I am trying to use git-annex as part of a datalad install on an Ubuntu 18.04 VirtualBox VM to set up a repository linking to some datasets at a public ftp server. When I use git-annex version 7.20190219 (installed with apt-get install git-annex-standalone from NeuroDebian) to add the links, I get the following error:

emmet@emmet-VirtualBox:~/conp-dataset/projects/1KGP-22May2019$ git annex addurl ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz –file=1KGP_chr6-vcf.gz addurl ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz Unsupported url scheme ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz

download failed: Unsupported url scheme ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz failed git-annex: addurl: 1 failed

However when I use git-annex v 6.20180227 (installed from the apt repository that came with Ubuntu 18.04) the link is generated correctly:

emmet@emmet-VirtualBox:~/conp-dataset/projects/1KGP-22May2019$ git annex addurl ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz --file=1KGP_chr6-vcf.gz addurl ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz ALL.chr6.phase3_shapeit2_mvncall_integrated_ 100%[===========================================================================================>] 953.14M 9.81MB/s in 93s
2019-05-22 16:10:44 URL: ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr6.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz [999436830] -> "/home/emmet/conp-dataset/projects/1KGP-22May2019/.git/annex/tmp/URL-s999436830--ftp&c%%ftp.1000genomes.ebi.ac.uk-dda0592051d00d19f6c947a6b3aae6a5" [1] (to 1KGP_chr6-vcf.gz) ok (recording state in git...)

Similarly, after publishing the correctly formed links generated with 6.20180227 to github, trying to download them on another Ubuntu VirtualBox VM with git-annex 7.20190219 installed causes the following error:

emmet@emmet-VirtualBox:~/conp-dataset/projects/1KGP-22May2019$ git annex get 1KGP_chr2-vcf.gz get 1KGP_chr2-vcf.gz (from web...) download failed: Unsupported url scheme ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr2.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz

Unable to access these remotes: web

Try making some of these repositories available: 00000000-0000-0000-0000-000000000001 -- web 2cc9a182-813f-4740-8157-c7a560560238 -- emmet@emmet-VirtualBox:~/conp-dataset/projects/1KGP-22May2019

(Note that these git remotes have annex-ignore set: origin) failed git-annex: get: 1 failed

but downloading them on that same VM with git-annex v 6.20180227 installed instead works:

emmet@emmet-VirtualBox:~/conp-dataset/projects/1KGP-22May2019$ git annex get 1KGP_chr2-vcf.gz get 1KGP_chr2-vcf.gz (from web...) ALL.chr2.phase3_sha 100%[===================>] 1.22G 6.99MB/s in 3m 11s 2019-05-22 16:25:44 URL: ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr2.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz [1312735578] -> "/home/emmet/conp-dataset/projects/1KGP-22May2019/.git/annex/tmp/MD5E-s1312735578--774370621affaf125e6994089f975ed7.gz" [1] (checksum...) ok (recording state in git...)

Would you know why this is happening? And how I can avoid this failure mode in the more up-to-date version of git-annex?

Added back support for ftp urls with no special configuration needed. done --Joey