Set your pref to send email in rich text html. Get a message that has a paragraph of html Press reply Note that message is quoted, with blue bar at left (HTML quoting) Using mouse, place curson in the middle of a line of text, around the middle of a paragraph Press enter (new line) Note that quoting splits the sentence in the middle, and places cursor at left margin. Type two key strokes, such as 12 Note that the characters appear where you typed them. Press a third key, such as 3 Note that the 3 appears after the remainder of the quoted paragraph, several lines later in the message The expected result was of course that the 3 would appear just after the 1 and 2. This regression just happened in yesterday's build. It is present in the afternoon build for 3/8/2000 I don't know if it is too late to get a fix in for beta... but having a fix in hand will certainly help to have a shot at making this change for beta1.
Nominating for beta1... just in case it can make it :-/
Tested on commercial Mac bits from yesterday afternoon (2pm). I don't see this problem on Mac. Can anyone else reproduce this problem? Since it is not a crasher, I recommend that it be PDT-.
I'm able to reproduce this on a release build from today on linux. For what it's worth, after this triggers I get it to happend again if I click on a line *above* the one I clicked on (and press return and start typing), but not if I click on a line below the one I clicked on the first time.
strange: to my knowledge no related editor code has changed across the last few days. I wonder if this can happen with earlier builds?
I don't see this, on the commercial 3/9 linux build. I wonder if this could be some sort of timing related problem? Not that I can think of anything that *should* be timing related here ...
Oh, cool. A few minutes after I added that comment, replying to chofmann's "things that block branching" text/plain message with a multipart/alternative attachment, hitting return at the end of one of the plaintext lines, I saw the behavior: on the third character, just as described, the caret jumped to the end of the message (but before the attachment) and split the blockquote there. I can reproduce this every time on this particular message. No immediate clue why it's happening. Joe, any suggestions on where I should set breakpoints or put printfs?
i've been able to reproduce this on akkana's mac, and get some initial insight there. I'm going to try to debug it further on my machine. stay tuned...
Putting on PDT- radar for beta1.
I was able to reproduce this on a Win32 build yesterday as per jar's instructions.
This is actually a "caret" issue (not a cursor bug).