I have been noticing a change in how votes are being cast for things like adding code changes, node cleanup, and even title changes (or perhaps I am just now paying attention). A small, but growing number of votes are being cast for keep instead of edit, which effectively blocks an edit.

I originally was thinking that this was perhaps the work of a few people seeing their role as a spoiler, but after reading Editing pirated content links in reaped nodes, I have come to the realization that perhaps these keep votes are being cast to encourage personal responsibility rather than the governing structure taking care of the mistakes poor wording or spelling choices of the node authors.

I also looked at my own values, and discovered that I have been casting moderation votes against what I try to encourage in real life -- the individual is responsible and expected to correct mistakes, possibly with help or assistance from the community.

We used to have a parade around my neck of the woods where a large percentage of the parade units were equine. Some of them brought their own scooper crew or horse daipers with them, and other relied on the parade officials to run through with cleanup crews every so often. As a member of a marching unit (many moons ago), I know which equine groups I appreciated more. (ewwww)

How does this relate to my topic? Should we be encouraging posters to clean up after themselves, or should we be providing a service and cleaning up after them? I believe that the former is more likely to promote a sense of responsibility from the poster in the future, while the latter does not, and may even encourage maintaining the same behavior. Unless a kind monk /msgs the poster, will the poster even know their node has been moderated? [honest question, I don't know that I have ever had a moderated node, so I cannot say]

Perhaps there is a way to have something in the moderation system besides keep (do nothing) and edit (community does something). I will often times moderate a message to add code tags. At the same time I will /msg the OP to let them know that it was moderated and why. Perhaps, if this does not happen already, a message could be generated to the OP if their node was moderated. Perhaps there might also be a distinction made between keep/edit and "this should be edited, but it should not be the responsibility of the community to do it". I also believe that I will be voting keep on all but possibly the AM posts, and sending a /msg to the OP as a (polite) request to clean it up.

That is all I have at this time. I am interested in hearing other's thoughts.

I am having a hard time with sending SOAP / .NET "complex" structures. If I send the following response, my .NET (well, .NET WebServiceStudio) does not recognize the response -- it comes back with a "null" as the value of the string.

If I hardcode the return string to the above, I get the same response. However, if I change the attributes on the "Test1Response" tag above to be <namesp1:Test1Response xmlns:namesp1="urn:TESTServices">, all works fine.

My function in my handler is:

Does anyone have a good example of how to add the explicit namespace tag to the service name (the first child of the Body tag) using SOAP::Lite? I am not specifying the tag, I think it is supplied by SOAP::Lite.

Thanks,

update: Missed name space on closing tag.

PDF:API2 used to have a method pdfimage that would allow you to grab a page of a PDF file and treat the return value like an image. This appears to no longer be viable.

What would be the new way to pull a page from one document, scale it, and place it into an existing page in another document? I have seen CAD::PDF, but I don't have it installed (yet), and would prefer not to have to load yet another PDF library. CAD::PDF does look useful, however, so I will be looking at it in the future.

I have a third-party library that uses shell-style return codes (0 == success, anything else == error). An example chunk of code might be:

my $libobj = Third::Party::Library->new(...);
my $item = $libobj->NewItem(...);
$item->frobinize(...) and die($libobj->GetErrorCode);
$item->blatinate(...) and die($libobj->GetErrorCode);
[download]

The ... and ... just does not seem very perlish. Are there other patterns that are more compact, perlish, and easy to maintain? It does seem as compact and readable as can be, but the ... and ... just seems wrong.

Any thoughts?

Update:I think that I may have tried to generalize my original post too much, when I should have focused it explicitly on just HTML_TEMPLATE_ROOT. Please read this in the context of H::T and the HTML_TEMPLATE_ROOT variable.

I am running into a methodology question:

• Do I set my search paths from the environment
• Do I point to a configuration file where I set up my environment

In particular, HT's HTML_TEMPLATE_ROOT variable only allows a single path. If you have a set of paths that you wish to search, you need to make sure to put the paths into the ->new() call for the template.

This may not always be possible or desireable.

Take, for example, a scenerio where you tie parts of a system together with environment variables. You pick multiple paths for certain versions of libraries, PERL5LIB for the same, and you also want to pick some templates from a set of choices. I can create a configuration file and point to that, but now my configuration is in two locations, the shell script script starting the whole thing (both perl and non-perl parts) and the second configuration file.

Additionally, if I require my HTML::Template->new call to include the path attribute, I now need to make sure that every call to ->new, including those that I do not have control over, allow me to tweak the include path.

On the down side, there may be a performance penalty if the search path on new is combined with a multi-path HTML_TEMPLATE_PATH environment variable.

There is also the potential for security issues if someone else can tweak your environment. But if that is the case, then PATH is also suspect.

Any thoughts on this? I am leaning heavily toward patching HT (and submitting it, of course), but am not sure if there are other implications.

Back in Loading a 32-bit library via DynaLoader into 64-bit Perl 5.8.8 on HP-UX 11i, I found that I needed to recompile my Perl 5.8.8 in 32-bit mode due to an API library that I need only being available as a precompiled wrapper to an application we use only being available in 32-bit mode.

I have the 32-bit version of gcc 3.4.5 (from the hp development center), but the Configure script in the Perl distribution is forcing the lp64 flag to the compiler, which isn't recognized as a valid flag under the 32-bit version.

Any idea what flags I need to pass to configure to force 32-bit mode on this platform?

Loading a 32-bit library via DynaLoader into 64-bit Perl 5.8.8 on HP-UX 11i appears to be difficult, at least with a library from one of my vendors. I am getting a generic DynaLoader message, that when running under the debugger turns out being a "Bad magic number" message, which I am assuming is indicative of trying to load a 32-bit library into a 64-bit perl.

The vendor does not have a 64-bit version of the library available (although they have "certified" the application to run under 64-bit HP-UX 11i). I have applied appropriate pressure, but, as is typical with this vendor, the solution will be a long time coming.

I guess I am looking for build flag advice or possible tips on getting the vendor's library to load. Perl is compiled with HP's hp-gcc64-3.4.5 distribution. If I do need to revert to a 32 bit version of gcc, then so be it, but I would rather not.

Any ideas?

I am finding a use for a relocatable perl on HP-UX. I have looked at supersearch, found the announcement that TPF approved a grant to implement relocation in Feb 2006.

I am curious as to the status of this. Is there any update page that I can look at to find details?

Thanks all.

Update: Specifically, I am looking for 5.8.8 :). A piece of info that might be useful.

All,

Using XEmacs 21.4, cperl-mode, and Test::Inline blocks between =begin testing / =end testing blocks, has anyone made emacs treat the stuff between the =begin testing / =end testing blocks font lock and indent as perl code?

I would like a block like this...

=begin testing

my $obj = foo();
is($obj, "blatz", "beer run(s)");

=end testing
[download]

... to act like ...

=begin testing

=cut

my $obj = foo();
is($obj, "blatz", "beer run(s)");

=end testing
[download]

... as far as cperl is concerned, but not use the =cut explicitly, since then perl chokes on it :)

Before I dig into the cperl source, I want to see if this river has already been crossed by another user.

Is this a reasonable approach to injecting a check into the "warn" call to remove certain warnings from your error stream? I would like to fix the warnings, but this is from 3rd party code, so the best I can do without rewriting their code (and introducing that mess) is to filter the errors out.

I don't want to filter all of these warnings out, just the ones from Foo::Bar.

#!perl -w
...
use Foo::Bar;
...
{
 my $lastwarn = $SIG{__WARN__};
 # Filter out garbage warnings
 $SIG{__WARN__} = sub {
   return if ($_[0] =~ m/Use of uninitialized value in subroutine.*Foo
+\/Bar.*/);
   (defined($lastwarn) && $lastwarn) ? &$lastwarn(@_) : warn(@_);
 };
}
[download]

Thanks all.

Update: Fixed error in regex.

Update 2: It appears (silly me) that the -w in the shebang line has implications (like making warnings global, for instance). I guess this is one of those "looking at the problem too long" type solutions (thanks tye).