Does not clean URL on Google's sites

[RepoMoveBot]

Issue by hayohayo
Tuesday Jul 03, 2018 at 18:35 GMT

---

Neat URL does not delete query parameters from links on Google. The simplest URL that reproduces my issue is: https://www.google.com/?utm_source=x. utm_source is removed just fine on other domains.

Add-on: Neat URL 4.1.5
OS: MacOS High Sierra
Browser: Firefox 61.0

[RepoMoveBot]

Comment by hayohayo
Tuesday Jul 03, 2018 at 18:46 GMT

---

I believe, it has something to do with domain-specific blocked parameters. Once I've removed ved@google.*, ei@google.*, sei@google.*, gws_rd@google.*, utm_source is properly removed.

[RepoMoveBot]

Comment by crssi
Tuesday Jul 03, 2018 at 20:22 GMT

---

It must be different on MacOS then on Windows 10, since here is working fine.
This is from browser console:
[Neat URL]: (type main_frame): 'https://www.google.com/?utm_source=x' has been changed to 'https://www.google.com/'

[RepoMoveBot]

Comment by hayohayo
Tuesday Jul 03, 2018 at 20:39 GMT

---

Hmm... have just tried on Debian - the same here. Works for, say, amazon.com, but not on google.com. Also, not sure how to see the logs - I don't see them in browser's console, even though I enabled them in add-on's settings.

OS: Linux, Debian 9.4
Browser: Firefox ESR 52.9.0

[my-wan]

Had the same issue on Lubuntu 17.10 (32 bit) LXDE with Firefox 61.0. Removing gws_rd@google.* by itself from the blocked parameters fixed the issue. However, setting gws_rd as a global block does not cause any problems, only when given as a Google specific block. This appears related to the workaround for issue# 76 on line 288 of background.js. The only thing that's returning true even when it apparently shouldn't is detailsUrl.contains("gws_rd=cr"). There is no gws_rd parameter in the Google URLS being skipped. I haven't got deep enough to know why yet. But this appears to be where the getMatch function is returning prematurely and skipping the Google domain when it shouldn't.

[tartpvule]

Duplicate: #146

This is a bug in 53eb44b, specifically the one in background.js supposedly for fixing #76, like @my-wan said.
But the problem is not that it returns true, it is that throws detailsUrl.contains is not a function.

In background.js line 501, let match = getMatch(gbp, domain, rootDomain, domainMinusSuffix, url);.
"url" here is a URL object, not a string.

The end result is that the presence of "gws_rd@google.*" in "Blocked parameters" causes Neat URL to not process google.com URLs altogether.

Please fix this.