IE8 (and I assume other browsers that use a doctype pre-html5) interpret the span outside of the link, as in HTML4.01 you needed the closing tag.
http://www.w3.org/TR/REC-html40/struct/global.html#h-7.5.4

The code was causing a NPE in IE8 and users could not sort the columns.

>> cell.innerHTML = '<a href="#" class="sortheader">'+this.getInnerText(cell)+'<span class="sortarrow" /></a>';

>> cell.innerHTML ;
"<A class=sortheader href="#"> S<SPAN class=sortarrow></A></SPAN>"

>> cell.innerHTML = '<a href="#" class="sortheader">'+this.getInnerText(cell)+'<span class="sortarrow"></span></a>'

>> cell.innerHTML ;
"<A class=sortheader href="#"> S<SPAN class=sortarrow></SPAN></A>"

[FIXED JENKINS-22681] Fix of JENKINS-19418 should have used getById, not load, to avoid reloading the last build gratuitously.

As per discussion under PR #1194, it should go into separate plugin.

[FIXED JENKINS-22631] Do not just warn about multiple builds with the same number; note this in the official problem list, and offer to fix it.

[FIXED JENKINS-21622] Release workspace leases (and avoid @2) even if post-build actions throw LinkageError.

[FIXED JENKINS-22607] Avoid loading recent builds when deletion of builds older than some date has been requested.

Integrated v2.38

(Forgot that Run’s are unmarshalled in place after a special constructor is called, so there is no need for readResolve or other tricks.)
Also calling RunListener.onDeleted outside of the Run lock to avoid problems with things like ChangeLogHistoryRunListener.

The actual fix has been implemented over the last dozen or so commits

[FIXED JENKINS-18364] Functions.getRelativeLinkTo is called many times during rendering, so skip checking View.getItems which is expensive.

Avoid extracting war file under /tmp. When jar files and resource files disappear from under
us, all sorts of strange errors happen.

[FIXED JENKINS-22382] Need to restore deprecated overload of BuildAuthorizationToken.checkPermission for binary compatibility.

User.current() can return null, so this better check for null