Further update on this. Emacspeak has built in support to masquerade as a different browser. The variable emacspeak-eww-masquerade is true when masquerading is on and the function emacspeak-eww-masquerade can be used to toggle masquerading on and off. When masquerading is on, which I think is the default, the variable emacspeak-eww-masquerade-as is used to set the User-Agent header. So, my guess is that if we find the correct value for the user agent and set emacspeak-eww-masquerade-as to that value, google search will work again. However, Raman does tend to make a fair bit of use of XSLT transformations to massage results into a more accessible structure. It is possible that when we modify the user-agent string that this will modify the structure of the returned document and cause issues with the XSLT parsing, so there could be some further knock on side effects that will need to be diagnosed and fixed once we do get google to respond with data. Tim "peter.julien.rayner" (via emacspeak Mailing List) <emacspeak@xxxxxxxxxxxxx> writes: > Has anyone else had google searches stop working in the last few days? > The sequence > ctrl-e <ret> testing google search<ret> > yields > Summary: 0 Nodes Matching //\*\[\(\@id\="main"\)\] in document. > > the same search returns results when run from chrome. > This happened about the same time I was configuring rclone and getting > a client id for rclone but that is probably coincidence. More likely > google just changed format slightly so the regexp isn't working. > Before I dig can someone else let me know if they can reproduce? This > is with emacs 29.3 under ubuntu 24.04 and using emacspeak master from > github. > cheers > Peter
|Full archive May 1995 - present by Year|Search the archive|
If you have questions about this archive or had problems using it, please contact us.