I've been looking at a question I asked 3 years ago in Refactoring webcode to use templates

How things have changed since then. We've closed down the part of the business that I refactored all the code for, but I certainly learnt a lot in the process.

One of the things I refactored and now do as standard is to have pretty much all common code in modules. Although they were common in my code until a few years back, I now never use the require keyword. This got me thinking...is require ever still required or is it obsolete in the modern world?

I'm writing an XML Sitemap generator based around WWW::Crawl

I want to record the priority to set each entry in the sitemap. My first thought was to use a CSV or similar text file but it could become huge and cumbersome. So what are the alternatives?

I could write this server-side where I have a MariaDB instance running so storage is no problem. But I'm thinking I want to run it client side although I don't really know why. So my choice seems to be to hold the data in a Storable object. Run MariaDB, MySQL or similar locally or use DBD::SQLite from within Perl. No doubt there are other choices...

Which would you do and why?

What would you definitely avoid doing and why?

This has absolutely nothing to do with Perl but I know the vast range of knowledge and experience in The Monastery might save me a lot of searching...

A client is using Ecwid e-commerce platform and wants to sent the data to a Go High Level (GHL) CRM. I've set up a webhook that notifies GHL when an order is placed. GHL then calls the Ecwid API to get email address, name, etc. Everything works fine except the email addresses get mutilated in the process!

abc123@example.com     -> abc123@example.comabc
andrew.test@gmail.com  -> andrew.test@gmail.comandrew.test
a.test123@example.com  -> a.test123@example.coma.test
[download]

From the limited dataset I have it appears that everything from the start of the email address to the @ or the first numeric digit is appended to the end of the email address...

Have you ever seen anything like this before?
Do you have any suggestions before I spend too long trying to work out where the problem is and what is causing it, then dealing with two lots of customer support teams who are likely to blame each other.

I've seen various options for where to place POD:

• Next to the methods
• At the end of the module
• In a separate file

For a module I am going to publish on CPAN, I put the POD at the end of the file. I figure that the current version of the module will be 'finished' before it gets uploaded to CPAN.

I assume that a separate POD file is only really helpful for large modules that have several packages that need to be covered by one POD file.

But, for one project, I have a module of helper functions - this is a private module, not something I will publish. The module deals with things I either want to access in different places or that it makes sense to take out of the main scripts. Currently, this file has comments to document the methods, which is OK for a short module. But it's now exceeded 700 lines with a couple of dozen methods, and finding my way around it is becoming frustrating and time-consuming.

I have started to create POD for this module to make it easier for me to find the method and syntax I need when I add some new functionality or update existing functionality.

Because the module is never really 'finished' and gets added to whenever I need a new method, it seems sensible to add the POD next to each method. But I am sure there is more to it than this...am I opening myself up to future problems if I spread the POD through the module and document each method next to that method's code?

Where do you place your POD, and why do you do it that way?

I've been checking CPAN Tester results for one of my modules and noticed a single failure amidst a sea of passes. On checking it, is from Test::More is returning 0.99999999999999996 instead of 1.

My best guess is that the machine doing the testing has a peculiar build somewhere that is affecting the floating point arithmetic.

• Is that a fair assumption?
• Is there anything I can do about that?
• Is it worth doing anything about this?

I realise that we've had a very long recent thread about Unicode...sorry if this drags the issue out further! However, I have little understanding of Unicode, and I rarely needed to know. But I could do with some help, please...

My partner runs a dog care business and I have built the booking platform for her. Part of this provides a URL to Google Calendar to update our mobile calendars from the booking system. This all works fine but I'll ask a couple of more general questions at the end.

I decided it would be nice to have a Dog Face Emoji as the first character of the title of the calendar entry. But I cannot get this to display. The script uses Template to generate the ICS feed for Google Calendar.

BEGIN:VCALENDAR
VERSION:2.0
PRODID:Pawsies Calendar 1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
[% FOREACH event IN day %]BEGIN:VEVENT
SUMMARY:\x{e052} [% event.type %][% IF event.dog %] - [% event.dog %][
+% END %]
[% IF event.note %]DESCRIPTION: [% event.note %]
[% END %]UID:pawsies[% event.idBooking %][% event.id %]@pawsies.uk
SEQUENCE:[% event.sequence %]
DTSTAMP:[% event.dtstart %]
DTSTART:[% event.dtstart %]
DTEND:[% event.dtend %]
URL:https://www.pawsies.uk/admin/calendar/?day=[% event.date %]
COLOR:[% event.color %]
END:VEVENT
[% END %]END:VCALENDAR
[download]

Everything works as expected except the emoji is printed as a literal \x{e052} instead of a 🐶 - I have use utf8; at the top of the script and the HTTP Header is:

Content-type: text/calendar; charset=utf-8
[download]

A couple of extra questions if you are experienced at feeding data to Google Calendar:
1 - Is it possible to force Google to refresh the feed? Waiting for 24 hours or so makes debugging slow and tedious.
2 - Is it possible to set the colour of the event from the ICS feed so we can have multiple colours from one feed? Currently, we have two feeds just to get two different colours. I've tried the COLOR property but it seems to be ignored.

Updated to correct MIME type

I've released an updated version of AI::Embedding - version 1.1

The CPAN documentation shows the latest version. However, CPAN Testers is showing N/A for all Perl versions, not just the ones before the minimum Perl version. This seems very strange. Have you come across this before?

Also, when I try to upgrade to the latest version, it doesn't happen

C:\Users\bod>cpanm AI::Embedding
AI::Embedding is up to date. (1.01)
[download]

I've checked that I haven't inadvertently specified a minimum Perl version that doesn't yet exist.

Any suggestions for what to check that might be causing this?

Since Aldebaran found a problem with AI::Embedding in Re^2: perl hooks for AI, I have updated the module.

But, when I use prove to run my tests, it is using the installed version of AI::Embedding and not the one in my test directory. How can I get prove to ignore the version from CPAN and use the test version?

Recently, I've noticed a few places where anonymous subroutines have been declared for no apparent reason:

my $function = sub {
    # ...do stuff...
};

# Call the sub
$function->('param');
[download]

sub function {
    # ...do stuff...
}

# Call the sub
function('param);
[download]

I don't think I have ever written a stand-alone anonymous subroutine like this. The only place I generally use anonymous subs is as a 'value' in a hash. Typically, to pass to Template to keep logic separated from display like this:

my $vars = {
    'date'    => $date,
    'format'  => sub { return ucfirst lc $_[0]; },
};
$template->process('example.tt', $vars);
[download]

I have two email-sending processes. Both are seemingly the same but they produce different results. Both use MIME::Lite.

One sends emails immediately, the other is for larger numbers of emails and consists of an email queue held in a MariaDB database. The sending script runs every 20 minutes from CRON. The body text for both gets to them from an HTML textarea and they have been tested using identical body text.

If we add an emoji character to the body text, the queue sends it correctly. But the immediate sending script produces ��� instead of the emoji.

As far as I can tell the only differences between the scripts is that one runs under CRON and stores and retrieves the body using a database. Other than that they seem identical.

Both scripts use utf8;

I am at a loss how to debug the problem...what would you try to get to the root of this problem?