the ChangeLog entry is better.  The X-Forwarded-For
> > change is a bug fix and is not mentioned in either.  I thought we'd
> > already agreed that non-trivial changes should have an RFE or BUG
> > posted to the tracker?
> 
> No, i did not add anythig new,  X-Forwarded-For was handled by old nslog
> as well,
> i just restructure the module to support more options
> 
> > It's unreasonable to expect end users to examine the source code for
> > documentation, so we should not clutter up the code with it.  There
> > are no other source files I'm aware of which do this.
> 
> nslog used to have index.html, in the CVS i do not see it, so i put the
> docs in the nslog.c for now and
> we do not have docs for modules yet.
> 
> 
> > The members of the Log struct have been re-arranged.  Variable
> > declarations beginning a function have moved around.  Many of the
> > comments have been reformatted.  Correctly formatted code has had
> > braces and spacing removed, etc.  This all makes it very tricky to
> > read diffs to figure out what has logically changed.
> >
> Yes, that's my fault, i will change it to follow naviserver code style.

Re: [naviserver-devel] Roadmap

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

> 
>>- using -b 0.0.0.0 will listen on all interfaces
> 
> 
> I'm afraid this is still not clear to me, but I'll have
> to give you some examples why this is making me trouble.
> I will prepare the switch to naviserver in our development
> branch in couple of days and then this will be clear


More correct solution:

nsd ... -b 0.0.0.0

and

ns_section      "ns/server/${server}/module/nssock"
ns_param        address         0.0.0.0
...

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/