User Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:51.0) Gecko/20100101 Firefox/51.0 Build ID: 20170129012138 Steps to reproduce: I enter a keyword (shortcut key) in the address bar to chose one of my pre-defined search engines. Anything I type in after that is transmitted to that search engine. Actual results: When I enter "w something" or "W something" into the address bar, I am taken to the German Wikipedia page. Expected results: For years, I have had the same key(word) in uppercase and lower case for the international and the German version of a specific web site, e.g. w - Wikipedia (German) W - Wikipedia (International) a - amazon.de A - Amazon.com This was very handy, as I did not have to remember different keys for the same service. All of the sudden this has stopped and w and W as well as a and A lead me to the same search page.

Another side effect of bug 1306639: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=64cd7d87e78cf2e4b919aad10fb9f4962834b004&tochange=70d19ed29c980ba3d6825e5286bda08fa7dc676f

It was a confusing feature. I can understand that you were relying on that, but for most users it would be confusing if "w something" works and "W something" doesn't, when they only set one of those. So, imho, that was a bug fix, not a regression, and this is a wontfix. Btw, I'm asking Florian, as Search module owner, if he thinks search aliases should be case sensitive or not, also for future reference when we'll start unifying the features. For bookmark keywords my answer is that the same keyword should point to the same url, regardless of casing.

Case insensitive keywords are less error prone, so I agree with Marco that wontfix makes sense here.

Hi Marco and Florian, I understand your argument and even if it is a regression for me, I agree that for most people it is rather the opposite. Just one idea: Would it be very difficult to build in a check, that verifies whether or not the same key was actually used twice? That way, if only "w" was attached to a search site, "W" would also work. But if "w" and "W" were specified for different sites, it would honor the case.

(In reply to Uwe Trenkner from comment #4) > Just one idea: Would it be very difficult to build in a check, that verifies > whether or not the same key was actually used twice? That way, if only "w" > was attached to a search site, "W" would also work. But if "w" and "W" were > specified for different sites, it would honor the case. It would be possible, and I thought about it already. The problem is that the fact the UI allows to set cased keywords on different engines is a bug by itself that we should fix (I will file it now). So, even if we add such a special casing, when we actually fix the UI to disallow cased aliases, that special code becomes pointless.

I filed bug 1336395 to fix the UI, atm I don't think we plan to add special code to handle this case, I'm sorry if this is not satisfying your needs, unfortunately decisions always tend to leave someone upset :\

I do not quite see why allowing to set cased keywords should be considered a UI bug. Case sensitive shortcut keys are quite common in standard software products - even Firefox uses it: "Ctrl+z" is for "undo" and "Ctrl+Z" is for "redo". It does not hurt to allow it and on the contrary, I am sure you will hurt other long-time users of this feature, too. For other users the above discussed safety net could be provided.

(In reply to Uwe Trenkner from comment #7) > I do not quite see why allowing to set cased keywords should be considered a > UI bug. Case sensitive shortcut keys are quite common in standard software > products - even Firefox uses it: "Ctrl+z" is for "undo" and "Ctrl+Z" is for > "redo". The latter is actually CTRL+SHIFT+z, it's quite different. To make a proper comparison, on your keyboard you should have both z and Z keys, close to each other, ctrl+z would do a different thing from ctrl+Z. For most people it would be confusing. > It does not hurt to allow it and on the contrary, I am sure you will hurt > other long-time users of this feature, too. For other users the above > discussed safety net could be provided. Yes, we are aware this may break some edge workflows, I'm sorry about that.

Sorry, not trolling: Isn't for the user "Ctrl+Shift+z" exactly the same as "Ctrl+Z" - the upper case Z implying the shift key? If it is different, then maybe we could allow this difference also for search? In my case it would be: "W" for Wikipedia (German) "Shift+W" for Wikipedia (International)

(In reply to Uwe Trenkner from comment #9) > Isn't for the user "Ctrl+Shift+z" exactly the same as > "Ctrl+Z" - the upper case Z implying the shift key? No, shift is a specific key on the keyboard. The case difference could be due to eg. Caps lock, so no it's not the same.

Thank you for the explanation! And for your patience! Good luck with the changes!