Перейти к файлу
Matt Wobensmith 5cea7af08e Updated status message to reflect remaining number of sites to be loaded 2015-01-02 12:25:29 -08:00
.gitignore Adding .gitignore 2014-04-30 13:00:41 -07:00
README.md Updating documentation 2014-10-08 16:28:50 -07:00
compare.sh Improving more 2014-10-08 16:28:51 -07:00
getXHRSSLStatus.js Updated status message to reflect remaining number of sites to be loaded 2015-01-02 12:25:29 -08:00
pulse-domains-master.txt Removing blank lines 2014-10-08 16:28:01 -07:00
run.sh Add driver, remove unused 2014-04-28 15:28:33 -07:00
runforcompare.sh Removing early exit 2014-10-08 16:29:50 -07:00
saverun.sh Changes to support running and re-running large surveys with multiple options 2014-10-08 16:28:01 -07:00

README.md

pkix-compatibility-testing

Running a scan

Given an xpcshell instance, run the following:

$ .../xpcshell getXHRSSLStatus.js domains.txt errors.txt ev.txt >output.txt 2>errors.txt

For the full list of domains, this could take a long time. Include a shorter set if you need.

Additional command line options determine which file any errors are logged to, and which file EV certificates are logged to.

Additional tweaks can be made by editing the script. This isn't a full-service operation :)

Comparing gecko builds

If you have a local build of gecko in a git repository, you can use runforcompare.sh to set this up. Simply create a domains.txt file and point the script to your gecko repository.

$ ln -s pulse-domains-master.txt domains.txt
$ ./runforcompare.sh ../gecko-dev

This saves files that use the branch name of your build: domains.<branch>.errors and domains.<branch>.ev. This allows you to compare different builds easily.

Comparing runs from different builds

The compare.sh script compares the output of two runs. Simply pass it the names of the branches involved.

$ ./compare.sh master bug1024576

This produces files including a comparison of the two named runs, including a list of domains that produce different error codes between the scans, and a file containing the domains that differ between runs.

Re-running scans

Sometimes results are a little noisy, so re-running a scan over the set of domains that might be different is useful.

The saverun.sh script saves the current run and builds a list of domains for rechecking. This list includes all domains that produced different responses between all the current runs (for which there might be more than two). The (hopefully shorter) list is automatically symlinked to domains.txt in preparation for the next run.