Now cleaning old firefox archives.

Now cleaning old firefox archives.

We have a few gigs of old archives out on the bots, and nothing
to clean them up. There's no reason to keep more than one old
archive around, so this makes the download script prune them.

I considered making this a separate script, but arguably it's the
job of the download script to deal with its byproducts (and I
could share some code) so I made the existing script do it.

BUG=360516
R=kjellander@chromium.org

Committed: https://chromium.googlesource.com/chromium/deps/webrtc/webrtc.DEPS/+/293897

[phoglund_chromium, 2015-02-02 09:30:25]

phoglund@chromium.org changed reviewers:
+ kjellander@chromium.org

[kjellander_chromium, 2015-02-02 09:45:00]

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py
File download_firefox_nightly.py (right):

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py#...
download_firefox_nightly.py:45: if not firefox_archives:
Add:
or len(firefox_archives) < 2 
condition here?

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py#...
download_firefox_nightly.py:57: pass
I think it's unlikely we will fail to delete these. Isn't it better to have such
a failure actually be reported?

[phoglund_chromium, 2015-02-02 10:04:11]

PTAL

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py
File download_firefox_nightly.py (right):

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py#...
download_firefox_nightly.py:45: if not firefox_archives:
On 2015/02/02 09:45:00, kjellander wrote:
> Add:
> or len(firefox_archives) < 2 
> condition here?

Done.

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py#...
download_firefox_nightly.py:57: pass
On 2015/02/02 09:45:00, kjellander wrote:
> I think it's unlikely we will fail to delete these. Isn't it better to have
such
> a failure actually be reported?

I don't think so. If we fail here it's not enough of a big deal to fail the
build, I think. We may as well make this code robust if it ever runs on Windows
(where a delete will fail if the archive unpacker has the file open but has
hanged, for instance).

[kjellander_chromium, 2015-02-02 10:13:46]

lgtm

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py
File download_firefox_nightly.py (right):

https://codereview.chromium.org/891303002/diff/1/download_firefox_nightly.py#...
download_firefox_nightly.py:57: pass
On 2015/02/02 10:04:11, phoglund wrote:
> On 2015/02/02 09:45:00, kjellander wrote:
> > I think it's unlikely we will fail to delete these. Isn't it better to have
> such
> > a failure actually be reported?
> 
> I don't think so. If we fail here it's not enough of a big deal to fail the
> build, I think. We may as well make this code robust if it ever runs on
Windows
> (where a delete will fail if the archive unpacker has the file open but has
> hanged, for instance).

Fair enough. Worst case we'll just run out of disk, which should hopefully be
detected by the Chrome infra's monitoring efforts before it happens.

[phoglund_chromium, 2015-02-02 10:30:05]

Committed patchset #2 (id:20001) manually as r293897 (presubmit successful).