"Cutchins, Stephen" <stephen.cutchins@accenturefederal.com>, 2015-07-24 14:03 +0000: > Archived-At: <http://www.w3.org/mid/DM2PR10MB030133DDBE2C1D909C424E88F6810@DM2PR10MB0301.namprd10.prod.outlook.com> > > Hi, > > We have been using Validator with an automated test script for a Federal > project. Roughly a week ago or so, the counts were removed from > Validator which is causing our automated test scripts to fail since this > is what we were checking against. > > Is there a way to add the counts back? We would very much appreciate it! You can use the JSON output from the W3C HTML Checker and programatically count the number of message types you’re interested in; for example: https://validator.w3.org/nu/?doc=https://accenturefederal.com&out=json If you’re interested only in the error count, your test scripts can programatically count just the “type: error” items in that JSON output. If you’re interested in warnings as well, your test scripts can also count the “type: info” items that have a “subtype: warning” property. Depending on what kind of test scripts you’re using, there are other machine-parsable output formats you might want to try: https://validator.w3.org/nu/?doc=https://accenturefederal.com&out=xml https://validator.w3.org/nu/?doc=https://accenturefederal.com&out=gnu The out=gnu format puts each message on a single line, so your test script could get an error count from that just by piping a few commands together: curl -s "https://validator.w3.org/nu/?doc=https://accenturefederal.com&out=gnu" | grep ": error:" | wc -l Or to get a count of the warnings: curl -s "https://validator.w3.org/nu/?doc=https://accenturefederal.com&out=gnu" | grep ": info warning:" | wc -l -- Michael[tm] Smith https://people.w3.org/mike
Received on Monday, 27 July 2015 00:18:33 UTC