Minecraft
  1. Minecraft
  2. MC-9582

Whole-word selection behaves erratically in chat

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Resolution: Fixed
    • Affects Version/s: Minecraft 1.4.7, Snapshot 13w06a, Minecraft 1.5, Snapshot 13w11a, Minecraft 1.7.4
    • Fix Version/s: Minecraft 14w04a
    • Security Level: Minecraft - Public (Viewable by everyone)
    • Labels:
    • Confirmation Status:
      Confirmed

      Description

      Trying to select a word using ctrl+shift+left or ctrl+shift+right will work for one word only. It appears to attempt to match the selection starting from the wrong selection anchor.
      Two otherwise extremely useful key combinations, "home, shift+[end, ctrl+left]" and "end, shift+[home, ctrl+right]" are both affected adversely by this.

        Activity

        Hide
        [Mod] Kumasasa added a comment -

        Confirmed (Windows 7)

        Show
        [Mod] Kumasasa added a comment - Confirmed (Windows 7)
        Hide
        Score_Under the Human added a comment -

        For the record (and for mojang): in the textbox class, the (overloaded) function used to find a word given a position and offset... ignores the offset. Looks like a copy/paste gone wrong.

        Show
        Score_Under the Human added a comment - For the record (and for mojang): in the textbox class, the (overloaded) function used to find a word given a position and offset... ignores the offset. Looks like a copy/paste gone wrong.
        Hide
        [Mod] Kumasasa added a comment -

        Confirmed in 1.5

        Show
        [Mod] Kumasasa added a comment - Confirmed in 1.5
        Hide
        Score_Under the Human added a comment - - edited

        Just so that it's 100% obvious exactly where the bug is and why it happens (added a picture)

        Show
        Score_Under the Human added a comment - - edited Just so that it's 100% obvious exactly where the bug is and why it happens (added a picture)
        Hide
        [Mod] Tails added a comment -

        Is this still a concern in the current Minecraft version 1.7.4 / Launcher version 1.3.7 or later? If so, please update the affected versions in order to best aid Mojang ensuring bugs are still valid in the latest releases/pre-releases.

        Show
        [Mod] Tails added a comment - Is this still a concern in the current Minecraft version 1.7.4 / Launcher version 1.3.7 or later? If so, please update the affected versions in order to best aid Mojang ensuring bugs are still valid in the latest releases/pre-releases.
        Hide
        Score_Under the Human added a comment -

        Added the current version, just remembered to actually test it. This bug has been present since the inception of the new chat system around version 1.2.5 IIRC.

        If only Mojang read bug reports... it is frustrating to know exactly the mistake made and exactly how to fix it in less than ten keystrokes in the IDE, and to report where the problem is exactly in the code, only to have this ignored for just shy of a year. It's not a matter of not knowing the solution, it's a matter of simply not reading this report.

        Show
        Score_Under the Human added a comment - Added the current version, just remembered to actually test it. This bug has been present since the inception of the new chat system around version 1.2.5 IIRC. If only Mojang read bug reports... it is frustrating to know exactly the mistake made and exactly how to fix it in less than ten keystrokes in the IDE, and to report where the problem is exactly in the code, only to have this ignored for just shy of a year. It's not a matter of not knowing the solution, it's a matter of simply not reading this report.
        Hide
        Torabi added a comment -

        They do read bug reports, because the various developers occasionally comment, assign themselves to bugs, fix the bugs, and mark them fixed. However, there are over 2000 open reports, and over 40000 total reports on the tracker. Most of which are garbage, either because they're duplicates, incomprehensible, or not actually bugs. It's a lot of work to sift through it all looking for valid issues to address. And despite how many employees Mojang now has, very few of them are actively developing Minecraft. Something like 5, now that [Mojang] Mog (Ryan Holtz) and [Mojang] Searge (Michael Stoyke) have joined the team.

        The vote count also figures into putting priority on fixing bugs. This one is also obviously not an issue for many people, as there are only two votes (one of them's mine!), and zero duplicate reports. The thing is, no matter how quick and easy it would be to fix something, there's still hundreds, if not thousands, of other things all demanding the same time and attention. Fixing this would mean something else gets put off, even if it's only by 5 minutes. With so many people filing reports, some of them are bound to be put off, no matter how fast they work.

        Also consider that fixes based on decompiled code are useless to the Mojang developers, because it's nothing like the actual source code. I'm kind of surprised this bug survived the big code review, however.

        Show
        Torabi added a comment - They do read bug reports, because the various developers occasionally comment, assign themselves to bugs, fix the bugs, and mark them fixed. However, there are over 2000 open reports , and over 40000 total reports on the tracker. Most of which are garbage, either because they're duplicates, incomprehensible, or not actually bugs. It's a lot of work to sift through it all looking for valid issues to address. And despite how many employees Mojang now has, very few of them are actively developing Minecraft. Something like 5, now that [Mojang] Mog (Ryan Holtz) and [Mojang] Searge (Michael Stoyke) have joined the team. The vote count also figures into putting priority on fixing bugs. This one is also obviously not an issue for many people, as there are only two votes (one of them's mine!), and zero duplicate reports. The thing is, no matter how quick and easy it would be to fix something, there's still hundreds, if not thousands, of other things all demanding the same time and attention. Fixing this would mean something else gets put off, even if it's only by 5 minutes. With so many people filing reports, some of them are bound to be put off, no matter how fast they work. Also consider that fixes based on decompiled code are useless to the Mojang developers, because it's nothing like the actual source code . I'm kind of surprised this bug survived the big code review, however.
        Hide
        Score_Under the Human added a comment -

        Right, I'm aware they do actually read reports, it's just generally frustrating that bugs like this go ignored for such long periods of time.

        As for the "nothing like the actual source code" claim, I'm not so sure. I know it'd be nothing like the source code visually, but when it comes to the structure it would be nearly the same, and recognisably so. Things like this broken delegation to another method would be almost exactly as they were in the original source code, but with some names changed. A good starting point would be to run IntelliJ's "unused parameter" inspection on classes handling textbox input, and assuming nothing out of the ordinary is going on here it would find the bug instantly (or rather, narrow it down to one of maybe 5 lines of code, from which it would be easy for a human to determine where the problem lies).

        Show
        Score_Under the Human added a comment - Right, I'm aware they do actually read reports, it's just generally frustrating that bugs like this go ignored for such long periods of time. As for the "nothing like the actual source code" claim, I'm not so sure. I know it'd be nothing like the source code visually, but when it comes to the structure it would be nearly the same, and recognisably so. Things like this broken delegation to another method would be almost exactly as they were in the original source code, but with some names changed. A good starting point would be to run IntelliJ's "unused parameter" inspection on classes handling textbox input, and assuming nothing out of the ordinary is going on here it would find the bug instantly (or rather, narrow it down to one of maybe 5 lines of code, from which it would be easy for a human to determine where the problem lies).
        Hide
        Torabi added a comment -

        Do you have any experience with obfuscation? The whole point is to mangle the code so that it's unrecognizable, and yet still functions. Compiler optimizations do much the same as a side-effect, with the actual goal being efficiency. So no, the structure of the code isn't necessarily all that similar. Dinnerbone has said that it's often more work to understand someone else's fix against the MCP codebase than to just fix it themselves. It's true that this is a very simple bug, and that your code snippet shows that, but it's really not relevant to it actually getting fixed. They have bigger fish to fry. Just because you (and me!) care about this doesn't make it important. Most people are not all that devoted to bug-free code (See "worse is better"). Once again, I'm surprised this didn't get fixed in the big code review, as something like an unused parameter search would be an obvious thing to run, and would have highlighted this problem.

        Show
        Torabi added a comment - Do you have any experience with obfuscation? The whole point is to mangle the code so that it's unrecognizable, and yet still functions. Compiler optimizations do much the same as a side-effect, with the actual goal being efficiency. So no, the structure of the code isn't necessarily all that similar. Dinnerbone has said that it's often more work to understand someone else's fix against the MCP codebase than to just fix it themselves. It's true that this is a very simple bug, and that your code snippet shows that, but it's really not relevant to it actually getting fixed. They have bigger fish to fry. Just because you (and me!) care about this doesn't make it important. Most people are not all that devoted to bug-free code (See "worse is better"). Once again, I'm surprised this didn't get fixed in the big code review, as something like an unused parameter search would be an obvious thing to run, and would have highlighted this problem.

          People

          • Assignee:
            [Mojang] Dinnerbone (Nathan Adams)
            Reporter:
            Score_Under the Human
          • Votes:
            2 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:
              CHK: