[8690] in athena10
[Debathena] #1109: XDG_CACHE_HOME should be pointed to a local
daemon@ATHENA.MIT.EDU (Debathena Trac)
Fri Nov 25 00:44:31 2011
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
From: "Debathena Trac" <debathena@MIT.EDU>
Cc: debathena@MIT.EDU
To: davidben@MIT.EDU
Date: Fri, 25 Nov 2011 05:44:25 -0000
Reply-To:
Message-ID: <044.3f745efd891d1e481d848ac5652b1db6@mit.edu>
Content-Transfer-Encoding: 8bit
#1109: XDG_CACHE_HOME should be pointed to a local directory
----------------------------+--------------------------------
Reporter: davidben | Owner:
Type: enhancement | Status: new
Priority: normal | Milestone: The Distant Future
Component: -- | Keywords:
Upstream bug: |
----------------------------+--------------------------------
The [http://standards.freedesktop.org/basedir-spec/basedir-spec-
latest.html XDG base directory spec] defines several variables including
XDG_CACHE_HOME where applications are supposed to stuff cache data. There
aren't all that many that create cache data of any significance, but
browsers do. Google Chrome (and unbranded Chromium builds) uses it.
Firefox does not, but they already split their profile into local/remote
halves for Windows. I intend to shepherd a
[https://bugzilla.mozilla.org/show_bug.cgi?id=239254 patch] in Firefox to
put the local half into XDG_CACHE_HOME.
Debathena should consider pointing XDG_CACHE_HOME to something local. That
would get local cache in Chrome for free, and same with Firefox once the
patch goes through. This would also fix issue #329 as
urlclassifier3.{sqlite,pset} are also part of the local profile. So are
the offline cache and startup cache. (This would not fix #1107 as
Thunderbird does not treat ImapMail as local data.) This isn't explicitly
stated as the intention in the spec (not that rigorous reading of a spec
is remotely meaningful for desktop APIs), but random searches on Google do
suggest that NFS homedir and local XDG_CACHE_HOME are common.
It's conceivable that bad things will happen with an app running in two
concurrent logins but each sees different caches, so it's probably worth
testing a few things. (Then again, I suspect they all already break on
concurrent logins anyway.)
--
Ticket URL: <http://athena10.mit.edu/trac/ticket/1109>
Debathena <http://debathena.mit.edu>
MIT Debathena Project
