block_copy errors; not sure how to proceed

I am upgrading from 1.1.9 to 20.1. I read System/UpgradeGuide#Copy_the_data_using_61tools_47bulk_copy.pl_61

I'm trying to run perl bulk_copy.pl.

I'm getting errors: Tasks/Item13704

I have Perl 5.10.0 running on Mac OS X, 10.8

Do I need to run this as root?
Do I just cross my fingers and hope for the best?
Shall I just copy / move directories the way I've always done?

-- VickiBrown - 16 Sep 2015

The task you created is a duplicate of a known issue running on older versions of perl. 5.10.0 is quite old.

About using bulk_copy.pl ... we are learning more about the conversion process as some of the larger sites begin the migration to utf-8. One thing that will really break bulk_copy is if users have cut/paste topic text from Microsoft word, or other products that use the MS "Smart quotes" and other windows-1292 codepage special characters. In that case, you don't actually have source data compliant with iso-8859-1, and conversions of those topics will fail.

The CharsetConverterContrib is quite a bit more forgiving in the conversion process, though there are recent updates that have not yet been released.

A 2nd issue that we have not yet addressed is topics with links to attachments that use "extended ASCII" characters - umlaut, etc. in their filename. The "insert link to attachment" inserts the links using entity encoding, so they appear as plain ascii during the conversion, and will be unreachable in the converted wiki. Currently the solution is to manually fix up the topic names and links, removing the umlauts and other characters from the name. We are looking into a more automated solution but it is a very complex issue.

-- GeorgeClark - 16 Sep 2015

Oh... and there are other issues on the OS X platform. OSX uses a different unicode normalization scheme from Linux. Currently there are known issues with unicode on OS X that have not been addressed. See Tasks.Item13405.

-- GeorgeClark - 16 Sep 2015