public static boolean getFormFields(
ResponseWrapper rw, List<NameValuePairString> hiddenFormFields, String formSelector) {
// --- analisi della pagina contente la form, specifica al sito
Document doc = rw.getJSoupDocument();
Elements els = doc.select(formSelector); // per debug, dovrebbe essere uo
if (els == null || els.size() <= 0) {
log.error("unable to find form at selector: " + formSelector);
System.exit(1);
return false;
}
Element loginForm = els.get(0);
if (loginForm == null) {
log.error("failed to get form to analyze at: " + rw.dump());
System.exit(1);
}
// log.info("login form OUTER HTML\n" + loginForm.outerHtml());
Elements inputFields = loginForm.select("input");
// display all
for (Element e : inputFields) {
String type = e.attr("type");
if (type.equals("submit")) {
continue;
}
String attrName = e.attr("name");
hiddenFormFields.add(new NameValuePairString(attrName, e.val()));
log.debug("captured form input: " + attrName + " = " + e.val());
}
return false;
}
public static boolean isSubPathOf(String sub, String path, boolean acceptEqual) {
if (!(isPathCorrect(sub) && isPathCorrect(path))) {
log.error("exiting invalid path(s): " + sub + " altro " + path);
isPathCorrect(path);
isPathCorrect(sub);
System.exit(1);
return false;
}
if (path.endsWith("/") && path.length() > 1) path = path.substring(0, path.length() - 1);
if (sub.endsWith("/") && sub.length() > 1) sub = sub.substring(0, sub.length() - 1);
// da qui path corretti ed eventualmente normalizzati
if (acceptEqual) {
if (sub.equals(path)) return true;
}
if (sub.length() <= path.length()) return false;
String[] subElems = sub.split("/");
String[] pathElems = path.split("/");
for (int i = 0; i < subElems.length && i < pathElems.length; i++) {
if (!subElems[i].equals(pathElems[i])) return false;
}
return true;
}
public static boolean isSubDomainOf(String sub, String dom, boolean acceptEqual) {
if (!isDomainCorrect(sub) || !isDomainCorrect(dom)) {
log.error("exiting invalid domain(s): " + dom + " altro " + sub);
System.exit(1);
}
if (acceptEqual && sub.equals(dom)) return true;
if (!sub.endsWith(dom)) return false;
boolean ret = sub.matches(subDomainLeftRegex + dom);
return ret;
}
public static Result textSelList(Element elsPar, ArrayList<String> jsoupSelectors) {
Result res = new Result();
Elements els;
Elements elsTemp;
els = elsPar.select(jsoupSelectors.remove(0));
for (String sel : jsoupSelectors) {
elsTemp = els.select(sel);
if (elsTemp != null) els = elsTemp;
}
if (els == null || els.size() != 1) {
log.error("jsoup selector on elements does not match 1");
System.exit(1);
return res.setContinua(false);
}
return res.setRetStr(els.get(0).text());
}
public static boolean isRelativeURL(String url) {
boolean ret = false;
if (!WEBUtils.isCorrectURL(url)) {
if (!WEBUtils.isCorrectURL("http://google.com" + url)) {
// rifiuta URL relative come /a
log.error("esco, not a correct URL: " + url);
System.exit(1);
}
}
if (url.contains("http:") || url.contains("https:")) {
return false;
}
// probably wrong
// check if it contains an hostname
return !url.matches("[\\w]+\\.?[\\w]+/[^/].*");
}