import: Avoid duplicate trailing '/' in computed import URL. Unlike for file names, duplicate slashes in URLs matter. For example attempting to refresh 'qemu' would previously not find any result because it'd try to find links from 'https://download.qemu.org//', which is unreachable. * guix/gnu-maintenance.scm (import-html-release): Use canonicalize-url to join BASE-URL and DIRECTORY, which handles trailing '/' correctly. Change-Id: I108a542c4eede85c78a47f62921184273dc1ea01
1 files changed, 1 insertions(+), 3 deletions(-) M guix/gnu-maintenance.scm
M guix/gnu-maintenance.scm => guix/gnu-maintenance.scm +1 -3
@@ 677,9 677,7 @@ also updated to the latest version, as explained in the doc of the \"rewrite-url\" procedure used." (let* ((current-version (package-version package)) (name (package-upstream-name package)) (url (if (string-null? directory) base-url (string-append base-url directory "/"))) (url (canonicalize-url directory base-url)) (url (if rewrite-url? (rewrite-url url current-version #:to-version version #:partial-version? partial-version?)