User:GijsKruitbosch/JS Debugging/Tasks: Difference between revisions
(Adding venkman fix) |
|||
| Line 70: | Line 70: | ||
<tr> | <tr> | ||
<td>8</td> | <td>8</td> | ||
<td>Venkman should try to load line-maps lazily (on-demand, and not demand them immediately like it does now).</td> | |||
<td>No</td> | |||
<td>-</td> | |||
<td>None yet</td> | |||
</tr> | |||
<tr> | |||
<td>9</td> | |||
<td>Changing variable values / evaluating js shouldn't require changing tabs away from the debugger (in Firebug)</td> | <td>Changing variable values / evaluating js shouldn't require changing tabs away from the debugger (in Firebug)</td> | ||
<td>No</td> | <td>No</td> | ||
| Line 77: | Line 85: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>10</td> | ||
<td>Evaluate js in different windows/tabs/iframes</td> | <td>Evaluate js in different windows/tabs/iframes</td> | ||
| Line 85: | Line 93: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>11</td> | ||
<td>Provide immediate access to profiling data</td> | <td>Provide immediate access to profiling data</td> | ||
| Line 93: | Line 101: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>12</td> | ||
<td>Anonymous functions should not be lumped together in profiling.</td> | <td>Anonymous functions should not be lumped together in profiling.</td> | ||
<td>Presumably not</td> | <td>Presumably not</td> | ||
| Line 101: | Line 109: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>13</td> | ||
<td>Double-check handling of about:blank and (related) running/stopped state tracking in Firebug.</td> | <td>Double-check handling of about:blank and (related) running/stopped state tracking in Firebug.</td> | ||
<td>No</td> | <td>No</td> | ||
| Line 111: | Line 119: | ||
<tr> | <tr> | ||
<td> | <td>14</td> | ||
<td>Figure out why I'm crashing on GMail with firebug enabled (hack/sneak/workaround?)</td> | <td>Figure out why I'm crashing on GMail with firebug enabled (hack/sneak/workaround?)</td> | ||
| Line 119: | Line 127: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>15</td> | ||
<td>Fix wrapping of variable/value tree</td> | <td>Fix wrapping of variable/value tree</td> | ||
| Line 127: | Line 135: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>16</td> | ||
<td>JS 1.7 improvements</td> | <td>JS 1.7 improvements</td> | ||
<td>Yes?</td> | <td>Yes?</td> | ||
| Line 135: | Line 143: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>17</td> | ||
<td>Debug mutation events (watch DOM elements)</td> | <td>Debug mutation events (watch DOM elements)</td> | ||
| Line 143: | Line 151: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>18</td> | ||
<td>Getters / Setters improvements</td> | <td>Getters / Setters improvements</td> | ||
| Line 151: | Line 159: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>19</td> | ||
<td>'real' stack list (not a dropdown)</td> | <td>'real' stack list (not a dropdown)</td> | ||
| Line 159: | Line 167: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>20</td> | ||
<td>Inline editing of functions etc.</td> | <td>Inline editing of functions etc.</td> | ||
| Line 167: | Line 175: | ||
</tr> | </tr> | ||
<tr> | <tr> | ||
<td> | <td>21</td> | ||
<td>Pretty print</td> | <td>Pretty print</td> | ||
Revision as of 14:35, 16 June 2006
Importance-ordered Task List
A small note about this list: importance != working order, given that some of these items need other items, and that some require back-end changes which will need to happen fast, if at all, on the 1.8.1 branch (which is what I'm trying to cater to).
Front end work is being done on: Firebug
| N | Task | Needs backend changes | Bug number | Result |
| 1 | Implement list of Breakpoints | No | - | None yet |
| 2 | Implement Watches | No | - | None yet |
| 3 | Make the originated document accessible from any error in the console | No? | bug 228205 | None yet |
| 4 | Frontend adapted to keep breakpoints and such separate on a per-document basis (for Firebug's console) (see above) | No? | bug 228205 | None yet |
| 5 | Be able to go from a var referencing a DOM element to that DOM element in DOMI | No? | - | None yet |
| 6 | Be able to debug from the slow script warning dialog. | Yes? | - | None yet |
| 7 | Newlines are lost in the variable/value Firebug view. | No | - | None yet |
| 8 | Venkman should try to load line-maps lazily (on-demand, and not demand them immediately like it does now). | No | - | None yet |
| 9 | Changing variable values / evaluating js shouldn't require changing tabs away from the debugger (in Firebug) | No | - | None yet |
| 10 | Evaluate js in different windows/tabs/iframes | No | - | None yet |
| 11 | Provide immediate access to profiling data | No | - | None yet |
| 12 | Anonymous functions should not be lumped together in profiling. | Presumably not | - | None yet |
| 13 | Double-check handling of about:blank and (related) running/stopped state tracking in Firebug. | No | - | None yet |
| 14 | Figure out why I'm crashing on GMail with firebug enabled (hack/sneak/workaround?) | Hopefully not. | - | None yet |
| 15 | Fix wrapping of variable/value tree | No | - | None yet |
| 16 | JS 1.7 improvements | Yes? | - | None yet |
| 17 | Debug mutation events (watch DOM elements) | Yes | - | None yet |
| 18 | Getters / Setters improvements | Yes? | - | None yet |
| 19 | 'real' stack list (not a dropdown) | No | - | None yet |
| 20 | Inline editing of functions etc. | No, hopefully. | - | None yet |
| 21 | Pretty print | No | - | None yet |
Getting to the document that error X originated from
Shaver should write this sometime (see below)
Discussion
I would probably order these things higher:
- immediate profiling
- newlines in watches view (but there's no watch support yet, so?)
- eval JS in tab/frame/etc.
And these things lower:
- JS 1.7
- inline function editing
But I think you're generally on the right track. I need to write here how we can track what document an error came from without biting off everything that's in the current bug 228205, too. Basically involves finding the current document (via cx->contextPrivate->scriptOwner->frameElement or some such) from the handler called for exceptions and errors, and keeping track of it. We could also "just" have the console error object QI to nsIException, but I bet that would make some people API-sad, so I'm interested in us keeping the hacks to our side of the line if we can.
-- shaver
- Done, I think. GijsKruitbosch 07:20, 16 June 2006 (PDT)