Bug report #17181
Crash using QgsComposerAttributeTableV2 after setting columns
| Status: | Closed | ||
|---|---|---|---|
| Priority: | Normal | ||
| Assignee: | - | ||
| Category: | Python bindings / sipify | ||
| Affected QGIS version: | 2.18.13 | Regression?: | No |
| Operating System: | Easy fix?: | No | |
| Pull Request or Patch supplied: | No | Resolution: | invalid |
| Crashes QGIS or corrupts data: | Yes | Copied to github as #: | 25080 |
Description
QGis crash when you use this code into the console (with a real project):
table = QgsComposerAttributeTableV2(iface.activeComposers()[0].composition(), False)
cols = table.columns()
newCol = QgsComposerTableColumn("test_for_crash")
newCol.setAttribute("$id")
cols.append(newCol)
table.setColumns(cols)
[i.heading() for i in table.columns()]
Tested on Windows 7 QGis 2.18, Debian QGis 2.14 and FreeBSD 2.18.
Seems to be a bad_alloc. Here some logs from Debian:
Warning: QObject::connect: Cannot connect (null)::layerModified() to QgsComposerAttributeTableV2::refreshAttributes() Segmentation fault
FreeBSD show a windows with this error:
cxxrt::bad_alloc
Related issues
History
#1
Updated by Giovanni Manghi over 6 years ago
- Status changed from Open to Feedback
Is this related to #17180 ?
#2
Updated by Loïc BARTOLETTI over 6 years ago
No, they're different issues.
I'm working on several projects needing to use the composer and found these two bugs.
#3
Updated by Giovanni Manghi over 6 years ago
- Status changed from Feedback to Open
Loïc BARTOLETTI wrote:
No, they're different issues.
I'm working on several projects needing to use the composer and found these two bugs.
thanks, so please add a description to #17180 just not leave the link to stackexchange, thanks.
#4
Updated by Jürgen Fischer over 6 years ago
- Related to Bug report #17180: Atlas crash using python added
#5
Updated by Denis Rouzaud over 6 years ago
- Assignee deleted (
Denis Rouzaud)
#6
Updated by Nyall Dawson over 6 years ago
- Resolution set to invalid
- Status changed from Open to Closed
Your code is a bit wrong:
cols = table.columns()
...
cols.append(newCol)
table.setColumns(cols)
table.columns() is returning a pointer to the table's columns, so any edits (such as appending columns) happens immediately. There's no need to call table.setColumns() after. This is causing the crash, because the table already has ownership of that columns object.
Admittedly, not the nicest api!