DateTime crashed Devel::ptkdb

[autarch]

Migrated from rt.cpan.org #116924 (status was 'new')

Requestors:

• [email protected]

From [email protected] on 2016-08-12 20:17:05:

DateTime doesn't seem to work nicely with Devel::ptkdb.

Firstly if I monitor a DateTime object it reports a stringified version of the object.

Secondly if a DateTime object is part of a bigger structure - adding the structure to the expression to be monitored simply crashes the GUI.

This is how I reproduced the error.

$ cat t.pl use DateTime;

my $dt = DateTime->new(
year => 1964,
month => 10,
day => 16,
hour => 16,
minute => 12,
second => 47,
);

%arr = (k => $dt);

print %arr;

If I step up to line 14 then $dt shows 1964-10-16T16:12:47 and if I try to see %arr then the GUI crashes with the following error.

A DateTime object can only be compared to another DateTime object (DateTime=HASH(0x3066188), HASH(0x2a1c9c8)). at /cigdev64/borec/ExtData/2.0/opt/perl/lib/perl5/Devel/ptkdb.pm line 2658.

I can run the same script with perl -d and DataDumper correctly shows the complete hash structure of the object both in $dt and %arr without crashing.

I am using DateTime 1.34 Devel:,:ptkdb 1.1091 on perl 5.22.2

• Mithun

---

CONFIDENTIALITY AND SECURITY NOTICE

The contents of this message and any attachments may be confidential and proprietary. If you are not an intended recipient, please inform the sender of the transmission error and delete this message immediately without reading, distributing or copying the contents.

[matthewpersico]

There are two ways to solve this:

1. Overload == to call DateTime->compare()
2. Fix Devel::Ptkdb to not require == overloading.

I am working on the latter. But I don't know why the former is not implemented. @autarch , PR's accepted?

[autarch]

DateTime implements <=> with fallback enabled, so == overloading should just work. Is Devel::Ptkdb doing something weird?

I'm not keen on adding things to DateTime just to support a tool which hasn't been updated in 13 years.

[matthewpersico]

Well now that I know you're not at fault I'll put it on my list of things to investigate about ptkdb.

On Wed, Feb 22, 2017 at 16:55 Dave Rolsky ***@***.***> wrote: DateTime implements <=> with fallback enabled, so == overloading should just work. Is Devel::Ptkdb doing something weird? I'm not keen on adding things to DateTime just to support a tool which hasn't been updated in 13 years. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#24 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC-_duPQyfNsijHPzq47Sj2m6V5zAQqsks5rfK7tgaJpZM4Kwero> .

-- Matthew O. Persico

[autarch]

Just to confirm that this does work ...

$ perl -MDateTime -E 'my $now = DateTime->now; say $now == $now ? 1 : 0; say $now == DateTime->today ? 1: 0'
1
0

So whatever ptkdb is doing, it's something that somehow breaks the overloading fallback or ignores it. Maybe it checks directly if the object implements a specific overloading?

[matthewpersico]

Yeah. Something like that. Long story short I fixed this and a whole bunch of other nonsense in this module in 2013. But it was behind a corporate firewall and I had to leave it when I left. Been trying to find the time to recreate all those fixes since.

…

Sent from my iPhone

On Feb 22, 2017, at 18:05, Dave Rolsky ***@***.***> wrote: Just to confirm that this does work ... $ perl -MDateTime -E 'my $now = DateTime->now; say $now == $now ? 1 : 0; say $now == DateTime->today ? 1: 0' 1 0 So whatever ptkdb is doing, it's something that somehow breaks the overloading fallback or ignores it. Maybe it checks directly if the object implements a specific overloading? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[preaction]

So... It sounds like this is an issue with ptkdb, there's nothing to be done to DateTime, and we can close this?

[matthewpersico]

Go ahead

On Sat, Feb 25, 2017 at 00:26 Doug Bell ***@***.***> wrote: So... It sounds like this is an issue with ptkdb, there's nothing to be done to DateTime, and we can close this? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#24 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC-_diWN4m2yPT2rFI4DQE7nQKpGjPUfks5rf7uDgaJpZM4Kwero> .

-- Matthew O. Persico