naviserver-devel — Developer discussion

Re: [naviserver-devel] Explanation of last changes

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> >
> > It isn't a dummy server, it's needed.  The current example-config.tcl
> > is wrong in a number of respects, for example this comment:
> >
> > #
> > # Server-level configuration
> > #
> > #  There is only one server in naviserver, but this is helpful when multiple
> > #  servers share the same configuration file.  This file assumes that only
> > #  one server is in use so it is set at the top in the "server" Tcl variable.
> > #  Other host-specific values are set up above as Tcl variables, too.
> > #
> >
> > There isn't only one server!  All Tcl commands and procs *have* to run
> > in the context of a virtual server, even if there's only one.
>
> I can put it in other words: there is always virtual server in
> NaviServer, if not defined in config file, internal server "default" is
> used instead. It is impossible to run nsd without virtual server, so
> instead of exiting with error internal server is used.
>
> I am not insisting on doing it the way i did but it seemed to me very
> logical and not breaking overall NS architecture or model.
>
> But of course, suggestions are welcome, this is development version
> under discussion.


The stuff about virtual servers in the config file just isn't hard.
And the problem with hiding them is that it really gives people the
wrong mental-model about how things work.


> > Yes, you can have global modules.  But not if they register Tcl
> > commands!  Which virtual server context do they run with?
> >
> > I guess you could write a module such that when it's loaded globally,
> > it roots around for all virtual servers and adds it's Tcl commands to
> > each. But they don't do that...
> >
> > The idea of creating a default virtual server magically hides this
> > important distinction, which makes it harder over all for people to
> > understand what's going on.
>
> This part actuall out of sync with overall NS model and logic, the only
> module that does it is nssock and it was made so for enabling host based
> virtual server support. I think this is ill and needs to be removed or
> re-build, if we always have virtual server, its nssock can do host
> mapping, no need to load it globally and global modules should be
> eliminated.


Hmm, not sure.  The nsdb module is also a global module.  But it's
kinda weird...

Re: [naviserver-devel] Explanation of last changes

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

> 
> It isn't a dummy server, it's needed.  The current example-config.tcl
> is wrong in a number of respects, for example this comment:
> 
> #
> # Server-level configuration
> #
> #  There is only one server in naviserver, but this is helpful when multiple
> #  servers share the same configuration file.  This file assumes that only
> #  one server is in use so it is set at the top in the "server" Tcl variable.
> #  Other host-specific values are set up above as Tcl variables, too.
> #
> 
> There isn't only one server!  All Tcl commands and procs *have* to run
> in the context of a virtual server, even if there's only one.

I can put it in other words: there is always virtual server in 
NaviServer, if not defined in config file, internal server "default" is 
used instead. It is impossible to run nsd without virtual server, so 
instead of exiting with error internal server is used.

I am not insisting on doing it the way i did but it seemed to me very 
logical and not breaking overall NS architecture or model.

But of course, suggestions are welcome, this is development version 
under discussion.

> Yes, you can have global modules.  But not if they register Tcl
> commands!  Which virtual server context do they run with?
> 
> I guess you could write a module such that when it's loaded globally,
> it roots around for all virtual servers and adds it's Tcl commands to
> each. But they don't do that...
> 
> The idea of creating a default virtual server magically hides this
> important distinction, which makes it harder over all for people to
> understand what's going on.

This part actuall out of sync with overall NS model and logic, the only 
module that does it is nssock and it was made so for enabling host based 
virtual server support. I think this is ill and needs to be removed or 
re-build, if we always have virtual server, its nssock can do host 
mapping, no need to load it globally and global modules should be 
eliminated.


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

Re: [naviserver-devel] contrib

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

How about we do this after the next release, but before we start
making changes for the next one?  That way we minimise the noise while
we're working in stuff.


On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> Ok, good point
>
>
> PS: On another note, if nobody objects, can i run all source files and
> remove trailing spaces from every line, it is like garbage anyway and i
> will not disturb CVS anymore with too many changes:-))) ???
>
> Stephen Deasey wrote:
> > On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> >> That was the reason i chnaged nsstats.tcl into page, not module, now you
> >> can put it anywhere under pages/ and point browser and use it instead of
> >> restarting server just or change something or enable/disable it.
> >>
> >> That was the main reason not to use it on existing production servers
> >> because it is a module
> >
> >
> > OK.  Great reason.  It's still a module -- a module with one page.  It
> > should live in the same place as all out other modules, have a README,
> > an example-config.tcl, etc...
> >
> > -------------------------------------------------------------------------
> > Using Tomcat but need to do more? Need to support web services, security?
> > Get stuff done quickly with pre-integrated technology to make your job easier
> > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> > _______________________________________________
> > 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/
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] Explanation of last changes

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> Auto magic i do not like myself but it seems too restrictive to have
> some dummy server definition to start server if there will be only one
> server anyway. Virtual servers are good feature for advanced use, we
> used them before when had multiple servers.


It isn't a dummy server, it's needed.  The current example-config.tcl
is wrong in a number of respects, for example this comment:

#
# Server-level configuration
#
#  There is only one server in naviserver, but this is helpful when multiple
#  servers share the same configuration file.  This file assumes that only
#  one server is in use so it is set at the top in the "server" Tcl variable.
#  Other host-specific values are set up above as Tcl variables, too.
#

There isn't only one server!  All Tcl commands and procs *have* to run
in the context of a virtual server, even if there's only one.

Yes, you can have global modules.  But not if they register Tcl
commands!  Which virtual server context do they run with?

I guess you could write a module such that when it's loaded globally,
it roots around for all virtual servers and adds it's Tcl commands to
each. But they don't do that...

The idea of creating a default virtual server magically hides this
important distinction, which makes it harder over all for people to
understand what's going on.


> But to start one simple web server, i think it does not need to be that
> complicated, enabling module you want and have it running is good way to
> immediately see it working. I agree, making it open or less secure just
> to be able simply run is bad, but if security is not involved and no
> changes in the core for running default/complex configs, maybe it is not
> that bad once it documented and logged?


Unfortunately you can't tell when security is involved or important,
so defaults matter and must be conservative.  Luckily, it's trivial to
enable to enable modules.

I think that a simple-config.tcl will improve the situation 100%.
Without all the esoteric config values, which should default to
something sane anyway, the config file should fit on a couple of
screens or so, singe virtual server included.


> Stephen Deasey wrote:
> > On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> >> - if no servers defined, it will run server "default", i.e. it silently
> >> creates server "default" in config section "ns/servers"
> >
> >
> > Er, no.
> >
> > No crazy magic to auto-create config behind your back.  Virtual
> > servers just aren't complicated.  They're a feature, and they're
> > embedded deeply into the core code (all the references to char *server
> > in the C API) so it's not going any where quickly.
> >
> > I agree with your over all goal, which is to have a minimal config
> > file able to boot and run a functional server, right?  This is
> > definitely a good idea.
> >
> > The current sample-config.tcl doesn't handle this case well, and it
> > shouldn't -- it's supposed to be a complete reference to all config
> > values.
> >
> >
> > So, we need some other way to handle basic servers.  Magic config
> > construction just isn't the way to do it though.  How about we move
> > nsd-min.tcl you just added to contrib (dumping ground, grrr...) into
> > the top level, and call it simple-config.tcl?
> >
> > We'll have two config files ship by default:
> >
> >   sample-config.tcl -- Exhaustive list of all config variables
> >   simple-config.tcl -- Basic config needed to run a single virtual server.
> >
> > They will be identical in the sense that the resulting running server
> > will have the same properties, because the sample-config.tcl will be
> > illustrating *optional* config values, and possibly they''' all be
> > commented out.
> >
> > We can rename 'server1' to 'default' if you like.
> >
> > -------------------------------------------------------------------------
> > Using Tomcat but need to do more? Need to support web services, security?
> > Get stuff done quickly with pre-integrated technology to make your job easier
> > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> > _______________________________________________
> > 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/
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
>

Re: [naviserver-devel] Documentation & website question

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

This sounds better solution than using SF wiki but still site will not 
be distinctive and look like "very spare time and not often used project".

I am still thinking a dedicated server running naviserver is the way to 
go but as it happen before i am not convincing anybody, just my opinion

Stephen Deasey wrote:
> On 9/7/06, Bernd Eidenschink <eid...@we...> wrote:
>>> If the API doc is always automatically uptodate on the website then
>>> that can be the single canonical location to find everything.
>> A little grouching from me: Not now, but sometime later, we need a faster
>> webpage. No matter what:  a dedicated server, a new or version of the Wiki or
>> another Wiki that allows caching in the sense of static HTML pages, or no
>> Wiki at all and from whatever produced HTML pages.
>> But currently THIS wiki is so slow at times, so r e a l l y slow, it kills
>> time you could better invest in doing real work.
>>
>> Bernd.
> 
> It is annoyingly slow, isn't it.  The MediaWiki does already cache to disk...
> 
> Here's a weird idea:  Why don't we use the Tcl wiki?
> 
>   http://wiki.tcl.tk/
> 
> NaviServer *is* a Tcl application, and our examples are mainly going
> to be Tcl applications.  We have a whole bunch of Tcl (NaviServer)
> modules.  We probably should be a bit more integrated into the Tcl
> community, and I'm sure there a are Tcl folks who would be interested
> in what we're doing.  It doesn't feel like we'd be spamming some one
> else wiki.
> 
> Doing this would maintain a very important advantage we currently have
> with sourceforge: very little maintenance on our part. In fact it
> would be better: zero maintenance!  Plus, presumably there would be
> more people to help clean up after spammers etc.
> 
> A disadvantage might be that the Tcl wiki is a little plain. We also
> don't have any control, should we need to change something.  Well, I
> guess we have as much control as the rest of the Tcl community -- we
> can ask.
> 
> 
> If we did this, I imagine we would turn the current front page into a
> static portal page.  It would point to key pages within the Tcl wiki
> which we would maintain, along with pointers to out mailing list
> interface, cvs, etc.
> 
> The rest of our current website would be auto generated.  The API
> docs, we've mentioned.  We should also auto generate a page for each
> module in cvs, using the DOAP file.
> 
> 
> Is this a crazy idea?  Should we ask someone at the Tcl wiki?
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] contrib

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

Ok, good point


PS: On another note, if nobody objects, can i run all source files and 
remove trailing spaces from every line, it is like garbage anyway and i 
will not disturb CVS anymore with too many changes:-))) ???

Stephen Deasey wrote:
> On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
>> That was the reason i chnaged nsstats.tcl into page, not module, now you
>> can put it anywhere under pages/ and point browser and use it instead of
>> restarting server just or change something or enable/disable it.
>>
>> That was the main reason not to use it on existing production servers
>> because it is a module
> 
> 
> OK.  Great reason.  It's still a module -- a module with one page.  It
> should live in the same place as all out other modules, have a README,
> an example-config.tcl, etc...
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] contrib

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> That was the reason i chnaged nsstats.tcl into page, not module, now you
> can put it anywhere under pages/ and point browser and use it instead of
> restarting server just or change something or enable/disable it.
>
> That was the main reason not to use it on existing production servers
> because it is a module


OK.  Great reason.  It's still a module -- a module with one page.  It
should live in the same place as all out other modules, have a README,
an example-config.tcl, etc...

Re: [naviserver-devel] Documentation & website question

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

On 9/7/06, Bernd Eidenschink <eid...@we...> wrote:
> > If the API doc is always automatically uptodate on the website then
> > that can be the single canonical location to find everything.
>
> A little grouching from me: Not now, but sometime later, we need a faster
> webpage. No matter what:  a dedicated server, a new or version of the Wiki or
> another Wiki that allows caching in the sense of static HTML pages, or no
> Wiki at all and from whatever produced HTML pages.
> But currently THIS wiki is so slow at times, so r e a l l y slow, it kills
> time you could better invest in doing real work.
>
> Bernd.

It is annoyingly slow, isn't it.  The MediaWiki does already cache to disk...

Here's a weird idea:  Why don't we use the Tcl wiki?

  http://wiki.tcl.tk/

NaviServer *is* a Tcl application, and our examples are mainly going
to be Tcl applications.  We have a whole bunch of Tcl (NaviServer)
modules.  We probably should be a bit more integrated into the Tcl
community, and I'm sure there a are Tcl folks who would be interested
in what we're doing.  It doesn't feel like we'd be spamming some one
else wiki.

Doing this would maintain a very important advantage we currently have
with sourceforge: very little maintenance on our part. In fact it
would be better: zero maintenance!  Plus, presumably there would be
more people to help clean up after spammers etc.

A disadvantage might be that the Tcl wiki is a little plain. We also
don't have any control, should we need to change something.  Well, I
guess we have as much control as the rest of the Tcl community -- we
can ask.


If we did this, I imagine we would turn the current front page into a
static portal page.  It would point to key pages within the Tcl wiki
which we would maintain, along with pointers to out mailing list
interface, cvs, etc.

The rest of our current website would be auto generated.  The API
docs, we've mentioned.  We should also auto generate a page for each
module in cvs, using the DOAP file.


Is this a crazy idea?  Should we ask someone at the Tcl wiki?

Re: [naviserver-devel] Explanation of last changes

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

Auto magic i do not like myself but it seems too restrictive to have 
some dummy server definition to start server if there will be only one 
server anyway. Virtual servers are good feature for advanced use, we 
used them before when had multiple servers.

But to start one simple web server, i think it does not need to be that 
complicated, enabling module you want and have it running is good way to 
immediately see it working. I agree, making it open or less secure just 
to be able simply run is bad, but if security is not involved and no 
changes in the core for running default/complex configs, maybe it is not 
that bad once it documented and logged?

Stephen Deasey wrote:
> On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
>> - if no servers defined, it will run server "default", i.e. it silently
>> creates server "default" in config section "ns/servers"
> 
> 
> Er, no.
> 
> No crazy magic to auto-create config behind your back.  Virtual
> servers just aren't complicated.  They're a feature, and they're
> embedded deeply into the core code (all the references to char *server
> in the C API) so it's not going any where quickly.
> 
> I agree with your over all goal, which is to have a minimal config
> file able to boot and run a functional server, right?  This is
> definitely a good idea.
> 
> The current sample-config.tcl doesn't handle this case well, and it
> shouldn't -- it's supposed to be a complete reference to all config
> values.
> 
> 
> So, we need some other way to handle basic servers.  Magic config
> construction just isn't the way to do it though.  How about we move
> nsd-min.tcl you just added to contrib (dumping ground, grrr...) into
> the top level, and call it simple-config.tcl?
> 
> We'll have two config files ship by default:
> 
>   sample-config.tcl -- Exhaustive list of all config variables
>   simple-config.tcl -- Basic config needed to run a single virtual server.
> 
> They will be identical in the sense that the resulting running server
> will have the same properties, because the sample-config.tcl will be
> illustrating *optional* config values, and possibly they''' all be
> commented out.
> 
> We can rename 'server1' to 'default' if you like.
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] contrib

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

That was the reason i chnaged nsstats.tcl into page, not module, now you 
can put it anywhere under pages/ and point browser and use it instead of 
restarting server just or change something or enable/disable it.

That was the main reason not to use it on existing production servers 
because it is a module

Stephen Deasey wrote:
> On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
>> I agree overall, but i made them pages instead of tcl library and
>> because page cannot reside under tcl/ they ended up in contrib.
>> do we need special directory for index.adp and other pages, maybe pages/
>> in the CVS tree?
> 
> 
> There's only one page: index.adp.  We can just put that in the root directory.
> 
> nsstats.tcl is a module and should be in the modules/ directory of CVS.
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] contrib

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
> I agree overall, but i made them pages instead of tcl library and
> because page cannot reside under tcl/ they ended up in contrib.
> do we need special directory for index.adp and other pages, maybe pages/
> in the CVS tree?


There's only one page: index.adp.  We can just put that in the root directory.

nsstats.tcl is a module and should be in the modules/ directory of CVS.

Re: [naviserver-devel] White space changes

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

Oops, yes, my editor removes extra spaces at the end of line

Stephen Deasey wrote:
> There's something wrong with your editor Vlad. e.g.:
> 
> 
>>   /*
>> !  * More typedefs of functions
>>    */
>>
>> --- 564,568 ----
>>
>>   /*
>> !  * More typedefs of functions
>>    */
> 
> 
> The only change here is white space. But there's hundreds of them on
> every commit, which makes it completely impossible to figure what
> actually changed by following the cvs commit mailing list.
> 
> Can you fix this?
> 
> 
> 
> On 9/7/06, Vlad Seryakov <ser...@us...> wrote:
>> Update of /cvsroot/naviserver/naviserver/include
>> In directory sc8-pr-cvs7.sourceforge.net:/tmp/cvs-serv32608/include
>>
>> Modified Files:
>>         ns.h
>> Log Message:
>> A lot of small changes that will not affect normal server but will allow to
>> have minimal nsd.tcl config with working defaults
>>
>>
>> Index: ns.h
>> ===================================================================
>> RCS file: /cvsroot/naviserver/naviserver/include/ns.h,v
>> retrieving revision 1.102
>> retrieving revision 1.103
>> diff -C2 -d -r1.102 -r1.103
>> *** ns.h        14 Aug 2006 14:35:31 -0000      1.102
>> --- ns.h        7 Sep 2006 04:26:49 -0000       1.103
>> ***************
>> *** 12,16 ****
>>    * The Original Code is AOLserver Code and related documentation
>>    * distributed by AOL.
>> !  *
>>    * The Initial Developer of the Original Code is America Online,
>>    * Inc. Portions created by AOL are Copyright (C) 1999 America Online,
>> --- 12,16 ----
>>    * The Original Code is AOLserver Code and related documentation
>>    * distributed by AOL.
>> !  *
>>    * The Initial Developer of the Original Code is America Online,
>>    * Inc. Portions created by AOL are Copyright (C) 1999 America Online,
>> ***************
>> *** 170,174 ****
>>   #define NS_INT_64_FORMAT_STRING "%ld"
>>   #elif defined(_WIN32)
>> ! typedef int                   mode_t;  /* Bug: #703061 */
>>   typedef __int64                       ns_int64;
>>   typedef unsigned __int64      ns_uint64;
>> --- 170,174 ----
>>   #define NS_INT_64_FORMAT_STRING "%ld"
>>   #elif defined(_WIN32)
>> ! typedef int                   mode_t;  /* Bug: #703061 */
>>   typedef __int64                       ns_int64;
>>   typedef unsigned __int64      ns_uint64;
>> ***************
>> *** 462,466 ****
>>    * The defined value has no meaning.
>>    * It just have to exist and it needs
>> !  * to be accepted by any strerror() call.
>>    */
>>
>> --- 462,466 ----
>>    * The defined value has no meaning.
>>    * It just have to exist and it needs
>> !  * to be accepted by any strerror() call.
>>    */
>>
>> ***************
>> *** 564,568 ****
>>
>>   /*
>> !  * More typedefs of functions
>>    */
>>
>> --- 564,568 ----
>>
>>   /*
>> !  * More typedefs of functions
>>    */
>>
>> ***************
>> *** 585,589 ****
>>       (Ns_DString *dsPtr, CONST char *server, CONST char *url);
>>
>> ! typedef int (Ns_Url2FileProc)
>>       (Ns_DString *dsPtr, CONST char *url, void *arg);
>>
>> --- 585,589 ----
>>       (Ns_DString *dsPtr, CONST char *server, CONST char *url);
>>
>> ! typedef int (Ns_Url2FileProc)
>>       (Ns_DString *dsPtr, CONST char *url, void *arg);
>>
>> ***************
>> *** 659,664 ****
>>
>>   NS_EXTERN Ns_Entry *
>> ! Ns_CacheWaitCreateEntry(Ns_Cache *cache, CONST char *key, int *newPtr,
>> !                         Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1)
>>        NS_GNUC_NONNULL(2) NS_GNUC_NONNULL(3);
>>
>> --- 659,664 ----
>>
>>   NS_EXTERN Ns_Entry *
>> ! Ns_CacheWaitCreateEntry(Ns_Cache *cache, CONST char *key, int *newPtr,
>> !                         Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1)
>>        NS_GNUC_NONNULL(2) NS_GNUC_NONNULL(3);
>>
>> ***************
>> *** 687,691 ****
>>
>>   NS_EXTERN void
>> ! Ns_CacheSetValueExpires(Ns_Entry *entry, void *value, size_t size,
>>                           Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1);
>>
>> --- 687,691 ----
>>
>>   NS_EXTERN void
>> ! Ns_CacheSetValueExpires(Ns_Entry *entry, void *value, size_t size,
>>                           Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1);
>>
>> ***************
>> *** 827,830 ****
>> --- 827,833 ----
>>   Ns_ConfigGetSection(CONST char *section);
>>
>> + NS_EXTERN Ns_Set *
>> + Ns_ConfigCreateSection(CONST char *section);
>> +
>>   NS_EXTERN void
>>   Ns_GetVersion(int *major, int *minor, int *patch, int *type);
>> ***************
>> *** 1075,1079 ****
>>   Ns_ExitEventQueue(Ns_EventQueue *queue)
>>       NS_GNUC_NONNULL(1);
>> !
>>
>>   /*
>> --- 1078,1082 ----
>>   Ns_ExitEventQueue(Ns_EventQueue *queue)
>>       NS_GNUC_NONNULL(1);
>> !
>>
>>   /*
>> ***************
>> *** 1154,1158 ****
>>    * see macros above for:
>>    *
>> !  * Ns_IndexCount(X)
>>    */
>>
>> --- 1157,1161 ----
>>    * see macros above for:
>>    *
>> !  * Ns_IndexCount(X)
>>    */
>>
>> ***************
>> *** 1196,1201 ****
>>   NS_EXTERN double Ns_DRand(void);
>>
>> ! /*
>> !  * task.c:
>>    */
>>
>> --- 1199,1204 ----
>>   NS_EXTERN double Ns_DRand(void);
>>
>> ! /*
>> !  * task.c:
>>    */
>>
>> ***************
>> *** 1309,1315 ****
>>
>>   NS_EXTERN void Ns_TclSetTimeObj(Tcl_Obj *objPtr, Ns_Time *timePtr);
>> ! NS_EXTERN int Ns_TclGetTimeFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>>                                      Ns_Time *timePtr);
>> ! NS_EXTERN int Ns_TclGetTimePtrFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>>                                         Ns_Time **timePtrPtr);
>>
>> --- 1312,1318 ----
>>
>>   NS_EXTERN void Ns_TclSetTimeObj(Tcl_Obj *objPtr, Ns_Time *timePtr);
>> ! NS_EXTERN int Ns_TclGetTimeFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>>                                      Ns_Time *timePtr);
>> ! NS_EXTERN int Ns_TclGetTimePtrFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>>                                         Ns_Time **timePtrPtr);
>>
>> ***************
>> *** 1365,1369 ****
>>
>>   NS_EXTERN void
>> ! Ns_SetLogFlushProc(Ns_LogFlushProc *procPtr) NS_GNUC_DEPRECATED;
>>
>>   NS_EXTERN void
>> --- 1368,1372 ----
>>
>>   NS_EXTERN void
>> ! Ns_SetLogFlushProc(Ns_LogFlushProc *procPtr) NS_GNUC_DEPRECATED;
>>
>>   NS_EXTERN void
>> ***************
>> *** 1794,1804 ****
>>   NS_EXTERN int Ns_SockWait(SOCKET sock, int what, int timeout);
>>   NS_EXTERN int Ns_SockTimedWait(SOCKET sock, int what, Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockRecv(SOCKET sock, void *vbuf, size_t nrecv,
>>                             Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockSend(SOCKET sock, void *vbuf, size_t nsend,
>>                             Ns_Time *timeoutPtr);
>>   NS_EXTERN int Ns_SockRecvBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>>                                 Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockSendBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>>                                 Ns_Time *timeoutPtr);
>>
>> --- 1797,1807 ----
>>   NS_EXTERN int Ns_SockWait(SOCKET sock, int what, int timeout);
>>   NS_EXTERN int Ns_SockTimedWait(SOCKET sock, int what, Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockRecv(SOCKET sock, void *vbuf, size_t nrecv,
>>                             Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockSend(SOCKET sock, void *vbuf, size_t nsend,
>>                             Ns_Time *timeoutPtr);
>>   NS_EXTERN int Ns_SockRecvBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>>                                 Ns_Time *timeoutPtr);
>> ! NS_EXTERN int Ns_SockSendBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>>                                 Ns_Time *timeoutPtr);
>>
>> ***************
>> *** 1813,1817 ****
>>   NS_EXTERN SOCKET Ns_SockAsyncConnect2(char *host, int port, char *lhost, int lport);
>>   NS_EXTERN SOCKET Ns_SockTimedConnect(char *host, int port, Ns_Time *timeoutPtr);
>> ! NS_EXTERN SOCKET Ns_SockTimedConnect2(char *host, int port, char *lhost, int lport,
>>                                         Ns_Time *timeoutPtr);
>>
>> --- 1816,1820 ----
>>   NS_EXTERN SOCKET Ns_SockAsyncConnect2(char *host, int port, char *lhost, int lport);
>>   NS_EXTERN SOCKET Ns_SockTimedConnect(char *host, int port, Ns_Time *timeoutPtr);
>> ! NS_EXTERN SOCKET Ns_SockTimedConnect2(char *host, int port, char *lhost, int lport,
>>                                         Ns_Time *timeoutPtr);
>>
>>
>>
>> -------------------------------------------------------------------------
>> Using Tomcat but need to do more? Need to support web services, security?
>> Get stuff done quickly with pre-integrated technology to make your job easier
>> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
>> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>> _______________________________________________
>> naviserver-commits mailing list
>> nav...@li...
>> https://lists.sourceforge.net/lists/listinfo/naviserver-commits
>>
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] Explanation of last changes

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
>
> - if no servers defined, it will run server "default", i.e. it silently
> creates server "default" in config section "ns/servers"


Er, no.

No crazy magic to auto-create config behind your back.  Virtual
servers just aren't complicated.  They're a feature, and they're
embedded deeply into the core code (all the references to char *server
in the C API) so it's not going any where quickly.

I agree with your over all goal, which is to have a minimal config
file able to boot and run a functional server, right?  This is
definitely a good idea.

The current sample-config.tcl doesn't handle this case well, and it
shouldn't -- it's supposed to be a complete reference to all config
values.


So, we need some other way to handle basic servers.  Magic config
construction just isn't the way to do it though.  How about we move
nsd-min.tcl you just added to contrib (dumping ground, grrr...) into
the top level, and call it simple-config.tcl?

We'll have two config files ship by default:

  sample-config.tcl -- Exhaustive list of all config variables
  simple-config.tcl -- Basic config needed to run a single virtual server.

They will be identical in the sense that the resulting running server
will have the same properties, because the sample-config.tcl will be
illustrating *optional* config values, and possibly they''' all be
commented out.

We can rename 'server1' to 'default' if you like.

[naviserver-devel] White space changes

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

There's something wrong with your editor Vlad. e.g.:


>   /*
> !  * More typedefs of functions
>    */
>
> --- 564,568 ----
>
>   /*
> !  * More typedefs of functions
>    */


The only change here is white space. But there's hundreds of them on
every commit, which makes it completely impossible to figure what
actually changed by following the cvs commit mailing list.

Can you fix this?



On 9/7/06, Vlad Seryakov <ser...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/include
> In directory sc8-pr-cvs7.sourceforge.net:/tmp/cvs-serv32608/include
>
> Modified Files:
>         ns.h
> Log Message:
> A lot of small changes that will not affect normal server but will allow to
> have minimal nsd.tcl config with working defaults
>
>
> Index: ns.h
> ===================================================================
> RCS file: /cvsroot/naviserver/naviserver/include/ns.h,v
> retrieving revision 1.102
> retrieving revision 1.103
> diff -C2 -d -r1.102 -r1.103
> *** ns.h        14 Aug 2006 14:35:31 -0000      1.102
> --- ns.h        7 Sep 2006 04:26:49 -0000       1.103
> ***************
> *** 12,16 ****
>    * The Original Code is AOLserver Code and related documentation
>    * distributed by AOL.
> !  *
>    * The Initial Developer of the Original Code is America Online,
>    * Inc. Portions created by AOL are Copyright (C) 1999 America Online,
> --- 12,16 ----
>    * The Original Code is AOLserver Code and related documentation
>    * distributed by AOL.
> !  *
>    * The Initial Developer of the Original Code is America Online,
>    * Inc. Portions created by AOL are Copyright (C) 1999 America Online,
> ***************
> *** 170,174 ****
>   #define NS_INT_64_FORMAT_STRING "%ld"
>   #elif defined(_WIN32)
> ! typedef int                   mode_t;  /* Bug: #703061 */
>   typedef __int64                       ns_int64;
>   typedef unsigned __int64      ns_uint64;
> --- 170,174 ----
>   #define NS_INT_64_FORMAT_STRING "%ld"
>   #elif defined(_WIN32)
> ! typedef int                   mode_t;  /* Bug: #703061 */
>   typedef __int64                       ns_int64;
>   typedef unsigned __int64      ns_uint64;
> ***************
> *** 462,466 ****
>    * The defined value has no meaning.
>    * It just have to exist and it needs
> !  * to be accepted by any strerror() call.
>    */
>
> --- 462,466 ----
>    * The defined value has no meaning.
>    * It just have to exist and it needs
> !  * to be accepted by any strerror() call.
>    */
>
> ***************
> *** 564,568 ****
>
>   /*
> !  * More typedefs of functions
>    */
>
> --- 564,568 ----
>
>   /*
> !  * More typedefs of functions
>    */
>
> ***************
> *** 585,589 ****
>       (Ns_DString *dsPtr, CONST char *server, CONST char *url);
>
> ! typedef int (Ns_Url2FileProc)
>       (Ns_DString *dsPtr, CONST char *url, void *arg);
>
> --- 585,589 ----
>       (Ns_DString *dsPtr, CONST char *server, CONST char *url);
>
> ! typedef int (Ns_Url2FileProc)
>       (Ns_DString *dsPtr, CONST char *url, void *arg);
>
> ***************
> *** 659,664 ****
>
>   NS_EXTERN Ns_Entry *
> ! Ns_CacheWaitCreateEntry(Ns_Cache *cache, CONST char *key, int *newPtr,
> !                         Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1)
>        NS_GNUC_NONNULL(2) NS_GNUC_NONNULL(3);
>
> --- 659,664 ----
>
>   NS_EXTERN Ns_Entry *
> ! Ns_CacheWaitCreateEntry(Ns_Cache *cache, CONST char *key, int *newPtr,
> !                         Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1)
>        NS_GNUC_NONNULL(2) NS_GNUC_NONNULL(3);
>
> ***************
> *** 687,691 ****
>
>   NS_EXTERN void
> ! Ns_CacheSetValueExpires(Ns_Entry *entry, void *value, size_t size,
>                           Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1);
>
> --- 687,691 ----
>
>   NS_EXTERN void
> ! Ns_CacheSetValueExpires(Ns_Entry *entry, void *value, size_t size,
>                           Ns_Time *timeoutPtr) NS_GNUC_NONNULL(1);
>
> ***************
> *** 827,830 ****
> --- 827,833 ----
>   Ns_ConfigGetSection(CONST char *section);
>
> + NS_EXTERN Ns_Set *
> + Ns_ConfigCreateSection(CONST char *section);
> +
>   NS_EXTERN void
>   Ns_GetVersion(int *major, int *minor, int *patch, int *type);
> ***************
> *** 1075,1079 ****
>   Ns_ExitEventQueue(Ns_EventQueue *queue)
>       NS_GNUC_NONNULL(1);
> !
>
>   /*
> --- 1078,1082 ----
>   Ns_ExitEventQueue(Ns_EventQueue *queue)
>       NS_GNUC_NONNULL(1);
> !
>
>   /*
> ***************
> *** 1154,1158 ****
>    * see macros above for:
>    *
> !  * Ns_IndexCount(X)
>    */
>
> --- 1157,1161 ----
>    * see macros above for:
>    *
> !  * Ns_IndexCount(X)
>    */
>
> ***************
> *** 1196,1201 ****
>   NS_EXTERN double Ns_DRand(void);
>
> ! /*
> !  * task.c:
>    */
>
> --- 1199,1204 ----
>   NS_EXTERN double Ns_DRand(void);
>
> ! /*
> !  * task.c:
>    */
>
> ***************
> *** 1309,1315 ****
>
>   NS_EXTERN void Ns_TclSetTimeObj(Tcl_Obj *objPtr, Ns_Time *timePtr);
> ! NS_EXTERN int Ns_TclGetTimeFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>                                      Ns_Time *timePtr);
> ! NS_EXTERN int Ns_TclGetTimePtrFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>                                         Ns_Time **timePtrPtr);
>
> --- 1312,1318 ----
>
>   NS_EXTERN void Ns_TclSetTimeObj(Tcl_Obj *objPtr, Ns_Time *timePtr);
> ! NS_EXTERN int Ns_TclGetTimeFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>                                      Ns_Time *timePtr);
> ! NS_EXTERN int Ns_TclGetTimePtrFromObj(Tcl_Interp *interp, Tcl_Obj *objPtr,
>                                         Ns_Time **timePtrPtr);
>
> ***************
> *** 1365,1369 ****
>
>   NS_EXTERN void
> ! Ns_SetLogFlushProc(Ns_LogFlushProc *procPtr) NS_GNUC_DEPRECATED;
>
>   NS_EXTERN void
> --- 1368,1372 ----
>
>   NS_EXTERN void
> ! Ns_SetLogFlushProc(Ns_LogFlushProc *procPtr) NS_GNUC_DEPRECATED;
>
>   NS_EXTERN void
> ***************
> *** 1794,1804 ****
>   NS_EXTERN int Ns_SockWait(SOCKET sock, int what, int timeout);
>   NS_EXTERN int Ns_SockTimedWait(SOCKET sock, int what, Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockRecv(SOCKET sock, void *vbuf, size_t nrecv,
>                             Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockSend(SOCKET sock, void *vbuf, size_t nsend,
>                             Ns_Time *timeoutPtr);
>   NS_EXTERN int Ns_SockRecvBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>                                 Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockSendBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>                                 Ns_Time *timeoutPtr);
>
> --- 1797,1807 ----
>   NS_EXTERN int Ns_SockWait(SOCKET sock, int what, int timeout);
>   NS_EXTERN int Ns_SockTimedWait(SOCKET sock, int what, Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockRecv(SOCKET sock, void *vbuf, size_t nrecv,
>                             Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockSend(SOCKET sock, void *vbuf, size_t nsend,
>                             Ns_Time *timeoutPtr);
>   NS_EXTERN int Ns_SockRecvBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>                                 Ns_Time *timeoutPtr);
> ! NS_EXTERN int Ns_SockSendBufs(SOCKET sock, struct iovec *bufs, int nbufs,
>                                 Ns_Time *timeoutPtr);
>
> ***************
> *** 1813,1817 ****
>   NS_EXTERN SOCKET Ns_SockAsyncConnect2(char *host, int port, char *lhost, int lport);
>   NS_EXTERN SOCKET Ns_SockTimedConnect(char *host, int port, Ns_Time *timeoutPtr);
> ! NS_EXTERN SOCKET Ns_SockTimedConnect2(char *host, int port, char *lhost, int lport,
>                                         Ns_Time *timeoutPtr);
>
> --- 1816,1820 ----
>   NS_EXTERN SOCKET Ns_SockAsyncConnect2(char *host, int port, char *lhost, int lport);
>   NS_EXTERN SOCKET Ns_SockTimedConnect(char *host, int port, Ns_Time *timeoutPtr);
> ! NS_EXTERN SOCKET Ns_SockTimedConnect2(char *host, int port, char *lhost, int lport,
>                                         Ns_Time *timeoutPtr);
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> naviserver-commits mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-commits
>

Re: [naviserver-devel] contrib

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

I agree overall, but i made them pages instead of tcl library and 
because page cannot reside under tcl/ they ended up in contrib.
do we need special directory for index.adp and other pages, maybe pages/ 
in the CVS tree?


Stephen Deasey wrote:
> Looks like nsstats got moved to the contrib directory.  Contrib is
> generally a bad idea because it becomes a dumping ground for random
> stuff.
> 
> Two files in there, nsstats.tcl and nsconfig.tcl, need to be turned
> into modules. They are in fact modules, but kind of crappy ones
> because there's no README file, no Makefile to place them in the
> correct position, etc.
> 
> Modules aren't second class citizens. Code in a module is not intended
> to imply it is lesser quality, or not supported, or whatever.  The
> only difference is the *advantage* that the release schedule for a
> module is not tied to that of the server.
> 
> (The name of nsconfig should probably be something less generic).
> 
> Also, the list of mime types in contrib is confusing (it confused one
> new list member at least) and now redundant. All mime types are
> compiled in, and there's a new command [ns_info mimetypes] for
> introspection.  Can this be removed?
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] contrib

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

Looks like nsstats got moved to the contrib directory.  Contrib is
generally a bad idea because it becomes a dumping ground for random
stuff.

Two files in there, nsstats.tcl and nsconfig.tcl, need to be turned
into modules. They are in fact modules, but kind of crappy ones
because there's no README file, no Makefile to place them in the
correct position, etc.

Modules aren't second class citizens. Code in a module is not intended
to imply it is lesser quality, or not supported, or whatever.  The
only difference is the *advantage* that the release schedule for a
module is not tied to that of the server.

(The name of nsconfig should probably be something less generic).

Also, the list of mime types in contrib is confusing (it confused one
new list member at least) and now redundant. All mime types are
compiled in, and there's a new command [ns_info mimetypes] for
introspection.  Can this be removed?

Re: [naviserver-devel] Explanation of last changes

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

On 9/7/06, Vlad Seryakov <vl...@cr...> wrote:
>
>   - if nscp loaded without settings, it will listen on 127.0.0.1 port
> 2080 and will accept connections with empty username and password. I
> think this is reasonable for development server.


This is insecure as a default and we have to change it.  The nscp
module should *not* be loaded by default, and we shouldn't be hard
coding default user names and passwords. We espescially shouldn't be
doing both at the same time.

It's no trouble at all to remove a single '#' to enable a module.

This also applies to the nsstats functionality -- off by default please.

Re: [naviserver-devel] command list

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

> > Or do you think of creating  a list of sections with lists of the
> > commands?
>
> Yes, like in any book:-))

ok, the file is now in doc/src/commandlist.man.

The name should be changed, of course. And I (mis)used the [arg] parameter to 
indicate TCL commands, this is also to be changed.

Bernd.

Re: [naviserver-devel] command list

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

> 
> Or do you think of creating  a list of sections with lists of the commands?
> 

Yes, like in any book:-))

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

Re: [naviserver-devel] command list

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

> why it will be deleted, make it TOC in doctools format, we will figure
> out ot to do crossreferencing/linking in doctools but TOc will be a good
> start, we will have to follow this from top up to individual command or
> other document. Commit it into doc/src.

ok... but I must admit I have not create a TOC before. Can you point me to an 
example TOC-doc-file where I can borrow the structure?

Or do you think of creating  a list of sections with lists of the commands?

Bernd.

Re: [naviserver-devel] command list

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

why it will be deleted, make it TOC in doctools format, we will figure 
out ot to do crossreferencing/linking in doctools but TOc will be a good 
start, we will have to follow this from top up to individual command or 
other document. Commit it into doc/src.

Bernd Eidenschink wrote:
> I have the list of commands now updated to the latest code 4.99.2 and 
> categorized most of them.
> 
> I would ask you to review the list, and re-categorize if I put some commands 
> where they not belong. Or delete, if the command is a "private" helper.
> 
> Guess it's better not to post the list (file) but to commit it. 
> Where to? Below doc/src, into doc, into contrib? It's just for now, the list 
> will be deleted finally.
> 
> After we consider the list to be complete and correct, we can allocate 
> duties :-)
> 
> Bernd.
> 
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> 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] command list

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

I have the list of commands now updated to the latest code 4.99.2 and 
categorized most of them.

I would ask you to review the list, and re-categorize if I put some commands 
where they not belong. Or delete, if the command is a "private" helper.

Guess it's better not to post the list (file) but to commit it. 
Where to? Below doc/src, into doc, into contrib? It's just for now, the list 
will be deleted finally.

After we consider the list to be complete and correct, we can allocate 
duties :-)

Bernd.

Re: [naviserver-devel] Documentation & website question

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

> I strongly beleive we should start with doc sources in
> doctools and then use converters to auto-generate other
> (handful of) formats.

I also vote for that. I'm creating an updated list of commands to be 
documented and will post them soon.

Bernd.

Re: [naviserver-devel] Documentation & website question

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

On 06.09.2006, at 23:15, Stephen Deasey wrote:

> On 9/5/06, Zoran Vasiljevic <zv...@ar...> wrote:
>>
>> On 05.09.2006, at 19:07, Bernd Eidenschink wrote:
>>
>>>
>>> http://wiki2man.sourceforge.net/
>>
>> Almost (although you are right)! What we need is
>> wiki->doctools man-pages not nroff man pages.
>> (what a stupid naming mess !!!)
>>
>> I wonder if there is a wiki->doctools as I know
>> there is doctools->wiki converter...
>
>
> If someone adds some Makefile magic to generate nroff and HTML from
> the doctools source, I'll set up sourceforge to auto generate the HTML
> straight from CVS and make it available on the website.

This may not be the problem. I will try (honestly!) to make some
directives in the makefile for that. It should be something like

  make doc
  make mandoc
  make htmldoc

where make doc will make both and the others will make
only the corresponding formats.
I will put the putput into the

   doc/man
   doc/html

and the source files (mainly templates for now) into

   doc/src

>
> There's a script in there now which creates symbolic links for all the
> commands embedded in a singe file, so if foo, bar, and baz are all
> documented in the foo.man file, you can still access bar and baz
> directly via 'man bar' etc.  Does doctools support anything like this?

I do not thing they can but I'm not 100% sure. I'm afraid it will
be 1:1 (one procedure per file)

>
> We can add some wiki-macros so that whenever you type [proc ns_foo] it
> creates a link to /doc/ns_foo.html (or wherever).
>
> If the API doc is always automatically uptodate on the website then
> that can be the single canonical location to find everything.
>
>
> Is there a script which will convert the existing man pages to
> doctools format?  We can't be the first people to ever have done this.

I do not know. I do not believe there are is such tool ready.
Mostly people either have no content and they start with some
or they already have large existing base in which case they stay
with it. We have few in man, few in wiki, none in doctools.

The reason why I'm for doctools is:

   its Tcl
   its easy to write
   its easy to convert to wiki/html

>
> If we can convert the existing stuff, then strip out the
> old/deprecated APIs and stubb out the new ones, I think finding time
> to sit down and write something will be much easier.

I believe that converting the existing things is always
going to require manual intervention.
As of today we have:

   html pages from old aolserver docs
   wiki pages from new aolserver docs
   some nroff pages from old aolserver docs

The task to convert all to doctools and then generate
html and man (and wiki) is not small, but once you have
done it, you can automate whatever you want. Plus: the
doctools itself is written in Tcl so it would be easy
for us to tweak it if necessary.

I strongly beleive we should start with doc sources in
doctools and then use converters to auto-generate other
(handful of) formats.


Zoran

Re: [naviserver-devel] Documentation & website question

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

> If the API doc is always automatically uptodate on the website then
> that can be the single canonical location to find everything.

A little grouching from me: Not now, but sometime later, we need a faster 
webpage. No matter what:  a dedicated server, a new or version of the Wiki or 
another Wiki that allows caching in the sense of static HTML pages, or no 
Wiki at all and from whatever produced HTML pages. 
But currently THIS wiki is so slow at times, so r e a l l y slow, it kills 
time you could better invest in doing real work.

Bernd.