FixBrowser

Forum: General Discussion: fixproxy Google search support

f
2025/04/25 09:06 When doing a search on google.com I am getting the "Outside link" dialog, and no clicking on any of the 2 Yes links brings me to search results.
The first line says:Confirm going to URL:
https://consent.google.com/ml?continue=https://www.google.com/search%3 ...
Using Image search works from google.com/imghp address.

jezek2
2025/04/28 18:41 Fixed in the latest developer version on GitHub.

f
2025/04/29 17:39 I was able to compile using 32-bit version of Haiku only when replacing all the gcc and strip commands with the gcc-x86 and strip-x86 in the compile_haiku.sh
Thanks for adding support for Google search, still missing is displaying of images for Video search that are on the "i.ytimg.com" server.

jezek2
2025/04/29 20:32 Both fixed. The proper way is to use the "setarch x86" command to switch to a newer GCC version. I've added automatic switching to the compile script.

f
2025/10/14 13:22 Thanks for providing the brief Google search support. Now I know that, at the beggining of 2025 the mechanism for serving results was changed to not provide results in the page but generate them with JavaScript. Changing the UserAgent for browser worked for some time and it does not work anymore.

jezek2
2025/10/14 13:54 Yeah it was inevitable. I will add support using the GateOpener service.

Users who haven't donated (and therefore don't have access to the service) will have to use an alternative search engine (DuckDuckGo).

There is also a possibility of including a JavaScript runtime, but I have faced this before when dealing with Cloudflare and other gatekeepers, that it's too easy for the websites to require a full web browser anyway. Also the various smaller JS runtimes seem to lag behind in features.

I will also look if there is a possibility to reimplement the client-side logic without relying on JS execution from the server. This depends on if Google has implemented any kind of protection like it does for YouTube downloads (but not for the website itself).

jezek2
2025/10/14 15:11 Found another User Agent string that still works, the fix is available on GitHub.

f
2025/10/25 19:38 I verified that setting this User Agent string is returning the results for Images, however when I try to change to Google Images tab I have the incompatible browser error. Is it possible that this user agent is somehow lost when I try image search (for example due to certificate or domain being different) in fixproxy?

Reply

The website was designed for modern browsers and IE4+.

Name:
Content:
Confirmation code:

	(type the letters shown in the picture)

f 2025/04/25 09:06	When doing a search on google.com I am getting the "Outside link" dialog, and no clicking on any of the 2 Yes links brings me to search results. The first line says:Confirm going to URL: https://consent.google.com/ml?continue=https://www.google.com/search%3 ... Using Image search works from google.com/imghp address.
jezek2 2025/04/28 18:41	Fixed in the latest developer version on GitHub.
f 2025/04/29 17:39	I was able to compile using 32-bit version of Haiku only when replacing all the gcc and strip commands with the gcc-x86 and strip-x86 in the compile_haiku.sh Thanks for adding support for Google search, still missing is displaying of images for Video search that are on the "i.ytimg.com" server.
jezek2 2025/04/29 20:32	Both fixed. The proper way is to use the "setarch x86" command to switch to a newer GCC version. I've added automatic switching to the compile script.
f 2025/10/14 13:22	Thanks for providing the brief Google search support. Now I know that, at the beggining of 2025 the mechanism for serving results was changed to not provide results in the page but generate them with JavaScript. Changing the UserAgent for browser worked for some time and it does not work anymore.
jezek2 2025/10/14 13:54	Yeah it was inevitable. I will add support using the GateOpener service. Users who haven't donated (and therefore don't have access to the service) will have to use an alternative search engine (DuckDuckGo). There is also a possibility of including a JavaScript runtime, but I have faced this before when dealing with Cloudflare and other gatekeepers, that it's too easy for the websites to require a full web browser anyway. Also the various smaller JS runtimes seem to lag behind in features. I will also look if there is a possibility to reimplement the client-side logic without relying on JS execution from the server. This depends on if Google has implemented any kind of protection like it does for YouTube downloads (but not for the website itself).
jezek2 2025/10/14 15:11	Found another User Agent string that still works, the fix is available on GitHub.
f 2025/10/25 19:38	I verified that setting this User Agent string is returning the results for Images, however when I try to change to Google Images tab I have the incompatible browser error. Is it possible that this user agent is somehow lost when I try image search (for example due to certificate or domain being different) in fixproxy?