While reading an article on MarketWatch using Firefox Nightly I was prompted to fill out a user survey. I’ve stopped responding to most of these types of requests since they have become more of a data gathering exercise about me rather than actually attempting to find out what I think of the site itself. Since I like MarketWatch and haven’t seen this request before, I accepted the prompt this time and prepared to see what they wanted to know. Unfortunately, I was blocked by the following:
Needless to say, I didn’t switch to Chrome to fill out their survey and now have a negative opinion of MarketWatch and UserZoom.
Update January 21
I did send in feedback to MarketWatch regarding the situation but did not receive a reply. Just now I visited again and received another prompt for a user survey and accepted in order to see if they had fixed the situation.
Close but no cigar.
I guess detecting Firefox is just too hard. For the record, my user agent string is Mozilla/5.0 (X11; Linux x86_64; rv:86.0) Gecko/20100101 Firefox/86.0
7:15 AM PT – Swapped Kingston Class 4 4G sdcards for 00_23_76_96_cc_6f_nexus-one and c8_aa_21_ac_0c_b5_droid-pro after repeated failures checking for the device root.
19:00 PM PT – Both 00_23_76_96_cc_6f_nexus-one and c8_aa_21_ac_0c_b5_droid-pro showed an inability to properly detect and use the external sdcard. I cycled through most of the previously used sdcards attempting to find ones that worked. After checking almost all of the remaining older cards, I found one that worked in the droid-pro but could not find a card that would work with the nexus-one. In my attempt to get a working patch for bug 933842 – Add ability to specify test root in SUTAgent.ini, I realized that mozdevice’s devicemanager.getDeviceRoot() caches the reported value of the test root. Since the autophone workers survive the rebooting of a device, the cached test root value is reused until the device’s worker is restarted. This means that once an incorrect test root is detected, it will not be corrected by rebooting the device. I am testing an preliminary version of the patch for bug 933842 on the nexus one which forces the test root to /mnt/sdcard. I have resubmitted all builds from 2013-10-31 to now for both nexus ones and the droid pro. Hopefully the devices will be more reliable. It may also be the case that the older sdcards are not as unreliable as initially thought.
I also noticed that Autophone’s workers use devicemanager’s getDeviceRoot() to determine if a device is still responsive. Since the value is cached, this test is ineffective in determining if a device is still responding.
bug 804865 recently introduced the requirement to use Python 2.7 when building Firefox Nightly (Firefox 21). Since only Python 2.6 is available on Redhat Enterprise Linux 6 (RHEL6) you will need to build Python. A working version can be built by:
sudo yum install bzip2-devel sqlite-devel ncurses-devel openssl-devel readline-devel tkinter tk-devel
curl -O http://www.python.org/ftp/python/2.7.3/Python-2.7.3.tar.bz2
if ! md5sum Python-2.7.3.tar.bz2 | grep c57477edd6d18bd9eeca2f21add73919; then
echo "Failed md5sum check"
tar xjvf Python-2.7.3.tar.bz2 && \
cd Python-2.7.3 && \
./configure && \
make && \
sudo make altinstall
This will produce a working version of Python that can be used to build Firefox Nightly though the following modules will not be built.