Nabble - mod_tcl

Re: mod_rivet breaks mod_cgi on apache2?

2009-11-23T10:50:45Z

I went for a quick run of 'gdb bin/httpd' to see what happened in
Rivet_SendContent and it seems to be working ok returning DECLINED upon
a request for a cgi script. For what I know this is what Rivet is
supposed to do. The request seems to be lost forever and apache even
doesn't log it down. Only the error log has entries about timeout
errors. Another thing that might be worth noticing is that quitting the
debugging session gdb warns about a child process that has to be killed
(never got that message debugging rivet/tcl requests). Therefore Apache
could have got to the stage of forking the process for the cgi script,
but for some reason the process is unable to produce any content (or
headers).

I wonder if other functions (hooks) of Rivet could be confusing apache.
In recent years other modules were the culprits of similar problems.
Some googling shows that mod_sql_log (IIRC) was blocking the content
generation of cgi-scripts and simply removing it restored the expected
behavior.

I don't have much time these days, i have to put it off for a while.

-- Massimo

David Welton wrote:
>
> Have you tried "eliminating" bits and pieces of mod_rivet (commenting
> them out or something) to see if you can figure out what exactly is
> causing the problem?
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: mod_rivet breaks mod_cgi on apache2?

2009-11-21T08:20:02Z

> I found out that when mod_rivet is loaded, cgi scripts (namely mailman's cgi
> script) stop working altogether letting a browser wait forever for output.
> Removing mod_rivet restores usual cgi scripts functionality. I checked also
> that it isn't specific to mailman's scripts (which are wrapped within binary
> code): even a typical "hello world" perl script would produce the same
> behavior.

> I've briefly checked in Rivet_SendContent how mod_rivet actually gives up a
> request to the framework and could not find any evident flaw in that.
> Before I dive into a debugging session I ask whether anyone has tips this
> issue. It would be very much appreciated.

Have you tried "eliminating" bits and pieces of mod_rivet (commenting
them out or something) to see if you can figure out what exactly is
causing the problem?

--
David N. Welton

http://www.welton.it/davidw/

http://www.dedasys.com/
Sent from Innsbruck, Tyrol, Austria

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

mod_rivet breaks mod_cgi on apache2?

2009-11-20T03:49:51Z

Hi fellow Rivetters

I found out that when mod_rivet is loaded, cgi scripts (namely mailman's
cgi script) stop working altogether letting a browser wait forever for
output. Removing mod_rivet restores usual cgi scripts functionality. I
checked also that it isn't specific to mailman's scripts (which are
wrapped within binary code): even a typical "hello world" perl script
would produce the same behavior.

I've briefly checked in Rivet_SendContent how mod_rivet actually gives
up a request to the framework and could not find any evident flaw in
that. Before I dive into a debugging session I ask whether anyone has
tips this issue. It would be very much appreciated.

-- Massimo

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

wrong subcommand list in Rivet_Upload

2009-11-03T10:03:44Z

I found out that in function Rivet_Upload the list of subcommands had 2
entries for 'tempname". In case a 'upload names' command had to be processed
Tcl_GetIndexfromObject mismatches the entry for the 'names' subcommand with
its numerical counterpart. This causes the child process to segfault because
Rivet_Upload tries to get one more argument whereas there is none.

The bug surfaced when I reviewed more closely the output of the failing tests.

i'm going to commit right away.

-- Massimo

Index: src/apache-2/rivetCore.c
===================================================================
--- src/apache-2/rivetCore.c (revision 832481)
+++ src/apache-2/rivetCore.c (working copy)
@@ -836,7 +836,6 @@
"filename",
"tempname",
"names",
- "tempname",
NULL
};

@@ -848,8 +847,8 @@
SIZE,
TYPE,
FILENAME,
- NAMES,
- TEMPNAME
+ TEMPNAME,
+ NAMES
};

rivet_interp_globals *globals = Tcl_GetAssocData(interp, "rivet", NULL);
@@ -863,7 +862,8 @@

/* If it's any of these, we need to find a specific name. */

- /* Excluded cases are EXISTS and NAMES. */
+ /* Excluded case is NAMES. */
+
if ((enum subcommand)subcommandindex == CHANNEL ||
(enum subcommand)subcommandindex == SAVE ||
(enum subcommand)subcommandindex == DATA ||

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

test suite: fixed apachetest.tcl

2009-11-02T02:48:12Z

I fixed a function in apachetest.tcl that failed to interpret correctly
some 'Include' directives and generated a crippled configuration file.
The changes I made fixed the test suite on my apache installation, I
have no knowledge of other problems that could have been observed before
2007 when I began to be involved in the project.

Several tests fail and I haven't looked yet into the specific problem
addressed by each of them. Before starting a review of the failing test
I wish to commit the fixed file, in case someone wants to try himself to
run it.

regards

-- Massimo

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: how to use apache_log_error?

2009-10-29T03:41:16Z

> I noticed Karl added apache_log_error, how would I use that in an ErrorScript
> in order to pass the entire interpreter error string to the apache log
> (basically pass what rivet normally spits to the screen)? (Basically, can
> I substitute some internal variable to the message argument of
> apache_log_error?)
>
> (There appears to be a dearth of documentation surrounding Rivet...)

Looks like it's used like so:

apache_log_error loglevel message

So in order to catch an error message and spit it out elsewhere, you
can use the same variables that are in the default error handler:

proc handle_error {} {
global errorInfo
global errorOutbuf

puts <PRE>
puts "<HR>$errorInfo<HR>"
puts "<P><B>OUTPUT BUFFER:</B></P>"
puts $errorOutbuf
puts </PRE>
}

Does that work?

--
David N. Welton

http://www.welton.it/davidw/

http://www.dedasys.com/
Sent from Padua, Veneto, Italy

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

how to use apache_log_error?

2009-10-27T06:23:03Z

I noticed Karl added apache_log_error, how would I use that in an ErrorScript
in order to pass the entire interpreter error string to the apache log
(basically pass what rivet normally spits to the screen)? (Basically, can
I substitute some internal variable to the message argument of
apache_log_error?)

(There appears to be a dearth of documentation surrounding Rivet...)

- newb at rivet
--
===========================================================
Peter C. Lai | Bard College at Simon's Rock
Systems Administrator | 84 Alford Rd.
Information Technology Svcs. | Gt. Barrington, MA 01230 USA
peter AT simons-rock.edu | (413) 528-7428
===========================================================

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: Question about interpreters

2009-10-26T08:50:07Z

> I have the following problem(s)

[...]

> I don't know what the heck is wrong, or how to use the interpreters
> or why they're needed, where they're needed .. how to combine them
> cause it seems not to display my pages correctly. I read the quick
> reference before but it took me lots of hours to figure out some
> things. Could it be because I'm on vista? Even sessions seemed to
> work weird, and without a proper interpreter I felt lost so I
> stopped using Websh.

I don't see anything that's obviously wrong in your code. (Well, there
isn't much code ;-)

Does the sample installation w/o changing anything (except maybe
settning webshroot correctly in websh.conf) work? What is the
output of using the sample as-is? Any logs in the Apache error log?

The poin twith the interpreters is: with

proc web::interpmap {file} {
# return interp class for request to file
return $file
}

you define which interpreter(s) handle which requests. The sample uses
2 defined interpreter classes (i.e. interpreterr that run the same
code for same set of requests) and all *.ws3 requests are handled by
their own class):

otherhandler.ws3: running requests to any file called other.html

htmlhandler.ws3: running requests to any *.html file (except the above)

*.ws3: each *.ws3 has it's own interpreter, which just runs the file itself

So the interpreter thing lets you decide which requests should be
handled by which code.

proc web::interpmap {file} {
return myCode.tcl
}

Would instead run every request through the same code no matter which
request is sent to mod_websh from the server (configured with the
AddHandler directive...)

hth
Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Question about interpreters

2009-10-24T10:26:58Z

Hi

Ok I finally compiled everything, I reinstalled Tcl so that .tcl was put in the extension path.
Then I started experimenting with Websh again, I'm sorry if my questions seem to be kinda stupid but the last time I ignored interpreters I got stuck.

I have the following problem(s)
When I try to visualize the configuration.html document it doesn't display full. I tried to change the code in htmlhandler.ws3 to this

web::initializer {
web::config putxmarkup tag
}

restart the server but it ends up only working for only ONE time, then no more. Apache seems to constantly crash and get back up on it's own. I remember having the same problems with the older versions too.
If i change the following:

web::interpclasscfg "$webshroot/conf/htmlhandler.ws3" maxrequests 1

or to another number then my Vista keeps asking me either to close apache or to "find a solution" for the program, Apache starts automatically up so it's kinda annoying..

I don't know what the heck is wrong, or how to use the interpreters or why they're needed, where they're needed .. how to combine them cause it seems not to display my pages correctly. I read the quick reference before but it took me lots of hours to figure out some things. Could it be because I'm on vista? Even sessions seemed to work weird, and without a proper interpreter I felt lost so I stopped using Websh.

Regards,
Andrei

From forum: Websh - User

Re: Compile problems

2009-10-23T06:36:11Z

Hi

> cl /D"WIN32" /D"VERSION=\"3.6.0b5\"" /D"_MBCS" /W3 /EHsc /O2 /Ob1 /DTHREAD_SAFE=1 -I"D:/Tcl/include" -I"D:/Tcl/Apache/include" -UUSE_TCL_STUBS -c ../generic/interpool.c /Fointerpool.obj

[...]

> interpool.c
> d:\tcl\websh-3.6.0b5\src\generic\mod_websh.h(23) : fatal error C1083: Cannot open include file: 'httpd.h': No such file or directory

Obviously the compiler doesn't find httpd.h. Try to locate your
httpd.h file and then configure your Makefile to point to the *parent*
dir of the include directory.

For me the following worked w/o any tweak:

- Install Apache 2.2.14 to
"C:\Program Files\Apache Software Foundation\Apache2.2.14\"
(make sure to also install the header files)

- Install the latest ActiveTcl to
"C:\Program Files\ActiveTcl\8.6.0\"

- Unpack websh-3.6.0b5, cd to src\win and run

nmake TCL_PREFIX="C:/Program Files/ActiveTcl/8.6.0" TCL_VERSION=86 HTTPD_PREFIX="C:/Program Files/Apache Software Foundation/Apache2.2.14" clean all test apachetest

-> That's it

hth
Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Compile problems

2009-10-23T02:34:09Z

Hello

I have (successfully) compiled Websh before (with a lot of searching and work..) but now it seems that i just can't compile the newer version. I'm using Tcl 8.6b1 and apache 2.2.14. I tried using the compiled apache and also the uncompiled one but it just won't work.

I always get errors, sometimes the arp.h is missing, other times the libhttpd.lib is missing although the libhttpd.dll exists.

Last time I tried to do everything i got this error:

Microsoft (R) Program Maintenance Utility Version 9.00.21022.08
Copyright (C) Microsoft Corporation. All rights reserved.

cl /D"WIN32" /D"VERSION=\"3.6.0b5\"" /D"_MBCS" /W3 /EHsc /O2 /Ob1 /DTHRE
AD_SAFE=1 -I"D:/Tcl/include" -I"D:/Tcl/Apache/include" -UUSE_TCL_STUBS -c ../gen
eric/interpool.c /Fointerpool.obj
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.21022.08 for 80x86
Copyright (C) Microsoft Corporation. All rights reserved.

interpool.c
d:\tcl\websh-3.6.0b5\src\generic\mod_websh.h(23) : fatal error C1083: Cannot ope
n include file: 'httpd.h': No such file or directory
NMAKE : fatal error U1077: '"C:\Program Files (x86)\Microsoft Visual Studio 9.0\
VC\BIN\cl.EXE"' : return code '0x2'
Stop.

Could you please be able to tell me how you did it precisely, cause i can't quite remember how i made it work the last time. Do I have to compile Apache separated or do I need the TCL source?

Thanks in advance.

From forum: Websh - User

Re: Regarding Google Analytics issue

2009-09-29T04:55:21Z

Hi Manu

> I just got stuck in the same issue as you with the google analytics.
>
> The cookie it generates have the value that consists of '=' sign,
> which is not parsed properly, and because of the same, our secured
> subdomain site doesn't work, and as soon as I delete the cookie, it
> starts working.

Are you talking about Websh? If so, please update to the latest
version (3.6.0b5), where this issue is solved.

If that is not possible for some reason, you can overwrite the
_getCookie proc of a cookiecontext (where yourContext is changed to
the name of your context) as follows:

web::cookiecontext yourContext
proc yourContext::_getCookie {id} {
::set cookie [web::request HTTP_COOKIE]
foreach v [split $cookie ";"] {
# get key and value: note: some cookies (e.g.. Google analytics)
# also use = in the value, so a simple split with =; does
# not always work...
set key [string trim [lindex [split $v =] 0]]
set value [string trim [join [lrange [split $v =] 1 end] =]]
lappend kvlist $key $value
}
# search for $datatag
foreach {key value} $kvlist {
if {[string compare $key $id] == 0} {
return $value
}
}
error "no matching cookie found"
}

> Could you kindly tell me what did you do to solve this issue.

And if you don't talk about Websh: make sure whatever pareses your
cookies properly splits at "," characters and then just takes the
everything before the first "=" as cookie name, the rest as value,
even if it contains "0" aswell...

hth
Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Re: ANNOUNCE: Websh3.6.0b5

2009-09-27T12:36:25Z

Hi Geroge

> Can websh be used to create an HTTP REST service in tcl?

Sure. But then, I guess you can create a REST service with just about
every technology.

Websh at least provides an easy interface to handle PATH_INFO and all
parameters or form vars. That's abot what you need to create a REST
service.

> Ideally I want to have a thread pool (the one in the thread package)
> with N threads already initialised. Can websh receive a request, place a
> task in the thread pool and return the result?

Well, I don't really know the thread package. Websh already provides
an interpreter pool (when run as mod_websh) that you can use to handle
mutliple requests per interpreter.

It should not be a problem to use a thread package, as Websh is "just"
an ordinary Tcl interpreter anyway. The only thing to onsider is
making sure that you understand what your apache process model is and
what you expect Websh to do, so that you scale where you actually plan
it.

That doesn't help much, I guess, but your question was pretty
vage... I still hth ;-)

Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Re: Building mod_websh DLL on Windows

2009-09-24T04:13:35Z

Hi

> Just saw your posting of the new websh on c.l.t and tried building on
> XP. Though the build went fine, I had the following questions -
>
> - websh links against both tcl86.lib and tclstub86.lib. Since the
> compile does not seem to define USE_TCL_STUBS except for interpool.obj,
> why is tclstub86.lib needed? Also why is only a single file built
> against the stubs? Somehow this seems inconsistent. AFAIK, you should
> link against one or the other, not both but perhaps I'm wrong about this.

It could actually be that it is not too consistent, but it definitly
needs to compile against both: basically libwebsh.so (or websh.dll)
are compiled against the stub library, as it is loaded into either a
tclsh or websh or mod_websh, where a tcl library is already
loaded. mod_websh.so however must link directly to tcl.lib, as there
is no Tcl anywhere else in the application. So mod_websh.so must
properly know how to load Tcl.

> - possibly more important, TCL_THREADS is not defined for the build as
> far as I can tell. As a result the tcl.h macros get resolved as not
> requiring locking (all the mutex calls become no-ops). Is that not going
> to be a problem? I'm asking because I got bitten by this early on
> working on my TWAPI extension causing occasional crashes that I tracked
> down to this issue.

Mhh. That's true. My impression was that if we need threads, we just
need to link against a threaded Tcl. I just added /DTCL_THREADS=1 to
the Makefile, but it doesn't seem to make a difference at least
concerning the reported instability of the test suite. (That doesn't
impley that it shouldn't be tehre, though ;-)

Unfortunately, I'm really bad with Windows debugging. If you have a
chance to look into this that would be wonderful. (nmake apachetest
for me just sometimes does not run through w/o error with Tcl85 or
Tcl86 on Apache2 and Apache 2.2) It helps if I sleep between calls to
the server and give it more time to start and shutdown, but I'm not
sure if that's actually the cause of the problem or if it's just a
latent timing problem that arises more often when requesting
earlier...)

Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@...
For additional commands, e-mail: websh-dev-help@...

From forum: Websh - Dev

ANNOUNCE: Websh3.6.0b5

2009-09-24T00:42:18Z

The Websh maintenance team is proud to announce Websh 3.6.0b5.

Websh is a Tcl extension to support web application development and
provides an Apache module for Apache 1.3, 2.0, and 2.2 servers. It
compiles against Tcl 8.3 and later, including Tcl 8.6.

Get the release from

http://tcl.apache.org/websh/download/

More info about Websh can be found on the project home page

http://tcl.apache.org/websh

Websh is known to compile under
Solaris (e.g. Solaris 8 to 10)
Linux (e.g. RedHat 6 and 8, SuSe 9.1, CentOS 4.3)
Windows XP and Vista(Visual Studion 6.0 nmake)
OS-X Tiger

Websh3.6.0b5 is considered the most stable version of Websh so far,
but it is still beta, as there are some issues regarding stability on
Windows. (See CHANGES below)

For a summary of Changes since Websh 3.5.0 refer to

http://svn.apache.org/repos/asf/tcl/websh/tags/3.6.0b5/CHANGES

A detailed lst of changes can be found in

http://svn.apache.org/repos/asf/tcl/websh/tags/3.6.0b5/src/ChangeLog

Any feedback welcome. Official mailing lists for questions:

websh-user@... (user questions/discussions)
websh-dev@... (development questions/discussion)

Ronnie Brunner
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

ANNOUNCE: Websh3.6.0b5

2009-09-24T00:42:18Z

The Websh maintenance team is proud to announce Websh 3.6.0b5.

Websh is a Tcl extension to support web application development and
provides an Apache module for Apache 1.3, 2.0, and 2.2 servers. It
compiles against Tcl 8.3 and later, including Tcl 8.6.

Get the release from

http://tcl.apache.org/websh/download/

More info about Websh can be found on the project home page

http://tcl.apache.org/websh

Websh is known to compile under
Solaris (e.g. Solaris 8 to 10)
Linux (e.g. RedHat 6 and 8, SuSe 9.1, CentOS 4.3)
Windows XP and Vista(Visual Studion 6.0 nmake)
OS-X Tiger

Websh3.6.0b5 is considered the most stable version of Websh so far,
but it is still beta, as there are some issues regarding stability on
Windows. (See CHANGES below)

For a summary of Changes since Websh 3.5.0 refer to

http://svn.apache.org/repos/asf/tcl/websh/tags/3.6.0b5/CHANGES

A detailed lst of changes can be found in

http://svn.apache.org/repos/asf/tcl/websh/tags/3.6.0b5/src/ChangeLog

Any feedback welcome. Official mailing lists for questions:

websh-user@... (user questions/discussions)
websh-dev@... (development questions/discussion)

Ronnie Brunner
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@...
For additional commands, e-mail: websh-dev-help@...

From forum: Websh - Dev

Re: Reconsider command name for error logger

2009-09-03T07:19:26Z

On Sep 3, 2009, at 8:54 AM, Karl Lehenbauer wrote:

> Since the table command is "apache_table", how about if the proposed
> error logging command is "apache_log_error"?

I don't have any problem with adding more of the Apache API to Rivet,
and I like keeping the function names the same if we can. Because all
of the Apache stuff is generally "namespaced" with the apache_ prefix,
we shouldn't really have to worry about name collisions, and it'll be
really easy for anyone familiar with Apache to pick up the API.

Go for it.

D

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Reconsider command name for error logger

2009-09-03T06:54:27Z

Since the table command is "apache_table", how about if the proposed
error logging command is "apache_log_error"?

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: bugzilla reports

2009-08-09T09:13:16Z

> I found in bugzilla database a report issued by
> David back in 2005. David lamented the Tcl build system
> was broken. Since we have switched to autotools i presume
> the report can be deleted.

Sure.

> I think also the we could use bugzilla to file a report
> (or more) as a memo for the missing features and open issues
> to be addressed before releasing. (fixing the test suite
> is the first that i can recall)

Sure, that works for me. What others come to mind?

--
David N. Welton

http://www.welton.it/davidw/

http://www.dedasys.com/
Sent from Padua, Veneto, Italy

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

bugzilla reports

2009-08-03T23:39:08Z

I found in bugzilla database a report issued by
David back in 2005. David lamented the Tcl build system
was broken. Since we have switched to autotools i presume
the report can be deleted.

I think also the we could use bugzilla to file a report
(or more) as a memo for the missing features and open issues
to be addressed before releasing. (fixing the test suite
is the first that i can recall)

-- Massimo

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: OK got AX_PREFIX_CONFIG_H

2009-06-26T09:56:13Z

On Fri, 26 Jun 2009 18:22:34 +0200, Massimo Manghi wrote
>
> svn: OPTIONS of 'https://svn.apache.org/repos/asf/tcl/rivet/trunk':
> could not connect to server (https://svn.apache.org)
>
> when i try to update my working copy,
> anyone else is experiencing the same problem?
>

I'm sorry to add noise to the list, but
it turned out this problem is specific to debian
'testing' release.

cheers

-- Massimo

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: OK got AX_PREFIX_CONFIG_H

2009-06-26T09:22:34Z

On Thu, 25 Jun 2009 10:36:22 -0500, Karl Lehenbauer wrote
> Yeah I was getting all those warnings too. With the updated files
> from libtool and TEA, it did silence the complaining on my machines.
>

good.

I wonder if the AC_CONFIG_MACRO_DIR is enough for enabling
the autotools to find all they need. Using the macro should be a more
stable solution....

by the way, I keep getting messages like

svn: OPTIONS of 'https://svn.apache.org/repos/asf/tcl/rivet/trunk': could not
connect to server (https://svn.apache.org)

when i try to update my working copy,
anyone else is experiencing the same problem?

-- Massimo

> On Jun 24, 2009, at 8:21 PM, Massimo Manghi wrote:
>
> > I can't remember what exactly I did introducing this macro
> > and svn.apache.org seems to be down at the moment to read
> > the logs. I'm quite certain I added ACLOCAL_AMFLAGS = -I m4
> > to Makefile.am
> >
> > I tried to rebuild the configure script in the way
> > you did and my libtool installation (2.2.6) suggested to put
> >
> > AC_CONFIG_MACRO_DIR([m4])
> >
> > in configure.ac. It seems that a working 'configure' comes out of it.
> >
> > more recent libtool versions complain a lot with warnings like this
> >
> > configure.ac:93: warning:
> > AC_CACHE_VAL(lt_prog_compiler_pic_works, ...):
> > suspicious cache-id, must contain _cv_ to be cached
> > tclconfig/libtool.m4:568: AC_LIBTOOL_COMPILER_OPTION is expanded
> > from...
> > tclconfig/libtool.m4:4559: AC_LIBTOOL_PROG_COMPILER_PIC is expanded
> > from...
> >
> > I hope the new libtool.m4 will silence these messages, but I still
> > can't get
> > to svn.apache.org to find out
> >
> > -- Massimo
> >
> > On Wed, 24 Jun 2009 14:34:24 -0500, Karl Lehenbauer wrote
> >> I added
> >>
> >> m4_include([m4/ax_prefix_config_h.m4])
> >>
> >> to acinclude.m4, and the configure script works now.
> >>
> >> -karl
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: rivet-dev-unsubscribe@...
> >> For additional commands, e-mail: rivet-dev-help@...
> >
> >
> > --
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: rivet-dev-unsubscribe@...
> For additional commands, e-mail: rivet-dev-help@...

--

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: OK got AX_PREFIX_CONFIG_H

2009-06-25T08:36:22Z

Yeah I was getting all those warnings too. With the updated files
from libtool and TEA, it did silence the complaining on my machines.

On Jun 24, 2009, at 8:21 PM, Massimo Manghi wrote:

> I can't remember what exactly I did introducing this macro
> and svn.apache.org seems to be down at the moment to read
> the logs. I'm quite certain I added ACLOCAL_AMFLAGS = -I m4
> to Makefile.am
>
> I tried to rebuild the configure script in the way
> you did and my libtool installation (2.2.6) suggested to put
>
> AC_CONFIG_MACRO_DIR([m4])
>
> in configure.ac. It seems that a working 'configure' comes out of it.
>
> more recent libtool versions complain a lot with warnings like this
>
> configure.ac:93: warning:
> AC_CACHE_VAL(lt_prog_compiler_pic_works, ...):
> suspicious cache-id, must contain _cv_ to be cached
> tclconfig/libtool.m4:568: AC_LIBTOOL_COMPILER_OPTION is expanded
> from...
> tclconfig/libtool.m4:4559: AC_LIBTOOL_PROG_COMPILER_PIC is expanded
> from...
>
> I hope the new libtool.m4 will silence these messages, but I still
> can't get
> to svn.apache.org to find out
>
> -- Massimo
>
> On Wed, 24 Jun 2009 14:34:24 -0500, Karl Lehenbauer wrote
>> I added
>>
>> m4_include([m4/ax_prefix_config_h.m4])
>>
>> to acinclude.m4, and the configure script works now.
>>
>> -karl
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: rivet-dev-unsubscribe@...
>> For additional commands, e-mail: rivet-dev-help@...
>
>
> --
>

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: OK got AX_PREFIX_CONFIG_H

2009-06-24T18:21:01Z

I can't remember what exactly I did introducing this macro
and svn.apache.org seems to be down at the moment to read
the logs. I'm quite certain I added ACLOCAL_AMFLAGS = -I m4
to Makefile.am

I tried to rebuild the configure script in the way
you did and my libtool installation (2.2.6) suggested to put

AC_CONFIG_MACRO_DIR([m4])

in configure.ac. It seems that a working 'configure' comes out of it.

more recent libtool versions complain a lot with warnings like this

configure.ac:93: warning: AC_CACHE_VAL(lt_prog_compiler_pic_works, ...):
suspicious cache-id, must contain _cv_ to be cached
tclconfig/libtool.m4:568: AC_LIBTOOL_COMPILER_OPTION is expanded
from...
tclconfig/libtool.m4:4559: AC_LIBTOOL_PROG_COMPILER_PIC is expanded
from...

I hope the new libtool.m4 will silence these messages, but I still can't get
to svn.apache.org to find out

-- Massimo

On Wed, 24 Jun 2009 14:34:24 -0500, Karl Lehenbauer wrote

> I added
>
> m4_include([m4/ax_prefix_config_h.m4])
>
> to acinclude.m4, and the configure script works now.
>
> -karl
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: rivet-dev-unsubscribe@...
> For additional commands, e-mail: rivet-dev-help@...

--

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

OK got AX_PREFIX_CONFIG_H

2009-06-24T12:34:24Z

I added

m4_include([m4/ax_prefix_config_h.m4])

to acinclude.m4, and the configure script works now.

-karl

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

trouble producing a working configure script

2009-06-24T12:09:19Z

I am having a hellacious time getting a working configure script for
Rivet. If I run aclocal (1.10), it gives a ton of errors complaining
about stuff in tclconfig/libtool.m4.

OK, so libtool.m4 is old. I copied a newer version out of libtool.

I then ran aclocal, and "libtoolize --force --copy" to install new
config.guess, config.sub and ltmain.sh in tclconfig dir.

Ran autoheader, automake, autoconf. Cool.

Then try to run configure, it barfs...

./configure: line 2747: syntax error near unexpected token
`rivet_config.h,rivet'
./configure: line 2747: `AX_PREFIX_CONFIG_H(rivet_config.h,rivet)'

This is because AX_PREFIX_CONFIG_H isn't defined.

It's defined in m4/ax_prefix_config_h.m4 but I haven't found anywhere
where that file is getting pulled in.

I have also pulled in newer Tcl Extension Architecture (TEA) support
files into tclconfig.sh and edited configure.ac to reference the
slightly newer version of TEA (3.7 versus 3.6).

Mostly probably I just need to get that AX_PREFIX_CONFIG_H macro
pulled in somehow.

Any suggestions?

-karl

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: Rivet website

2009-06-24T11:11:24Z

> I'd like to commit the scripts and pages to Rivet's svn repository as an
> example of possible development with Rivet and other Tcl related tools. I
> would like to commit also the html pages to the website.

I'm a bit late to the game, but by all means go ahead!

--
David N. Welton

http://www.welton.it/davidw/

http://www.dedasys.com/

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

Re: websh crash - possible cause

2009-06-22T22:31:58Z

After fixing the logdest/close problem, the rest of the tests from 'make
test' ran fine. I have not gotten around to running apachetest. It
failed when I tried it because Ubuntu Apache does not have its
configuration layout in the format expected by the test suite. I need to
uninstall it and install a standard Apache distribution first and retry.
Maybe today or tomorrow.

/Ashok

Ronnie Brunner wrote:

>> It appears as though websh will crash if "logdest delete" is called on
>> log destination that is a channel after closing the channel via Tcl's
>> close command. In particular, test log-2.6 has the following sequence:
>> close $fh
>> web::logdest delete
>>
>
> Yes, web::logdest has a problem if the channel is closed. I'll be
> looking at this in the next few days. Thanks for finding that one.
>
> In the mean time: if you just put the [close $fh] statement after the
> [web::logdest delete] statement in this test: does your setup still
> crash? Do you get other errors?
>
>
>> Please verify if that might the problem. Of course, that may or may not
>> be the cause of the crashes you are seeing.
>>
>
> It is not the cause for my problems.
>
>
>> Funnily enough on XP and Tcl 8.6 the test suite runs fine. I suppose it
>> depends when that memory gets reused.
>>
>
> Yeah, probably. (I never had this test fail on my RedHat boxes either).
>
> Ronnie
>

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Re: websh crash - possible cause

2009-06-22T10:43:30Z

> It appears as though websh will crash if "logdest delete" is called on
> log destination that is a channel after closing the channel via Tcl's
> close command. In particular, test log-2.6 has the following sequence:
> close $fh
> web::logdest delete

Yes, web::logdest has a problem if the channel is closed. I'll be
looking at this in the next few days. Thanks for finding that one.

In the mean time: if you just put the [close $fh] statement after the
[web::logdest delete] statement in this test: does your setup still
crash? Do you get other errors?

> Please verify if that might the problem. Of course, that may or may not
> be the cause of the crashes you are seeing.

It is not the cause for my problems.

> Funnily enough on XP and Tcl 8.6 the test suite runs fine. I suppose it
> depends when that memory gets reused.

Yeah, probably. (I never had this test fail on my RedHat boxes either).

Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Rivet website

2009-06-22T04:17:55Z

while i was recovering from a flue i worked on the scripts i used to
regenerate the website. I originally wrote these scripts along with my
students as a case of study for developing a small, slowly changing
website and keeping it manageable.

the experimental website is (temporarily) hosted at

http://rivet.biol.unipr.it/index.rvt

or at

http://rivet.biol.unipr.it/

through the latter link the same pages are shown but their local path
turns into a static html page (mod_rewrite maps these urls to the
template :e.g. /about.html -> /index.rvt?show=about). By pointing a
mirroring tool to the static form of the site should generate a static
image of it.

I haven't done much on the content of the pages, except for creating a
web page for each example already in the docs (by clicking the
'examples' link the 'hello world' example appears and a new menu linking
to all the examples shows up).

The core of the site generation is a template and a site map, an xml
document where links are organized into groups incidentally called menus
;-). Page contents are stored in XML files along with other metadata
(Author, Last Changes timestamp etc). In every XML file a central role
is played by the "content" tag elements. These elements *must* be nodes
to which a valid xhtml fragment representing the page main content is
attached. Actually there are some extensions (e.g. to include program
code, rivet templates or other non xml-compliant html code) that are
handled and reworked internally into valid xhtml. (More on this in docs
I am writing)

For those who might be interested in testing the scripts: a feature
available is that the website can go multilingual. An example is shown here:

http://rivet.biol.unipr.it/index.rvt?show=shaded_table

the page is visible also in italian

http://rivet.biol.unipr.it/index.rvt?show=shaded_table&lang=it

At the moment the site doesn't map properly static pages of the form

(page).(language).html

into the corresponding parameters for the template, but it should be
easy and straightforward to do it.

I'd like to commit the scripts and pages to Rivet's svn repository as an
example of possible development with Rivet and other Tcl related tools.
I would like to commit also the html pages to the website.

Comments and suggestions are welcome. If a competent person in styling a
website had a better template (and associated style sheet) it would be
an excellent way to prepare the website for the new release of Rivet.

-- Massimo

---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscribe@...
For additional commands, e-mail: rivet-dev-help@...

From forum: Rivet - Dev

websh crashes - possible cause

2009-06-20T04:35:03Z

(apologies if duplicate - my previous email did not show up on the list)

Ronnie,

Regarding the "make test" crashes on Ubuntu -

It appears as though websh will crash if "logdest delete" is called on
log destination that is a channel after closing the channel via Tcl's
close command. In particular, test log-2.6 has the following sequence:
close $fh
web::logdest delete
This results eventually in a call to Tcl_Flush from
destroyLogToChannelData with a channel pointer that is no longer valid.
The end-result is likely to be random crashes as Tcl_Flush does not seem
to do any validation.

Please verify if that might the problem. Of course, that may or may not
be the cause of the crashes you are seeing.

Funnily enough on XP and Tcl 8.6 the test suite runs fine. I suppose it
depends when that memory gets reused.

Regards

/Ashok

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

websh crash - possible cause

2009-06-20T01:27:46Z

It appears as though websh will crash if "logdest delete" is called on
log destination that is a channel after closing the channel via Tcl's
close command. In particular, test log-2.6 has the following sequence:
close $fh
web::logdest delete
This results eventually in a call to Tcl_Flush from
destroyLogToChannelData with a channel pointer that is no longer valid.
The end-result is likely to be random crashes as Tcl_Flush does not seem
to do any validation.

Please verify if that might the problem. Of course, that may or may not
be the cause of the crashes you are seeing.

Funnily enough on XP and Tcl 8.6 the test suite runs fine. I suppose it
depends when that memory gets reused.

Regards

/Ashok

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

Re: mod_websh DLL on Windows and other questions

2009-06-17T06:40:10Z

> On another note, I would like to make precompiled binaries, for Windows
> and possibly Linux, available from my Woof (woof.sourceforge.net) site.
> Would you have any objections to that? Primary purpose is to allow users
> to install easily without needing a compiler and tools, at least on
> Windows. FYI, Woof is a Rails-like framework that is server-independent.
> For Apache the preferred interface is mod_websh for performance reasons
> although it supports CGI and SCGI modes as well. Let me know if you're
> ok with my redistributing binaries.

I'm totally OK with that. (I actually plan to provide some on the
offical web site as well, if I ever get to release 3.6.0, but that
certainly shouldn't prevent you from doing so...

Small note: I usually test Websh on Window, Linux and Solaris for
Tcl8.4, 8.5 and 8.6 with Apache 1.3, 2.0 and 2.2

- Websh with Tcl8.6 does not work yet on Linux and Solaris: the apache
module crashes (standalone test suite works)
- Apache 1.3.41 does not perfectly work on Windows:
[web::config server_root] does not work on my machines
- Apache 2.2.11 sometimes crashes on Windows (nmake apachetest
regularly crashes with Tcl8.5 and Tcl8.6)

Maybe some tests on your environment might either help track the bugs
or at least confirm the problems... So I wouldn't mind if you could run
make test as well as make apachetest on your system(s) and let me know
the outcome.

Regards
Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@...
For additional commands, e-mail: websh-dev-help@...

From forum: Websh - Dev

Re: mod_websh DLL on Windows and other questions

2009-06-17T00:40:01Z

> In 3.6.0b4 (released version), copying the .so file to the modules
> directory was sufficient on Windows. In the current CVS head, the .dll
> also needs to be copied. The README specifically says the DLL is not
> needed to run the Apache module and should be fixed. Only a nit, but
> someone might spend a half-hour trying to figure out why Apache will not
> load the module (like I did).

Sorry for that one. I just committed a fix in the REAMDE.

> Also, another small request - would it be possible to change the build
> or patch number in CVS to distinguish from the released 3.60.b4 ?

Yeah, I guess it's about time to release at least 3.6.0b5. Unfortunately,
I was reproved by Apache for releasing anything without following the
proper ASF rules. It's about time I find out how to do it properly, I
guess ;-) (Just using some "-SNAPSHOT" version does not really solve
the problem of knowing what code is in use, it just makes sure that
you don't mix up the released version with any other build... I'll try
to think of something.)

> Finally two questions, on both Windows and *ix, does mod_websh need to
> be built differently based on the Apache process model (MPM etc.) ? How
> about whether the Tcl binaries are built threaded or not ?

No, mod_websh does not care about MPMs or threads in Tcl (i.e. you
don't have to change anything when building). Websh itself is thread
safe. But it certainly makes difference when it's run in different
environments:
- If Apache runs in any threaded MPM, Tcl Must be threaded as well.
- CGI case is independent of the model (for obvious reasonsn ;-)
- If your Websh app uses threads: Tcl must be threaded too, regardless
of the MPM Apache runs
- Interpreter handling is different when running in threaded
vs. non-threaded MPMs: For every (Apache) process there is an
interpreter pool. While requests can share interpreters within a
process, they cannot do so across processes. (For example this needs
to be considered when you try to create sessions kept in memory of
the Apache process, which only works properly if you run Apache in a
single process mode.)

Hope that helps
Ronnie
--
Ronnie Brunner | ronnie.brunner@...
phone +41-44-247 79 79 | fax +41-44-247 70 75
Netcetera AG | 8040 Zürich | Switzerland | http://netcetera.ch

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User

mod_websh DLL on Windows and other questions

2009-06-16T20:01:42Z

In 3.6.0b4 (released version), copying the .so file to the modules
directory was sufficient on Windows. In the current CVS head, the .dll
also needs to be copied. The README specifically says the DLL is not
needed to run the Apache module and should be fixed. Only a nit, but
someone might spend a half-hour trying to figure out why Apache will not
load the module (like I did).

Also, another small request - would it be possible to change the build
or patch number in CVS to distinguish from the released 3.60.b4 ?

Finally two questions, on both Windows and *ix, does mod_websh need to
be built differently based on the Apache process model (MPM etc.) ? How
about whether the Tcl binaries are built threaded or not ?

Thanks

/Ashok

---------------------------------------------------------------------
To unsubscribe, e-mail: websh-user-unsubscribe@...
For additional commands, e-mail: websh-user-help@...

From forum: Websh - User