You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This gives us more accurate version scraping that's not reliant on the cache of the HTML file listing (which instead matches how `yum`/`dnf` query these version numbers, so should stay in sync better).
This should fix our issues with CI sometimes returning a mismatch in 8.0 versions.
**However**, this *causes* a mismatch in 5.7 versions because the latest 5.7 release (that we were not picking up before due to the aforementioned HTML caching but are now with this updated code) is apparently not built/released for Debian (https://dev.mysql.com/downloads/mysql/), so we also need to decide whether we're going to introduce Debian/Oracle version skew or perhaps deprecate/remove the Debian variants of 5.7 (which is ostensibly EOL in ~3 months).
Copy file name to clipboardExpand all lines: versions.sh
+26-1Lines changed: 26 additions & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -23,7 +23,28 @@ fetch_rpm_versions() {
23
23
local oracleVersion="$1";shift
24
24
local package="$1";shift
25
25
26
-
curl -fsSL "$repo/$arch/"2>/dev/null \
26
+
local baseurl="$repo/$arch"
27
+
28
+
# *technically*, we should parse "repodata/repomd.xml", look for <data type="primary">, and use the <location href="..."> value out of it, but parsing XML is not trivial with only basic tools, it turns out, so instead we rely on MySQL's use of "*-primary.xml.*" as the filename we're after 👀
29
+
local primaryLocation
30
+
primaryLocation="$(
31
+
# 2>/dev/null in case "$arch" doesn't exist in "$repo" 🙈
0 commit comments