Hi,
we noticed a different number of (unfortunately very expensive) ocsp-requests shown in waterfalls delivered by version 126 and 130 of wptdriver. We tested with actual Firefox the following script:
logData 0
setCookie http://www.otto.de uniqueUserId = 806fd6a9-82e9-4f22-b1c0-ee262f1c4cfa; expires = 01-Jan-2080 00:00:00 GMT
/// Produkt 1 in Warenkorb
navigate http://www.otto.de/suche/?commit=true&ls=0&commit=true&fh_search=shirt&&fh_search_requested=shirt&sterm=shirt&stype=N
execAndWait document.querySelector('#p0 a.js_ProductListProductRef').click();
execAndWait document.getElementById('TconImgAddToBasket').click();
/// Produkt 2 in Warenkorb
navigate http://www.otto.de/suche/?commit=true&ls=0&commit=true&fh_search=brille&&fh_search_requested=brille&sterm=brille&stype=N&SelectRuleEngineVariants=0
execAndWait document.querySelector('#p0 a.js_ProductListProductRef').click();
execAndWait document.getElementById('TconImgAddToBasket').click();
/// PRodukt 3 in Warenkorb
navigate http://www.otto.de/suche/?commit=true&ls=0&commit=true&fh_search=bikini&&fh_search_requested=bikini&sterm=bikini&stype=N&SelectRuleEngineVariants=0
execAndWait document.querySelector('#p0 a.js_ProductListProductRef').click();
execAndWait document.getElementById('TconImgAddToBasket').click();
/// PRodukt 4 in Warenkorb
navigate http://www.otto.de/suche/?commit=true&ls=0&commit=true&fh_search=adidas&&fh_search_requested=adidas&sterm=adidas&stype=N&SelectRuleEngineVariants=0
execAndWait document.querySelector('#p0 a.js_ProductListProductRef').click();
execAndWait document.getElementById('TconImgAddToBasket').click();
/// Produkt 5 in Warenkorb
navigate http://www.otto.de/suche/?commit=true&ls=0&commit=true&fh_search=nike&&fh_search_requested=nike&sterm=nike&stype=N&SelectRuleEngineVariants=0
execAndWait document.querySelector('#p0 a.js_ProductListProductRef').click();
execAndWait document.getElementById('TconImgAddToBasket').click();
logData 1
execAndWait document.getElementById('TconImgGotoBasket').click();
In wptdriver 126-tests we had much more ocsp-requests than in actual wptdriver 130-tests (see hars in attachments). We think (and observed in wireshark manually) that wptdriver 126-results are more realisitic.
Could you tell us, what changes between 126 and 130 could be responsible for this. And what changed in measurement of HTTPS-sites with wptdriver the last weeks?
Thanks in advance,
regards, Nils