tclhttpd-users — Discussion and general Q&A for TclHttpd

Re: [Tclhttpd-users] tmp folder and Config gui

[Brent W. <we...@pa...>]

>>>"Luciano ES" said:

 > - I diff'ed the entire structure and noticed several changes. Shouldn't
 > it be assigned a new version number?

I usually wait a bit before changing the version number in CVS.
If I have time I'll do that.

 > In this new version, I still have to hack srvui.tcl with a "wm withdraw 
 > ." statement. I know there is a "Config gui" option in tclhttpd.rc, 
 > but it doesn"t withdraw the toplevel.
 > 
 > - What is the point of having that blank wish window hanging about if I
 > decide not to use a GUI?

I think I finally understand that issue - basically you have wish
but do not want the GUI.  This happens on Windows where wish is the
default way to run .tcl scripts.  That should be easy.
Something like this in bin/httpd.tcl ?

# Start up the user interface and event loop.

if {[info exists tk_version]} { 
  if {$Config(gui)} {
    package require httpd::srvui
    SrvUI_Init "Tcl HTTPD $Httpd(version)"
  } else {
    wm withdraw .
  }
}

or, perhaps always create the interface, but withdraw it:

if {[info exists tk_version]} { 
  if {$Config(gui)} {
  package require httpd::srvui
  SrvUI_Init "Tcl HTTPD $Httpd(version)"
  if {!$Config(gui)} {
    wm withdraw .
  }
}


--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
we...@pa...

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Erik L. <e.l...@hc...>]

Colin McCormack wrote:
> 
> md5 v2 offers a C implementation of md5, which is not something to be
> sniffed at, given that digest authentication requires at least 3 md5
> calculations per httpd interaction, and they're not cheap.
> 

Seems a rather convincing argument to me.

Erik Leunissen
==============

Re: [Tclhttpd-users] tclhttpd 3.5.1

[David G. <dav...@po...>]

Matthias Hoffmann <M.H...@hm...> wrote:

>- a cosmetic thing: tclkit-win32-sh tclhttp.kit -? shows the help, but also this error:
>
>     while executing
>"error [usage $optlist $usage]"
>     (procedure "cmdline::getoptions" line 15)
>     invoked from within
>"cmdline::getoptions argv $CommandLineOptions \
>     "usage: httpd.tcl options:""
>     invoked from within
>"array set Config [cmdline::getoptions argv $CommandLineOptions \
>     "usage: httpd.tcl options:"]"
>     (file "D:/PGM/WebSrv4/bin/tclhttpd.kit/bin/httpd.tcl" line 195)
>     invoked from within
>"source [file join $starkit::topdir bin/httpd.tcl]"
>     (file "tclhttpd.kit/main.tcl" line 3)
>     invoked from within
>"source tclhttpd.kit/main.tcl"
>     ("uplevel" body line 1)
>     invoked from within
>"uplevel [list source [file join $self main.tcl]]"
>

Yes, that can be fix with the following patch:

Index: httpd.tcl
===================================================================
RCS file: /cvsroot/tclhttpd/tclhttpd/bin/httpd.tcl,v
retrieving revision 1.55
diff -c -r1.55 httpd.tcl
*** httpd.tcl	19 May 2004 05:44:28 -0000	1.55
--- httpd.tcl	17 Jun 2004 08:16:59 -0000
***************
*** 192,199 ****
          [list gui.arg           [cget gui]      {flag for launching the
user interface}] \
          [list mail.arg           [cget MailServer]      {Mail Servers
for sending email from tclhttpd}] \
      ]
! array set Config [cmdline::getoptions argv $CommandLineOptions \
!     "usage: httpd.tcl options:"]
  
  if {[string length $Config(library)]} {
      lappend auto_path $Config(library)
--- 192,205 ----
          [list gui.arg           [cget gui]      {flag for launching the
user interface}] \
          [list mail.arg           [cget MailServer]      {Mail Servers
for sending email from tclhttpd}] \
      ]
! 
! if {[catch {
!     array set Config [cmdline::getoptions argv $CommandLineOptions \
!     "usage: httpd.tcl options:"]} err]
! } {
!     puts stderr $err
!     exit 1
! }
  
  if {[string length $Config(library)]} {
      lappend auto_path $Config(library)

-- 
David Gravereaux <dav...@po...>
[species: human; planet: earth,milkyway(western spiral arm),alpha sector]

[Tclhttpd-users] tclhttpd 3.5.1

[Matthias H. <M.H...@hm...>]

Hi all,

Just started experimenting with the new version of tclhttpd. Up until now I'm running (an unpacked) tclhttpd 3.4.2 with 
a multithreaded tclsh 8.4.4 and some additional modules unter w2k. I want to go one step further in using the 
tclkit-win32-sh-interpreter with tclhttpd.kit instead, but I don't want to change too much in the existing httproot and 
custom directories. In general, things are working well with this new environment, but a few questions arise:

- I can't find the package htmlutils anywhere it in the starkit - is it gone?
- after switching to my old -docRoot, 'package require mypage' fails (mypage.tcl is located in the httproot under 
/libtml); ok - I could move it to the custom dir, but then some dependency problems happen as a result of the order of 
SOURCes.
- What about encodings? tclkit-win32-sh only has a few encodings build in. If i store them in some auto_path lib, how 
could I use them?
- a cosmetic thing: tclkit-win32-sh tclhttp.kit -? shows the help, but also this error:

     while executing
"error [usage $optlist $usage]"
     (procedure "cmdline::getoptions" line 15)
     invoked from within
"cmdline::getoptions argv $CommandLineOptions \
     "usage: httpd.tcl options:""
     invoked from within
"array set Config [cmdline::getoptions argv $CommandLineOptions \
     "usage: httpd.tcl options:"]"
     (file "D:/PGM/WebSrv4/bin/tclhttpd.kit/bin/httpd.tcl" line 195)
     invoked from within
"source [file join $starkit::topdir bin/httpd.tcl]"
     (file "tclhttpd.kit/main.tcl" line 3)
     invoked from within
"source tclhttpd.kit/main.tcl"
     ("uplevel" body line 1)
     invoked from within
"uplevel [list source [file join $self main.tcl]]"

;-)

-- 
Hamburg Münchener Krankenkasse, Hauptverwaltung
Matthias Hoffmann
Systemtechnik HV 202.1
Schäferkampsallee 16
D-20357 Hamburg
Tel. +49 40 41535-232
Fax  +49 40 41535-359
Mail M.H...@hm...

Re: [Tclhttpd-users] tmp folder and Config gui

[Colin M. <co...@ch...>]

On Tue, 2004-06-15 at 04:35, Luciano ES wrote:

> I took so long to report the bug that it's been fixed already. I just got a new version from SF and there is no error, even without my tweaking. But I understand it because that very same line is still there.

Did you get it by CVS?  That'd explain the differences.

> Questions:
> 
> - How did the error go away if that problem line is still there?

It's no longer being read/used, so it's not causing the error.  Is it
still there?

> - Also, is it possible to tell tclhttpd where the temp file should be? I don't like to use C:\tmp.

As far as I can see, the only other /tmp references are in cgi demos and
in logging, where they are all configurable.

> - I diff'ed the entire structure and noticed several changes. Shouldn't it be assigned a new version number?

Not until release.  CVS just gets you the latest versions of everything.

> In this new version, I still have to hack srvui.tcl with a "wm withdraw ." statement.
> I know there is a "Config gui" option in tclhttpd.rc, but it doesn"t withdraw the toplevel.
> - What is the point of having that blank wish window hanging about if I decide not to use a GUI?

It's assumed that if you're not running a gui, you won't be running
wish, but rather tclsh which doesn't have any window interaction.

-- 
Colin McCormack <co...@ch...>

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Andreas K. <aku...@sh...>]

> Ugh - what a pain.  I didn't notice the use of md5hex in session.tcl.
> We can do:
> 
> (a) make session.tcl "package require md5 1" and have it simply call
> md5::md5 instead of md5hex

I consider that a bad idea. While we have fixed the problems with
smtp/mime other packages may use a different md5 and so we get another
conflict. For now I consider it better to be able to have no version
preferences at all when doing the require, and to work with both v1
and v2. This gives us maximum compatibility with any other package
using md5.

 
> (b) make session.tcl "package require httpd::digest" and call md5hex
> because it is currently (in CVS) defined in that module
> 
> (c) create a new file, lib/md5.tcl, that contains the md5hex shim
> and does "package provide httpd::md5" and then package require that
> module from digest.tcl and session.tcl
> 
> Arguably plan (c) has some merit of flexibility, but I'm annoyed enough
> with the incompatibilities between md5 1 and 2 that I'm leaning
> towards plan (a).  We don't really need whatever md5 2 is offering,
> so I feel like shunning it :-/

-- 
So long,
	Andreas Kupries <aku...@sh...>
			<http://www.purl.org/NET/akupries/>
	Developer @	<http://www.activestate.com/>
-------------------------------------------------------------------------------

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Gerald W. L. <Ger...@co...>]

Colin McCormack wrote:

>On Thu, 2004-06-17 at 09:10, Brent Welch wrote:
>  
>
> ...
>
>(d) explicitly depend on the latest version of tcllib 1.6.1, which
>obviates the need for a shim at all, and gives us md5 v2 at the cost of
>requiring people to download another 1-1.5Mb of code and run another
>make install (or use the kit.)  Is this a show-stopper impost on users?
>  
>
I believe this is a reasonable requirement.

-- 
+--------------------------------+---------------------------------------+
| Gerald W. Lester               | "The man who fights for his ideals is |
| Ger...@co...          |  the man who is alive." -- Cervantes  |
+--------------------------------+---------------------------------------+

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Colin M. <co...@ch...>]

On Thu, 2004-06-17 at 09:10, Brent Welch wrote:
> Ugh - what a pain.  I didn't notice the use of md5hex in session.tcl.

Yeah, it is a pain, but the new md5 interface is better - the old one
just assumed you wanted hex, and the new one gives you the option of
raw, which is an improvement.  Perhaps it ought to have defaulted to
hex, not raw, and made the option default to the old behavior for
minimal breakage.

> (a) make session.tcl "package require md5 1" and have it simply call
> md5::md5 instead of md5hex
> 
> (b) make session.tcl "package require httpd::digest" and call md5hex
> because it is currently (in CVS) defined in that module
> 
> (c) create a new file, lib/md5.tcl, that contains the md5hex shim
> and does "package provide httpd::md5" and then package require that
> module from digest.tcl and session.tcl

(d) explicitly depend on the latest version of tcllib 1.6.1, which
obviates the need for a shim at all, and gives us md5 v2 at the cost of
requiring people to download another 1-1.5Mb of code and run another
make install (or use the kit.)  Is this a show-stopper impost on users?

> Arguably plan (c) has some merit of flexibility, but I'm annoyed enough
> with the incompatibilities between md5 1 and 2 that I'm leaning
> towards plan (a).  We don't really need whatever md5 2 is offering,
> so I feel like shunning it :-/

md5 v2 offers a C implementation of md5, which is not something to be
sniffed at, given that digest authentication requires at least 3 md5
calculations per httpd interaction, and they're not cheap.

One thing I don't like about the [package require md5 1] approach is
that it moves the pain of installing a new tcllib (once, or once a month
:) to the pain of paying for a pure tcl md5 three times every GET, if
you prefer Digest to Basic authentication because you want to avoid
sending passwords over the net in plaintext.

-- 
Colin McCormack <co...@ch...>

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Don P. <dgp...@ve...>]

>
> As mentioned, I can certainly run a virtual host with tclhttpd HEAD and
> tcllib1.6.1, but it's not a minimal virtual host, as Don requires.

Just to clarify - I don't require anything.  I worked around the issue 
just
by hacking in a [catch].  Only mentioned the lack of need for md5 in
slave interps doing simple tasks because of old fashioned concerns
about efficiency.  I'm such a 20th century thinker.

I'm all for cleaner and/or more efficient solutions, but there's no
urgent need of mine left to satisfy.

Thanks.

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Brent W. <we...@pa...>]

Ugh - what a pain.  I didn't notice the use of md5hex in session.tcl.
We can do:

(a) make session.tcl "package require md5 1" and have it simply call
md5::md5 instead of md5hex

(b) make session.tcl "package require httpd::digest" and call md5hex
because it is currently (in CVS) defined in that module

(c) create a new file, lib/md5.tcl, that contains the md5hex shim
and does "package provide httpd::md5" and then package require that
module from digest.tcl and session.tcl

Arguably plan (c) has some merit of flexibility, but I'm annoyed enough
with the incompatibilities between md5 1 and 2 that I'm leaning
towards plan (a).  We don't really need whatever md5 2 is offering,
so I feel like shunning it :-/

>>>Colin McCormack said:
 > On Thu, 2004-06-17 at 03:26, Donald G Porter wrote:
 > > Brent Welch wrote:
 > > > Don - I moved the "package require md5 1" into digest.tcl -
 > > > I did not put it into bin/httpdthread.tcl
 > > 
 > > Ok, so it will be necessary for anything that plans to
 > > use digest.tcl to be sure to [source] it first so that
 > > [package require]'s greedy version selection algorithm
 > > does not doom it to fail against an md5 version 2
 > > loaded by some other module.
 > > 
 > > Alternatively, digest.tcl could be patched up to be
 > > able to work with either md5 release.
 > 
 > MD5 is used by session.tcl and digest.tcl, and indirectly by mail.tcl
 > via smtp.
 > 
 > The code in utils.tcl is there to provide an md5hex wrapper which
 > selects the appropriate invocation to the md5::md5 call depending on
 > which version happens to be available.  I think it mimics the tcllib
 > code, in that respect.  Don is saying, not unreasonably, that he doesn't
 > need the overhead if he's not using digest or session.
 > 
 > The cleanest solution would be for tclhttpd to require a sufficiently
 > new tcllib to avoid the problem altogether, then remove the md5hex
 > wrapper, and put a version check on tcllib itself somewhere at startup.
 > 
 > Next cleanest would be to duplicate the utils.tcl code into digest.tcl
 > and session.tcl (again, only called at startup.)
 > 
 > As mentioned, I can certainly run a virtual host with tclhttpd HEAD and
 > tcllib1.6.1, but it's not a minimal virtual host, as Don requires.
 > 
 > Isn't it time we added explicit package requires into each module?  Then
 > Don could source only what he wants, and expect to get a minimal system
 > - the problem now is one of cleanly and dynamically constructing a
 > minimal working set of .tcl files for a given set of desired functions.
 > -- 
 > Colin McCormack <co...@ch...>
 > 

--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
we...@pa...

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Colin M. <co...@ch...>]

On Thu, 2004-06-17 at 03:26, Donald G Porter wrote:
> Brent Welch wrote:
> > Don - I moved the "package require md5 1" into digest.tcl -
> > I did not put it into bin/httpdthread.tcl
> 
> Ok, so it will be necessary for anything that plans to
> use digest.tcl to be sure to [source] it first so that
> [package require]'s greedy version selection algorithm
> does not doom it to fail against an md5 version 2
> loaded by some other module.
> 
> Alternatively, digest.tcl could be patched up to be
> able to work with either md5 release.

MD5 is used by session.tcl and digest.tcl, and indirectly by mail.tcl
via smtp.

The code in utils.tcl is there to provide an md5hex wrapper which
selects the appropriate invocation to the md5::md5 call depending on
which version happens to be available.  I think it mimics the tcllib
code, in that respect.  Don is saying, not unreasonably, that he doesn't
need the overhead if he's not using digest or session.

The cleanest solution would be for tclhttpd to require a sufficiently
new tcllib to avoid the problem altogether, then remove the md5hex
wrapper, and put a version check on tcllib itself somewhere at startup.

Next cleanest would be to duplicate the utils.tcl code into digest.tcl
and session.tcl (again, only called at startup.)

As mentioned, I can certainly run a virtual host with tclhttpd HEAD and
tcllib1.6.1, but it's not a minimal virtual host, as Don requires.

Isn't it time we added explicit package requires into each module?  Then
Don could source only what he wants, and expect to get a minimal system
- the problem now is one of cleanly and dynamically constructing a
minimal working set of .tcl files for a given set of desired functions.
-- 
Colin McCormack <co...@ch...>

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Donald G P. <dgp...@ve...>]

Brent Welch wrote:
> Don - I moved the "package require md5 1" into digest.tcl -
> I did not put it into bin/httpdthread.tcl

Ok, so it will be necessary for anything that plans to
use digest.tcl to be sure to [source] it first so that
[package require]'s greedy version selection algorithm
does not doom it to fail against an md5 version 2
loaded by some other module.

Alternatively, digest.tcl could be patched up to be
able to work with either md5 release.

-- 
| Don Porter                                     dgp...@ve...  |
|           "Some days you just can't get rid of a bomb!"              |
|                       -- Adam West as BATMAN                         |
|______________________________________________________________________|

RE: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Andreas K. <and...@Ac...>]

> Glad to hear Tcllib is fixed up.
> Did you test tcllib 1.6.1 with tclhttpd3.5.1 ?

I worked with Colin (coldstore) on this and he tested 1.6.1 before I placed
the distribution archives on SourceForge.

> It was indeed some combination of md5, mime, smtp, and
> the digest module.  I don't know of other dependencies.

Thought so.

--
	Andreas Kupries <and...@Ac...>
	Developer @ http://www.ActiveState.com, a division of Sophos
	Tel: +1 604 484 6491

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Brent W. <we...@pa...>]

Glad to hear Tcllib is fixed up.
Did you test tcllib 1.6.1 with tclhttpd3.5.1 ?
It was indeed some combination of md5, mime, smtp, and
the digest module.  I don't know of other dependencies.

Don - I moved the "package require md5 1" into digest.tcl -
I did not put it into bin/httpdthread.tcl

>>>"Andreas Kupries" said:
 > 
 > > I don't think we need that block of code in utils.tcl
 > > I have bin/httpd.tcl do an explict
 > > package require md5 1
 > > to avoid the conflict with md5 2.0
 > > (I hate that it is not compatible, and that tcllib has
 > > inconsistent uses of the two packages - I couldn't figure
 > > out how the conflict crops up, by the way.  There is some
 > > sequence of package loading that grabs md5 2, where later
 > > packages do "package require md5 1" and raise errors.)
 > 
 > Brent, note that the use of md5 has been fixed in the mime and smtp packages
 > of Tcllib, in both CVS head, and in the recently released Tcllib 1.6.1
 > bugfix release. The latter was expressly done to fix this problem with md5,
 > mime, smtp, and tclhttpd. Both mime and smtp now simply package require md5,
 > without requiring a specific version, and work with both versions, switching
 > their behaviour at runtime to handle the changes in the API.
 > 
 > If there are other packages of tcllib involved in what you have seen then
 > please file a bug report at tcllib so that we can address this.
 > 
 > 
 > > Hmm - for virtual hosts and any threaded system we need
 > > to get that same
 > > package require md5 1
 > > into each thread, so it should be in bin/httpdthread.tcl
 > > I'm going to add that, and remove the code from utils.tcl
 > >
 > > Don - can you give me a short test plan/scripts for virtual
 > > hosts so we can ensure it gets tested before releases :-/ ?
 > > Something that works with the sample document tree, etc. etc.
 > 
 > 
 > --
 > 	Andreas Kupries <and...@Ac...>
 > 	Developer @ http://www.ActiveState.com, a division of Sophos
 > 	Tel: +1 604 484 6491
 > 

--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
we...@pa...

RE: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Andreas K. <and...@Ac...>]

> I don't think we need that block of code in utils.tcl
> I have bin/httpd.tcl do an explict
> package require md5 1
> to avoid the conflict with md5 2.0
> (I hate that it is not compatible, and that tcllib has
> inconsistent uses of the two packages - I couldn't figure
> out how the conflict crops up, by the way.  There is some
> sequence of package loading that grabs md5 2, where later
> packages do "package require md5 1" and raise errors.)

Brent, note that the use of md5 has been fixed in the mime and smtp packages
of Tcllib, in both CVS head, and in the recently released Tcllib 1.6.1
bugfix release. The latter was expressly done to fix this problem with md5,
mime, smtp, and tclhttpd. Both mime and smtp now simply package require md5,
without requiring a specific version, and work with both versions, switching
their behaviour at runtime to handle the changes in the API.

If there are other packages of tcllib involved in what you have seen then
please file a bug report at tcllib so that we can address this.


> Hmm - for virtual hosts and any threaded system we need
> to get that same
> package require md5 1
> into each thread, so it should be in bin/httpdthread.tcl
> I'm going to add that, and remove the code from utils.tcl
>
> Don - can you give me a short test plan/scripts for virtual
> hosts so we can ensure it gets tested before releases :-/ ?
> Something that works with the sample document tree, etc. etc.


--
	Andreas Kupries <and...@Ac...>
	Developer @ http://www.ActiveState.com, a division of Sophos
	Tel: +1 604 484 6491

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Donald G P. <dgp...@ve...>]

Brent Welch wrote:
 > I don't think we need that block of code in utils.tcl
 > I have bin/httpd.tcl do an explict
 > package require md5 1
 > to avoid the conflict with md5 2.0
...
 > Hmm - for virtual hosts and any threaded system we need
 > to get that same
 > package require md5 1
 > into each thread, so it should be in bin/httpdthread.tcl
 > I'm going to add that, and remove the code from utils.tcl

That's fine, and may help someone, but won't really address
the virtual hosts problem, I think.  The slave interp that
runs a virtual host might well never [source] either
of those files.  The virtual host may well have
its own [Config main] script.

One solution is just a requirement that
every [Config main] script ought to include
[package require md5 1].  Perhaps that requirement
is documented; I did not check.

The particular virtual host I was working on was one
that just does redirection [*].  The only tclhttpd
commands it needs are [Url_PrefixInstall] and
[Httpd_Redirect] (and whatever they depend on, of course).

I don't know where the md5 dependence comes from.  I'd sure
like to avoid it.  I think it would be better for
[package require] dependencies to be determined
locally in each file, so that each interpreter at least
has the chance to load only the packages it will actually
be using.  That may not be compatibile with [package require]'s
greedy algorithm for version selection though. :-(

 > Don - can you give me a short test plan/scripts for virtual
 > hosts so we can ensure it gets tested before releases :-/ ?
 > Something that works with the sample document tree, etc. etc.

Hmmm... maybe an entry that recognizes "localhost" as a virtual host?

Config virtual [list localhost [info script]]

That would not have caught this problem, but would at least
exercise the [Httpd_VirtualHosts] command.

[*] In particular, one that redirects all requests
sent to www.scriptics.com over to www.tcl.tk.

-- 
| Don Porter                                     dgp...@ve...  |
|           "Some days you just can't get rid of a bomb!"              |
|                       -- Adam West as BATMAN                         |
|______________________________________________________________________|

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Brent W. <we...@pa...>]

I don't think we need that block of code in utils.tcl
I have bin/httpd.tcl do an explict
package require md5 1
to avoid the conflict with md5 2.0
(I hate that it is not compatible, and that tcllib has
inconsistent uses of the two packages - I couldn't figure
out how the conflict crops up, by the way.  There is some
sequence of package loading that grabs md5 2, where later
packages do "package require md5 1" and raise errors.)

Hmm - for virtual hosts and any threaded system we need
to get that same
package require md5 1
into each thread, so it should be in bin/httpdthread.tcl
I'm going to add that, and remove the code from utils.tcl

Don - can you give me a short test plan/scripts for virtual
hosts so we can ensure it gets tested before releases :-/ ?
Something that works with the sample document tree, etc. etc.

>>>Donald G Porter said:
 > 
 > Is anyone using the virtual hosts feature with the latest tclhttpd
 > release?
 > 
 > It appears that the code near the end of /lib/utils.tcl makes a
 > call to [package present md5] and when the md5 package has never
 > been loaded (as will be the case in the slave interps that drive
 > virtual hosts) that throws a fatal error.
 > 
 > Am I missing something?
 > 
 > -- 
 > | Don Porter                                     dgp...@ve...  |
 > |           "Some days you just can't get rid of a bomb!"              |
 > |                       -- Adam West as BATMAN                         |
 > |______________________________________________________________________|
 > 
 > 
 > -------------------------------------------------------
 > This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
 > Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
 > Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
 > REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
 > _______________________________________________
 > TclHttpd-users mailing list
 > Tcl...@li...
 > https://lists.sourceforge.net/lists/listinfo/tclhttpd-users
 > 

--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
we...@pa...

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Colin M. <co...@ch...>]

On Wed, 2004-06-16 at 07:34, Donald G Porter wrote:
> Is anyone using the virtual hosts feature with the latest tclhttpd
> release?

Haven't tried it on the external facing machine - I guess I should ASAP.

> It appears that the code near the end of /lib/utils.tcl makes a
> call to [package present md5] and when the md5 package has never
> been loaded (as will be the case in the slave interps that drive
> virtual hosts) that throws a fatal error.

> Am I missing something?

Thanks for picking this up.

The files digest.tcl (which should be required by auth.tcl) and
session.tcl package require md5.  One solution (probably advisable) is
to have auth.tcl require digest and make utils.tcl require md5 (probably
reasonable, so I have checked in those mods.)  I've confirmed that this
works under my virtual hosts.

The code in utils is a hack to get around version-inconsistent md5
interfaces in tcllib.  Hopefully it can eventually be removed (say, when
we require the currently latest version of tcllib.)

I can't guarantee that tclhttpd will run with a very old tcllib.  It
might be worth grabbing the latest bugfixed version:
http://prdownloads.sourceforge.net/tcllib/tcllib-1.6.1.tar.bz2?download

The definitive solution would be to have each .tcl package explicitly
require everything it actually needs.
-- 
Colin McCormack <co...@ch...>

Re: [Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[<ede...@ie...>]

Yes I am.

I had the same problem and fixed it this way (not the smartest way):

-if {[package vcompare [package present md5] 2.0] > -1} {
+catch {package require md5} md5v
+if {[package vcompare $md5v 2.0] > -1} {

Solved my problem.

Edésio

On Tue, Jun 15, 2004 at 05:34:56PM -0400, Donald G Porter wrote:
> 
> Is anyone using the virtual hosts feature with the latest tclhttpd
> release?
> 
> It appears that the code near the end of /lib/utils.tcl makes a
> call to [package present md5] and when the md5 package has never
> been loaded (as will be the case in the slave interps that drive
> virtual hosts) that throws a fatal error.
> 
> Am I missing something?
> 
> -- 
> | Don Porter                                     dgp...@ve...  |
> |           "Some days you just can't get rid of a bomb!"              |
> |                       -- Adam West as BATMAN                         |
> |______________________________________________________________________|
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
> Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
> Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
> REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
> _______________________________________________
> TclHttpd-users mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tclhttpd-users

-- 
Grief can take care of itself, but to get the full value of a joy you must
have somebody to divide it with. -- Mark Twain

[Tclhttpd-users] Virtual Hosts broken in 3.5.1 release?

[Donald G P. <dgp...@ve...>]

Is anyone using the virtual hosts feature with the latest tclhttpd
release?

It appears that the code near the end of /lib/utils.tcl makes a
call to [package present md5] and when the md5 package has never
been loaded (as will be the case in the slave interps that drive
virtual hosts) that throws a fatal error.

Am I missing something?

-- 
| Don Porter                                     dgp...@ve...  |
|           "Some days you just can't get rid of a bomb!"              |
|                       -- Adam West as BATMAN                         |
|______________________________________________________________________|

Re: [Tclhttpd-users] templates aren't processed when docroot contains spaces

[Colin M. <co...@ch...>]

On Tue, 2004-06-15 at 16:36, Brent Welch wrote:
> within that tree named "Copy of templates" and I can access "form.tml"
> in there by any of "form.html", "form.htm", or "form.tml".  More precisely
> http://localhost:8014/Copy+of+templates/form.htm
> works OK for me.

I just tried it on HEAD, under linux.  A dir "Copy of templates" works. 
Interestingly, a symlink of the same name fails - thinks it's a
text/plain file ... I'll fix that.

> >>>"Jeff Rankin" said:
>  > Hi:
>  > 
>  > Here's an odd problem I'm seeing on Windows 2000 - when templates
>  > (*.tml) are contained in a directory name containing spaces, they're 
>  > not
>  > processed properly, they're just served up as plain text. So, you end 
>  > up
>  > seeing the unprocessed contents of the template.
>  > 
>  > There seems to be one way around this - if you call up the template
>  > without the .tml extension, it's served properly (this doesn't seem to
>  > work every time though). I'd prefer not to have to change my templates
>  > though, which typically reference the entire template name.
>  > 
>  > I'm seeing this issue in 3.5 and 3.5.1. Any ideas on how to address
>  > this? TIA -
>  > 
>  > W. Jeffrey Rankin
>  > Lead Web Application Developer
>  >
>  > O'NEIL & ASSOCIATES, INC. http://www.oneil.com
>  > 495 Byers Rd.
>  > Miamisburg, Ohio 45342-3662
>  > Phone: (937) 865-0846 ext. 3504
>  > Fax: (937) 865-5858
>  > E-mail: jr...@on...
> 
> --
> Brent Welch
> Software Architect, Panasas Inc
> Delivering the premier storage system for scalable Linux clusters
> 
> www.panasas.com
> we...@pa...
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
> Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
> Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
> REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
> _______________________________________________
> TclHttpd-users mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tclhttpd-users
-- 
Colin McCormack <co...@ch...>

Re: [Tclhttpd-users] templates aren't processed when docroot contains spaces

[Brent W. <we...@pa...>]

Hmm - I'm having trouble duplicating this problem.
I'm on my Windows 2000 laptop right now with tclhttpd3.5.1.
I have
   -docRoot "C:\A Folder\htdocs"
and that doesn't cause any problems.  I also created a directory
within that tree named "Copy of templates" and I can access "form.tml"
in there by any of "form.html", "form.htm", or "form.tml".  More precisely
http://localhost:8014/Copy+of+templates/form.htm
works OK for me.

>>>"Jeff Rankin" said:
 > Hi:
 > 
 > Here's an odd problem I'm seeing on Windows 2000 - when templates
 > (*.tml) are contained in a directory name containing spaces, they're 
 > not
 > processed properly, they're just served up as plain text. So, you end 
 > up
 > seeing the unprocessed contents of the template.
 > 
 > There seems to be one way around this - if you call up the template
 > without the .tml extension, it's served properly (this doesn't seem to
 > work every time though). I'd prefer not to have to change my templates
 > though, which typically reference the entire template name.
 > 
 > I'm seeing this issue in 3.5 and 3.5.1. Any ideas on how to address
 > this? TIA -
 > 
 > W. Jeffrey Rankin
 > Lead Web Application Developer
 >
 > O'NEIL & ASSOCIATES, INC. http://www.oneil.com
 > 495 Byers Rd.
 > Miamisburg, Ohio 45342-3662
 > Phone: (937) 865-0846 ext. 3504
 > Fax: (937) 865-5858
 > E-mail: jr...@on...

--
Brent Welch
Software Architect, Panasas Inc
Delivering the premier storage system for scalable Linux clusters

www.panasas.com
we...@pa...

[Tclhttpd-users] templates aren't processed when docroot contains spaces

[Jeff R. <Jr...@on...>]

Hi:

Here's an odd problem I'm seeing on Windows 2000 - when templates
(*.tml) are contained in a directory name containing spaces, they're not
processed properly, they're just served up as plain text. So, you end up
seeing the unprocessed contents of the template.

There seems to be one way around this - if you call up the template
without the .tml extension, it's served properly (this doesn't seem to
work every time though). I'd prefer not to have to change my templates
though, which typically reference the entire template name.

I'm seeing this issue in 3.5 and 3.5.1. Any ideas on how to address
this? TIA -

W. Jeffrey Rankin
Lead Web Application Developer
=20
O'NEIL & ASSOCIATES, INC. http://www.oneil.com
495 Byers Rd.
Miamisburg, Ohio 45342-3662
Phone: (937) 865-0846 ext. 3504
Fax: (937) 865-5858
E-mail: jr...@on...
=20


Confidentiality Notice
The information contained in this e-mail is confidential and intended for u=
se only by the person(s) or organization listed in the address. If you have=
 received this communication in error, please contact the sender at O'Neil =
& Associates, Inc., immediately. Any copying, dissemination, or distributio=
n of this communication, other than by the intended recipient, is strictly =
prohibited.

[Tclhttpd-users] tmp folder and Config gui

[Luciano E. <lu...@Br...>]

I tested tclhttpd 3.5.1 a few days ago and noticed a glitch: there used to=
 be this line in auth.tcl:

set Config(AuthDefaultFile) /tmp/tclhttpd.default

I got an error and noticed that tclhttpd would only run if I replaced the=
 line above with this:

set Config(AuthDefaultFile) D:/TclTk/httpd/debian/tclhttpd.default

i.e. the full path to tclhttpd.default.

I took so long to report the bug that it's been fixed already. I just got a=
 new version from SF and there is no error, even without my tweaking. But I=
 understand it because that very same line is still there.

Questions:

- How did the error go away if that problem line is still there?

- Also, is it possible to tell tclhttpd where the temp file should be? I=
 don't like to use C:\tmp.

- I diff'ed the entire structure and noticed several changes. Shouldn't it=
 be assigned a new version number?


In this new version, I still have to hack srvui.tcl with a "wm withdraw ."=
 statement. I know there is a "Config gui" option in tclhttpd.rc, but it=
 doesn"t withdraw the toplevel.

- What is the point of having that blank wish window hanging about if I=
 decide not to use a GUI?

Many thanks in advance.

-- 
Luciano Espirito Santo
Santos, SP - Brasil

Re: [Tclhttpd-users] Content Transformation/Filtering

[Nicolas B. <nic...@ma...>]

Colin McCormack wrote:

>I've added the facility to filter dynamic content en route to the client
>and the template cache.
>
>Now, just before the data goes out the spout, you have the chance to run
>a proc over it.
>  
>
One nice app that I can see for this is to generate specific (formatted) 
output conditionally based on dynamic data.
So one could basically server html, xml, text pdf...
Kinda like a simple decorator pattern for tclhttpd...Cool.....

>Url_PrefixInstall has a new option -filter $cmd which calls
>Httpd_Filter, which pushes a command prefix onto a per-socket stack.
>
>At a point immediately before dynamic data is returned to the client,
>each command prefix is popped from the stack, and invoked as [eval $cmd
>$sock $content].  The returned value becomes the new content.
>
>Similarly, filtering occurs before a template result is cached.
>
>This will be useful, for example, in providing a site-wide livery, or
>some form of URL specific, or doc domain directory specific
>transformation.
>  
>
Erh.., what's a livery?

regards,
nicolas boretos

>It may be possible to stack specific domains with this approach, I
>haven't tried it.
>
>  
>