I’m seeing occasional really annoying behaviour in the code pane of the debugger. It’s hard to pin down quite what might triggers it, so I’m not about to attempt to find an easy recipe right now.
Problem: sometimes the code pane gets into a state whereby no matter where you try to select - in order to printIt on a snippet of code, for example - it will jump the selection back to some other place. Right now for example I’m looking at startDrag: evt | rootForGrab | self break. rootForGrab := owner rootForGrabOf: self. rootForGrab ifNil: [^ self]. rootForGrab position: evt hand position + (self topLeft - evt cursorPoint). evt hand grabMorph: rootForGrab and if I try to select ‘evt cursorPoint’ to cmd-p, the selection jumps back to ‘hand’. It looks like it might be something to do with the #step code since if I’m really quick I can do cmd-a,cmd-c and copy the entire method (as I did above). Switching to another method in the stack and then back seems to make no difference. I’m not at all certain, but there is some possibility it only happens with methods which have the toggle-break on. They are after all likely to be methods I look at in the debugger. I’m not using a totally up to date image here (update #13297) so it’s possible the problem was fixed before the 4.5 final. I can’t do anything much to try this out in a final 4.5 image right now because I simply can’t spare the time on something so intermittent. I hate giving such vague bug reports but when the problem is sufficiently rare you can’t do much else. tim -- tim Rowledge; [hidden email]; http://www.rowledge.org/tim Strange OpCodes: SEXI: Sign EXtend Integer |
I'm not sure if I have seen the problem you are describing, but I just discovered
that I am getting a variety of wierd symptoms when I try to edit a class comment. Could this be related? Dave On Mon, Mar 31, 2014 at 05:11:01PM -0700, tim Rowledge wrote: > I?m seeing occasional really annoying behaviour in the code pane of the debugger. It?s hard to pin down quite what might triggers it, so I?m not about to attempt to find an easy recipe right now. > > Problem: sometimes the code pane gets into a state whereby no matter where you try to select - in order to printIt on a snippet of code, for example - it will jump the selection back to some other place. Right now for example I?m looking at > startDrag: evt > | rootForGrab | > self break. > rootForGrab := owner rootForGrabOf: self. > rootForGrab > ifNil: [^ self]. > rootForGrab position: evt hand position + (self topLeft - evt cursorPoint). > evt hand grabMorph: rootForGrab > and if I try to select ?evt cursorPoint? to cmd-p, the selection jumps back to ?hand?. It looks like it might be something to do with the #step code since if I?m really quick I can do cmd-a,cmd-c and copy the entire method (as I did above). > > Switching to another method in the stack and then back seems to make no difference. I?m not at all certain, but there is some possibility it only happens with methods which have the toggle-break on. They are after all likely to be methods I look at in the debugger. > > I?m not using a totally up to date image here (update #13297) so it?s possible the problem was fixed before the 4.5 final. I can?t do anything much to try this out in a final 4.5 image right now because I simply can?t spare the time on something so intermittent. > > I hate giving such vague bug reports but when the problem is sufficiently rare you can?t do much else. > > tim > -- > tim Rowledge; [hidden email]; http://www.rowledge.org/tim > Strange OpCodes: SEXI: Sign EXtend Integer > > |
In reply to this post by timrowledge
On 1 April 2014 01:11, tim Rowledge <[hidden email]> wrote:
> I’m seeing occasional really annoying behaviour in the code pane of the debugger. It’s hard to pin down quite what might triggers it, so I’m not about to attempt to find an easy recipe right now. > > Problem: sometimes the code pane gets into a state whereby no matter where you try to select - in order to printIt on a snippet of code, for example - it will jump the selection back to some other place. Right now for example I’m looking at > startDrag: evt > | rootForGrab | > self break. > rootForGrab := owner rootForGrabOf: self. > rootForGrab > ifNil: [^ self]. > rootForGrab position: evt hand position + (self topLeft - evt cursorPoint). > evt hand grabMorph: rootForGrab > and if I try to select ‘evt cursorPoint’ to cmd-p, the selection jumps back to ‘hand’. It looks like it might be something to do with the #step code since if I’m really quick I can do cmd-a,cmd-c and copy the entire method (as I did above). > > Switching to another method in the stack and then back seems to make no difference. I’m not at all certain, but there is some possibility it only happens with methods which have the toggle-break on. They are after all likely to be methods I look at in the debugger. > > I’m not using a totally up to date image here (update #13297) so it’s possible the problem was fixed before the 4.5 final. I can’t do anything much to try this out in a final 4.5 image right now because I simply can’t spare the time on something so intermittent. > > I hate giving such vague bug reports but when the problem is sufficiently rare you can’t do much else. I can only confirm that the bug exists, and has for a while (maybe at least a year?). I've seen it rarely. frank > tim > -- > tim Rowledge; [hidden email]; http://www.rowledge.org/tim > Strange OpCodes: SEXI: Sign EXtend Integer |
On 01-04-2014, at 5:59 AM, Frank Shearar <[hidden email]> wrote: >> >> I hate giving such vague bug reports but when the problem is sufficiently rare you can’t do much else. > > I can only confirm that the bug exists, and has for a while (maybe at > least a year?). I've seen it rarely. well at least it means I’m not going mad(der). tim -- tim Rowledge; [hidden email]; http://www.rowledge.org/tim Disclaimer: Any errors in spelling, tact, or fact are transmission errors. |
In reply to this post by timrowledge
No, I know exactly what you're talking about Tim. And, yes, it's an
impediment for working in the debugger, I hope we can identify a fix. I brought this up before: http://lists.squeakfoundation.org/pipermail/squeak-dev/2013-September/173549.html >From my vague memory, I started seeing this after there was a fix put in place to this problem: ---> That when the code was changed in the debugger and the method saved, it would not highlight the new method correctly when stepping. That problem was fixed, but I *think* it's what introduced this one.. On Mon, Mar 31, 2014 at 7:11 PM, tim Rowledge <[hidden email]> wrote: > I’m seeing occasional really annoying behaviour in the code pane of the debugger. It’s hard to pin down quite what might triggers it, so I’m not about to attempt to find an easy recipe right now. > > Problem: sometimes the code pane gets into a state whereby no matter where you try to select - in order to printIt on a snippet of code, for example - it will jump the selection back to some other place. Right now for example I’m looking at > startDrag: evt > | rootForGrab | > self break. > rootForGrab := owner rootForGrabOf: self. > rootForGrab > ifNil: [^ self]. > rootForGrab position: evt hand position + (self topLeft - evt cursorPoint). > evt hand grabMorph: rootForGrab > and if I try to select ‘evt cursorPoint’ to cmd-p, the selection jumps back to ‘hand’. It looks like it might be something to do with the #step code since if I’m really quick I can do cmd-a,cmd-c and copy the entire method (as I did above). > > Switching to another method in the stack and then back seems to make no difference. I’m not at all certain, but there is some possibility it only happens with methods which have the toggle-break on. They are after all likely to be methods I look at in the debugger. > > I’m not using a totally up to date image here (update #13297) so it’s possible the problem was fixed before the 4.5 final. I can’t do anything much to try this out in a final 4.5 image right now because I simply can’t spare the time on something so intermittent. > > I hate giving such vague bug reports but when the problem is sufficiently rare you can’t do much else. > > tim > -- > tim Rowledge; [hidden email]; http://www.rowledge.org/tim > Strange OpCodes: SEXI: Sign EXtend Integer > > > |
Free forum by Nabble | Edit this page |