broken links

[karenetheridge]

See #71 -- some broken links need to be fixed still, once replacements have been found.

[karenetheridge]

related to #7, #64

[vaitkus]

A few questions involving links to CPAN/packages.

lib/perlfaq4.pod:

• LinkedList::Single. The link, as it appears on GitHub and https://perldoc.perl.org/perlfaq4, does not resolve to a package on metaCPAN (https://metacpan.org/pod/LinkedList::Single). However, this module does exist on metaCPAN (https://metacpan.org/release/LEMBARK/LinkedList-Single-v1.3.0/view/lib/LinkedList/Single.pm). Should it be linked to directly using an full explicit link?
• Array::Iterator::Circular. Similar case to the LinkedList::Single. Interestingly enough, searching for it on metaCPAN via the search bar returns no results, even though this module exists (https://metacpan.org/dist/Array-Iterator/view/lib/Array/Iterator/Circular.pm). Should it be linked to directly using a full explicit link?

lib/perlfaq8.pod:

• TCP::Client. Is it stated that this package exists on CPAN, however, I was not able to find it neither via the CPAN search nor using a generic search engine. The package is not used in any of the examples. If a proper link to this package does not exist (I might have missed it), is it OK to remove the mention of it completely?

[Grinnz]

I would recommend leaving the link to LinkedList::Single. The latest version is not indexed in PAUSE and this must be corrected by the author, at which point the link will work.

For Array::Iterator, it seems the permissions of the sub-modules Array::Iterator::BiDirectional, Array::Iterator::Circular, and Array::Iterator::Reusable were not properly assigned to a new maintainer and so the main module of the distribution has been indexed but the sub modules have not. This is a permissions/maintainer issue that can also be corrected but may require intervention from the PAUSE admins.

If by Client::TCP you mean TCP::Client, there is an index entry for it but it is somehow indexed to T/TO/TOMC/scripts/ADVLABS/net/TCP/Client.pm.gz. This must be a historical artifact of CPAN (if you navigate there, it is Tom's Script Archive) and is not particularly usable.

[vaitkus]

I would recommend leaving the link to LinkedList::Single. The latest version is not indexed in PAUSE and this must be corrected by the author, at which point the link will work.

Understood. I would reach out to the author, however, I do not have sufficient experience to guide them through the required changes if they do answer.

For Array::Iterator, it seems the permissions of the sub-modules Array::Iterator::BiDirectional, Array::Iterator::Circular, and Array::Iterator::Reusable were not properly assigned to a new maintainer and so the main module of the distribution has been indexed but the sub modules have not. This is a permissions/maintainer issue that can also be corrected but may require intervention from the PAUSE admins.

Understood. Should anyone else in particular (e.g. PAUSE admins) be informed about this issue?

If by Client::TCP you mean TCP::Client, there is an index entry for it but it is somehow indexed to T/TO/TOMC/scripts/ADVLABS/net/TCP/Client.pm.gz. This must be a historical artifact of CPAN (if you navigate there, it is Tom's Script Archive) and is not particularly usable.

Yes, sorry, I somehow mistyped the name (I have now gone back and fixed in my original comment). Any suggestions how this could be best resolved (ask CPAN to reindex, link to the pm.gz using a full link, remove the mention, etc.)?

[Grinnz]

I have messaged the PAUSE admins regarding Array::Iterator's submodules. I think that the TCP::Client mention should just be removed.

[gwselke]

In PerlFAQ 5 there is a broken link to Numerical Recipes (http://www.nr.com). According to Wikipedia the domain nr.com was sold in 2015. The official website is now at http://nrbook.com/, including limited online access to the current version. Older versions are also available there or at https://archive.org/details/NumericalRecipes.

[davorg]

@gwselke I think you mean perlfaq4, not perlfaq5 - https://github.com/perl-doc-cats/perlfaq/blob/master/lib/perlfaq4.pod?plain=1#L414

The section we want to link to is "Random Numbers" at http://nrbook.com/book.html. But it's an annoying frame-driven site that makes it hard to deep-link to the correct page. It also doesn't use HTTPS, which makes it look dodgy on modern browsers.

I think we'd be better off looking for a better link (or just removing the link completely).

[guest20]

The Q there is about why the numbers from rand aren't random, so maybe squeezing the whole field of generating cryptographically random numbers in there as a foot note might be a little too ambitious?

It seems like the cpansec folks have a strong opinion about good random numbers, and even a document with references:
https://security.metacpan.org/docs/guides/random-data-for-security.html

... it could even be it's own Q. "Where can I find a decent random number around here?"‡

__
‡. Not that, but something like that

[gwselke]

@gwselke I think you mean perlfaq4, not perlfaq5 - https://github.com/perl-doc-cats/perlfaq/blob/master/lib/perlfaq4.pod?plain=1#L414
Ooops. Thanks for catching this sausage-finger-generated typo, Dave. It's perlfaq4 indeed.

The section we want to link to is "Random Numbers" at http://nrbook.com/book.html. But it's an annoying frame-driven site that makes it hard to deep-link to the correct page. It also doesn't use HTTPS, which makes it look dodgy on modern browsers.
:-( In any case, maybe just link to the general book URL and leave navigation to the reader?
I think we'd be better off looking for a better link (or just removing the link completely).
I, for one, still like the book (despite some controversy around it), it has very accessible presentations. So I think it's not bad. Otherwise, TAOCP is always a good source.

I think one cannot supply a link that will be always both up-to-date and authoritative. The important message is three-fold: a) random numbers aren't easy, don't try home-grown recipes; b) there are good solutions available in Perl; c) there's a whole science behind them, so start reading if you're really interested, and here you have some link to get you started.