naviserver-devel — Developer discussion

Re: [naviserver-devel] nslog

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

> > The last log entry is from yesterday 23:04 (when the last developer went
> > home), no new access file since then.
>
> Did the partition fill up?
>
>   $ df -h
>
> Or maybe the directory where the log files live was deleted/moved/had
> it's permissions changed...?

No, enough space; no logrotate stuff that could interfere, no cronjobs that 
change permissions, no rsync from other servers or things like that...
The access-log is in the same directory as the server.log and the pidfile
(logs/) on this server.

Re: [naviserver-devel] SLES 10

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

On 9/10/07, Bernd Eidenschink <eid...@we...> wrote:
>
> > Install autoconf >= 2.59c
> >
> > Suse 10.2 ships with autoconf-2.60
>
> This is a SuSE 10 Enterprise Server numbered 'autoconf-2.59-92.2'.
> It will not be that easy to update, but i'll try.
>
> > Hmm, maybe we just need to copy those macros into ./m4 for now. I
> > think Zoran had this problem as well...
>
> Ok, seems the easier way :-)
> Is it just copying the defs into a file within ./m4? Guess not...


Yep, that's all it should take.

Here's where they live within the autoconf distribution:

http://cvs.savannah.gnu.org/viewvc/autoconf/lib/autoconf/types.m4?root=autoconf&view=markup

Re: [naviserver-devel] nslog

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

On 9/10/07, Bernd Eidenschink <eid...@we...> wrote:
>
> Hi Stephen,
>
> > Are there any other log error messages?  I'm guessing either the
> > access log didn't open or didn't reopen after a role.
>
> That seems to be the case:
>
> > (This particular error message looks like it's always going to be wrong).
>
> Error: nslog: flush failed: 'No such file or directory'
> Error: nslog: flush failed: 'Resource temporarily unavailable'
> Error: nslog: flush failed: 'Inappropriate ioctl for device'
>
> The last log entry is from yesterday 23:04 (when the last developer went
> home), no new access file since then.


Did the partition fill up?

  $ df -h

Or maybe the directory where the log files live was deleted/moved/had
it's permissions changed...?

Re: [naviserver-devel] SLES 10

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

> Install autoconf >= 2.59c
>
> Suse 10.2 ships with autoconf-2.60

This is a SuSE 10 Enterprise Server numbered 'autoconf-2.59-92.2'.
It will not be that easy to update, but i'll try.

> Hmm, maybe we just need to copy those macros into ./m4 for now. I
> think Zoran had this problem as well...

Ok, seems the easier way :-)
Is it just copying the defs into a file within ./m4? Guess not...

Bernd.

Re: [naviserver-devel] nslog

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

Hi Stephen,

> Are there any other log error messages?  I'm guessing either the
> access log didn't open or didn't reopen after a role.

That seems to be the case:

> (This particular error message looks like it's always going to be wrong).

Error: nslog: flush failed: 'No such file or directory'
Error: nslog: flush failed: 'Resource temporarily unavailable'
Error: nslog: flush failed: 'Inappropriate ioctl for device'

The last log entry is from yesterday 23:04 (when the last developer went 
home), no new access file since then.

Bernd.

Re: [naviserver-devel] SLES 10

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

On 9/10/07, Bernd Eidenschink <eid...@we...> wrote:
>
> Hi,
>
> on a  SuSE Linux 10 Enterprise Server i ran into this when trying to compile
> from HEAD:
>
> ./autogen.sh --with-tcl=/usr/local/ns/lib --prefix=/usr/local/ns
>
> Running aclocal -I m4 -I /opt/gnome/share/aclocal
> Running autoheader
> Running autoconf
> configure.in:172: error: possibly undefined macro: AC_TYPE_INT8_T
>       If this token and others are legitimate, please use m4_pattern_allow.
>       See the Autoconf documentation.
> configure.in:173: error: possibly undefined macro: AC_TYPE_INT16_T
> configure.in:174: error: possibly undefined macro: AC_TYPE_INT32_T
> configure.in:175: error: possibly undefined macro: AC_TYPE_INT64_T
> configure.in:177: error: possibly undefined macro: AC_TYPE_UINT8_T
> configure.in:178: error: possibly undefined macro: AC_TYPE_UINT16_T
> configure.in:179: error: possibly undefined macro: AC_TYPE_UINT32_T
> configure.in:180: error: possibly undefined macro: AC_TYPE_UINT64_T
> configure.in:182: error: possibly undefined macro: AC_TYPE_INTMAX_T
> configure.in:183: error: possibly undefined macro: AC_TYPE_UINTMAX_T
> configure.in:185: error: possibly undefined macro: AC_TYPE_INTPTR_T
> configure.in:186: error: possibly undefined macro: AC_TYPE_UINTPTR_T
>
> Do you know what to do?


Install autoconf >= 2.59c

Suse 10.2 ships with autoconf-2.60


Hmm, maybe we just need to copy those macros into ./m4 for now. I
think Zoran had this problem as well...

Re: [naviserver-devel] nslog

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

On 9/10/07, Bernd Eidenschink <eid...@we...> wrote:
>
> Hi!
>
> I see some messages like
>
> [-conn:server1:8] Error: nslog: flush failed: 'Success'
> [-conn:server1:7] Error: nslog: flush failed: 'Success'
> [-conn:server1:6] Error: nslog: flush failed: 'Success'
> [-conn:server1:8] Error: nslog: flush failed: 'Success'
>
> in the server logfile.
>
> The nslog code part seems to be:
>
>     if (logPtr->maxlines == 0) {
>         status = LogFlush(logPtr, &ds);
>     } else {
>         Ns_DStringNAppend(&logPtr->buffer, ds.string, ds.length);
>         if (++logPtr->curlines > logPtr->maxlines) {
>             status = LogFlush(logPtr, &logPtr->buffer);
>             logPtr->curlines = 0;
>         } else {
>             status = NS_OK;
>         }
>     }
>
>     Ns_MutexUnlock(&logPtr->lock);
>     Ns_DStringFree(&ds);
>
>     if (status != NS_OK) {
>         Ns_Log(Error, "nslog: flush failed: '%s'", strerror(errno));
>     }
>
> Do you know why it fails with success? :-)


Why do birds, suddenly appear... ?   :-)

Are there any other log error messages?  I'm guessing either the
access log didn't open or didn't reopen after a role.

(This particular error message looks like it's always going to be wrong).

[naviserver-devel] SLES 10

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

Hi,

on a  SuSE Linux 10 Enterprise Server i ran into this when trying to compile 
from HEAD:

./autogen.sh --with-tcl=/usr/local/ns/lib --prefix=/usr/local/ns

Running aclocal -I m4 -I /opt/gnome/share/aclocal
Running autoheader
Running autoconf
configure.in:172: error: possibly undefined macro: AC_TYPE_INT8_T
      If this token and others are legitimate, please use m4_pattern_allow.
      See the Autoconf documentation.
configure.in:173: error: possibly undefined macro: AC_TYPE_INT16_T
configure.in:174: error: possibly undefined macro: AC_TYPE_INT32_T
configure.in:175: error: possibly undefined macro: AC_TYPE_INT64_T
configure.in:177: error: possibly undefined macro: AC_TYPE_UINT8_T
configure.in:178: error: possibly undefined macro: AC_TYPE_UINT16_T
configure.in:179: error: possibly undefined macro: AC_TYPE_UINT32_T
configure.in:180: error: possibly undefined macro: AC_TYPE_UINT64_T
configure.in:182: error: possibly undefined macro: AC_TYPE_INTMAX_T
configure.in:183: error: possibly undefined macro: AC_TYPE_UINTMAX_T
configure.in:185: error: possibly undefined macro: AC_TYPE_INTPTR_T
configure.in:186: error: possibly undefined macro: AC_TYPE_UINTPTR_T

Do you know what to do?

Bernd.

[naviserver-devel] nslog

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

Hi!

I see some messages like

[-conn:server1:8] Error: nslog: flush failed: 'Success'
[-conn:server1:7] Error: nslog: flush failed: 'Success'
[-conn:server1:6] Error: nslog: flush failed: 'Success'
[-conn:server1:8] Error: nslog: flush failed: 'Success'

in the server logfile.

The nslog code part seems to be:

    if (logPtr->maxlines == 0) {
        status = LogFlush(logPtr, &ds);
    } else {
        Ns_DStringNAppend(&logPtr->buffer, ds.string, ds.length);
        if (++logPtr->curlines > logPtr->maxlines) {
            status = LogFlush(logPtr, &logPtr->buffer);
            logPtr->curlines = 0;
        } else {
            status = NS_OK;
        }
    }

    Ns_MutexUnlock(&logPtr->lock);
    Ns_DStringFree(&ds);

    if (status != NS_OK) {
        Ns_Log(Error, "nslog: flush failed: '%s'", strerror(errno));
    }

Do you know why it fails with success? :-)

Bernd.

Re: [naviserver-devel] [naviserver-commits] naviserver ChangeLog, 1.678, 1.679

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

Nice catch, works perfectly now

Stephen Deasey wrote:
> Update of /cvsroot/naviserver/naviserver
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18813
> 
> Modified Files:
> 	ChangeLog 
> Log Message:
> 	* nsd/connio.c (ConnSend):
> 	* tests/http.test:
> 	* tests/testserver/pages/16480bytes: Add a call to
> 	Ns_ConnSetLengthHeader() at the start of ConnSend() with the total
> 	length of the file to send, otherwise Content-Length gets reported as
> 	8192 which is the size of the buffer (IOBUFSZ) used to read the file off
> 	disk.
> 
> 
> Index: ChangeLog
> ===================================================================
> RCS file: /cvsroot/naviserver/naviserver/ChangeLog,v
> retrieving revision 1.678
> retrieving revision 1.679
> diff -C2 -d -r1.678 -r1.679
> *** ChangeLog	21 Aug 2007 04:11:21 -0000	1.678
> --- ChangeLog	22 Aug 2007 16:17:08 -0000	1.679
> ***************
> *** 1,2 ****
> --- 1,12 ----
> + 2007-08-22  Stephen Deasey  <sd...@us...>
> + 
> + 	* nsd/connio.c (ConnSend):
> + 	* tests/http.test:
> + 	* tests/testserver/pages/16480bytes: Add a call to
> + 	Ns_ConnSetLengthHeader() at the start of ConnSend() with the total
> + 	length of the file to send, otherwise Content-Length gets reported as
> + 	8192 which is the size of the buffer (IOBUFSZ) used to read the file off
> + 	disk.
> + 
>   2007-08-20  Vlad Seryakov  <ser...@us...>
>   
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >>  http://get.splunk.com/
> _______________________________________________
> naviserver-commits mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-commits
> 

-- 
Vlad Seryakov
vl...@cr...
http://www.crystalballinc.com/vlad/

[naviserver-devel] Broken last version

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

Something is not right with the last CVS version, i compiled and now it 
works kind of weird, cannot even return JPEG images properly. Sometime 
it is fine, sometimes it is not, not sure where to look. Can it be 
something with latest changes in encoding/flushing?

Below is image i am trying to retrieve, i got only first 8K of it

-rw-r--r-- 1 root root 53955 2007-08-21 22:15 2098.jpg

vseryakov:vlad[22:17:49]#wget -d http://192.168.0.10/2098.jpg

DEBUG output created by Wget 1.10.2 on linux-gnu.

--22:17:56--  http://192.168.0.10/Covers/2098.jpg
            => `2098.jpg.1'
Connecting to 192.168.0.10:80... connected.
Created socket 3.
Releasing 0x08086b50 (new refcount 0).
Deleting unused 0x08086b50.

---request begin---
GET /Covers/2098.jpg HTTP/1.0
User-Agent: Wget/1.10.2
Accept: */*
Host: 192.168.0.10
Connection: Keep-Alive

---request end---
HTTP request sent, awaiting response...
---response begin---
HTTP/1.0 200 OK
MIME-Version: 1.0
Accept-Ranges: bytes
Server: NaviServer/4.99.2
Date: Wed, 22 Aug 2007 02:20:22 GMT
Last-Modified: Wed, 22 Aug 2007 02:15:21 GMT
Content-Type: image/jpeg
Content-Length: 8192
Connection: keep-alive

---response end---
200 OK
Registered socket 3 for persistent reuse.
Length: 8,192 (8.0K) [image/jpeg]

100%[============================================================================>] 
8,192         --.--K/s

22:17:56 (14.48 MB/s) - `2098.jpg.1' saved [8192/8192]


-- 
Vlad Seryakov
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] [naviserver-commits] naviserver/nsd adprequest.c, 1.26, 1.27 fastpath.c, 1.37, 1.38 nsd.h, 1.102, 1.103

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

No reason but i chose that one because TCL use it and our tcl.m4 file 
sets it with TCL.

Zoran Vasiljevic wrote:
> Am 21.08.2007 um 06:11 schrieb Vlad Seryakov:
> 
>> Update of /cvsroot/naviserver/naviserver/nsd
>> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18391/nsd
>>
>> Modified Files:
>> 	adprequest.c fastpath.c nsd.h
>> Log Message:
>>         * nsd/fastpath.c:
>>         * nsd/nsd.h: Added suppoirt for stat64 in FastStat. It uses  
>> TCL
>>         defined HAVE_STRUCT_STAT64
> 
> Why not using LARGEFILE64_SOURCE ?
> 
>      http://www.delorie.com/gnu/docs/glibc/libc_13.html
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >>  http://get.splunk.com/
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] [naviserver-commits] naviserver/nsd adprequest.c, 1.26, 1.27 fastpath.c, 1.37, 1.38 nsd.h, 1.102, 1.103

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

Am 21.08.2007 um 06:11 schrieb Vlad Seryakov:

> Update of /cvsroot/naviserver/naviserver/nsd
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18391/nsd
>
> Modified Files:
> 	adprequest.c fastpath.c nsd.h
> Log Message:
>         * nsd/fastpath.c:
>         * nsd/nsd.h: Added suppoirt for stat64 in FastStat. It uses  
> TCL
>         defined HAVE_STRUCT_STAT64

Why not using LARGEFILE64_SOURCE ?

     http://www.delorie.com/gnu/docs/glibc/libc_13.html

[naviserver-devel] Defaults in nsdb

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

Does it make sense to make them longer values?


     poolPtr->maxidle = Ns_ConfigIntRange(path, "maxidle", 600, 0, INT_MAX);
     poolPtr->maxopen = Ns_ConfigIntRange(path, "maxopen", 3600, 0, 
INT_MAX);

Re: [naviserver-devel] find -maxdepth

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

Hey, no problem, make sure it still works on Linux :-)))

Ibrahim Tannir wrote:
> Speaking of KISS:
> 
> for i in doc/html/*.html; do
> 
> Greetings,
> Ibrahim
> 
> 
> 
> Zoran Vasiljevic wrote:
>> Am 14.08.2007 um 11:15 schrieb Bernd Eidenschink:
>>
>>> ==========================
>>> for i in `ls doc/html/*.html`; do
>>> ==========================
>> Yup! Sometimes we are shooting with 120mm canons
>> on sparows...
>>
>> Cheers,
>> Zoran
> 
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >>  http://get.splunk.com/
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] find -maxdepth

[Ibrahim T. <it...@we...>]

Speaking of KISS:

for i in doc/html/*.html; do

Greetings,
Ibrahim



Zoran Vasiljevic wrote:
> Am 14.08.2007 um 11:15 schrieb Bernd Eidenschink:
> 
>> ==========================
>> for i in `ls doc/html/*.html`; do
>> ==========================
> 
> Yup! Sometimes we are shooting with 120mm canons
> on sparows...
> 
> Cheers,
> Zoran

Re: [naviserver-devel] find -maxdepth

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

Am 14.08.2007 um 11:15 schrieb Bernd Eidenschink:

> ==========================
> for i in `ls doc/html/*.html`; do
> ==========================

Yup! Sometimes we are shooting with 120mm canons
on sparows...

Cheers,
Zoran

Re: [naviserver-devel] find -maxdepth

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

Hi Zoran,

> <           for i in `find doc/html -maxdepth 1 -name "*.html" -
> print`; do \

> I want to remove the "-maxdepth" from Makefile as this is not
> supported on all platforms (Solaris for example).
> Is this OK for you?

or a simple

==========================
for i in `ls doc/html/*.html`; do
==========================

cu
BE

[naviserver-devel] find -maxdepth

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

Hi!

-bash-3.00$ cvs diff Makefile
vas...@na...'s password:
Index: Makefile
===================================================================
RCS file: /cvsroot/naviserver/naviserver/Makefile,v
retrieving revision 1.60
diff -r1.60 Makefile
111c111
<           for i in `find doc/html -maxdepth 1 -name "*.html" - 
print`; do \
---
 >           for i in `find doc/html -name "*.html" -print`; do \
114c114
<           for i in `find doc/html -maxdepth 1 -name "*.css" - 
print`; do \
---
 >           for i in `find doc/html -name "*.css" -print`; do \


I want to remove the "-maxdepth" from Makefile as this is not
supported on all platforms (Solaris for example).
Is this OK for you?

Cheers
Zoran

Re: [naviserver-devel] Proposal: just an idea about config syntax

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

Sure, i am not advocating for config syntax change

Stephen Deasey wrote:
> On 8/11/07, Vlad Seryakov <vl...@cr...> wrote:
>> If the only problem is comments that can be done easy.
>> Benefit to make config easier, if you check mailing list NS/As, most
>> question are related to configuration, for web admins who is not
>> familiar with Tcl this is a showstopper, plain text config would allow
>> just use the server even for static pages. Tcl config is cool developers
>> idea but not really practical for the admins and non-developers.
>>
> 
> 
> There certainly are a lot of questions about configuring the server. I
> don't think I've seen one about ns_section/ns_param.
> 
> The things that are hard about configuring the server are that there
> are a lot of knobs, it's not obvious what they do, or how they
> interact, and there's little documentation.
> 
> Some things are hard because we are asking basically unknowable
> questions, like how many threads should should there be. Who knows?
> 
> I don't think alternate syntax addresses any of these question, and
> worse, it introduces all kinds of gotchas so you can no longer explain
> it in a sentence as you can ns_section/ns_param.
> 
> But, it's 5 lines of code at the top of your config file so everyones
> options are still open. Right?
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >>  http://get.splunk.com/
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Proposal: just an idea about config syntax

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

On 8/11/07, Vlad Seryakov <vl...@cr...> wrote:
> If the only problem is comments that can be done easy.
> Benefit to make config easier, if you check mailing list NS/As, most
> question are related to configuration, for web admins who is not
> familiar with Tcl this is a showstopper, plain text config would allow
> just use the server even for static pages. Tcl config is cool developers
> idea but not really practical for the admins and non-developers.
>


There certainly are a lot of questions about configuring the server. I
don't think I've seen one about ns_section/ns_param.

The things that are hard about configuring the server are that there
are a lot of knobs, it's not obvious what they do, or how they
interact, and there's little documentation.

Some things are hard because we are asking basically unknowable
questions, like how many threads should should there be. Who knows?

I don't think alternate syntax addresses any of these question, and
worse, it introduces all kinds of gotchas so you can no longer explain
it in a sentence as you can ns_section/ns_param.

But, it's 5 lines of code at the top of your config file so everyones
options are still open. Right?

Re: [naviserver-devel] Proposal: just an idea about config syntax

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

On 8/11/07, Vlad Seryakov <vl...@cr...> wrote:
> >
> >
> > (btw. I updated the wiki software and SF updated their servers. Seems
> > pretty responsive.)
> >
>
> It is much faster, is it possibly to import our documentation so we will
> have one common place for all docs?
>


The documentation in doc/src/* that isn't really a man page?  Yeah,
sounds like a good idea.

dtplite can export in a wiki format I think...

Re: [naviserver-devel] Proposal: just an idea about config syntax

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

If the only problem is comments that can be done easy.
Benefit to make config easier, if you check mailing list NS/As, most 
question are related to configuration, for web admins who is not 
familiar with Tcl this is a showstopper, plain text config would allow 
just use the server even for static pages. Tcl config is cool developers 
idea but not really practical for the admins and non-developers.


Zoran Vasiljevic wrote:
> Am 11.08.2007 um 15:15 schrieb Stephen Deasey:
> 
>> One problem with this is you can't add comments:
>>
>> section fastpath {
>>     cache false    ;# Turn off caching
>>     ...
>> }
> 
> Well, you could be overriding the [unknown] and
> eval'in the whole block. This would implicitly
> mean that you should not have config option named
> after any built-in Tcl command, yes.
> 
> Altogether, I do not know if this all is worth
> the trouble. It looks nicer, true. But is it
> worth it?
> 
> Cheers
> Zoran
> 
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >>  http://get.splunk.com/
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel
> 

Re: [naviserver-devel] Proposal: just an idea about config syntax

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

> 
> 
> (btw. I updated the wiki software and SF updated their servers. Seems
> pretty responsive.)
> 

It is much faster, is it possibly to import our documentation so we will 
have one common place for all docs?

[naviserver-devel] Tcl inits (continued)

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

Hi!

Too bad we do not consider that lazyloader
in more detail...

(introspective script on plain NS install)

server1:nscp 2> string len [ns_ictl get]
142393

(trace-based script on plain NS install)

server1:nscp 1> string len [ns_ictl get]
40982

Needles to say that it loads much quicker
and consumes far less memory...

I would really like to get some exposure to
"real world" apps with that (hey, our works
perfectly with it)... I'm sure there will be
some "rough-edges" but I'm prepared to go
thru to fix them.

On the other side, I will continue to improve
the introspective script generator and define
callbacks within, that packages can register
and/or override, so we can better handle stuff
like XOTcl and similar.

Cheers
Zoran