[Qgis-developer] Composer grid issue

[Marco Hugentobler]

Hi John

Good to hear that the offset issue seems to be fixed now.

> For some reason, my old project now crashes if I try to open in qgis.
> I suspect this might be from changes in the composer structure for qgs
> files, but not certain.

Oh, this should not happen. Project reading should be backwards compatible. 
Unfortunately, it does not crash with my test projects. Could you send me the 
project file for testing?

> 1) If there is no composer, the "Map Composer" menu item is empty.
> Perhaps this should be a submenu under "File" rather than it's own menu?

I'll do that. Though the problem might be that users don't find it easily at 
this place. I'll wait what the feedback on the list is.

> 2) If I try to close my composer window, I am confronted with a dialog
> that says, "Do you really want to remove the composer instance,
> 'name'?" Canceling will keep the composer, but you have the window
> occupying desktop space. If you say ok, the composer is gone for good.
> My expectation would be to close the window, receive no dialog, but
> have that composer available under the "Map Composer" menu item.
> Perhaps this is a work in progress.

Yes, the behaviour is different now because it is possible to have multiple 
composers in a project. The composers are deleted on window close (similar to 
most apps that are deleted on main window close). Otherwise the user cannot 
get rid of composer instances. You could minimize the window to have more 
screen space.
An option would be to have the close button and menu entry back that minimises 
(or even hides?) the composer window.

Regards,
Marco

Am Donnerstag, 29. Oktober 2009 16.55:49 schrieb John C. Tull:
> Hi Marco,
> 
> Excellent work with the grid composer. Things lined up in a new test I
> ran, so it appears you have handled the offset issue.
> 
> For some reason, my old project now crashes if I try to open in qgis.
> I suspect this might be from changes in the composer structure for qgs
> files, but not certain.
> 
> The control over precision and the option for labels following
> boundary direction are brilliant! Great work, as usual.
> 
> New issues from the other changes I have seen so far:
> 
> 1) If there is no composer, the "Map Composer" menu item is empty.
> Perhaps this should be a submenu under "File" rather than it's own menu?
> 
> 2) If I try to close my composer window, I am confronted with a dialog
> that says, "Do you really want to remove the composer instance,
> 'name'?" Canceling will keep the composer, but you have the window
> occupying desktop space. If you say ok, the composer is gone for good.
> My expectation would be to close the window, receive no dialog, but
> have that composer available under the "Map Composer" menu item.
> Perhaps this is a work in progress.
> 
> Also see https://trac.osgeo.org/qgis/ticket/2042
> 
> Cheers,
> John
> 
> On Oct 29, 2009, at 6:57 AM, Marco Hugentobler wrote:
> > Hi John
> >
> > Do you still have the issue with the shift between composer map
> > contents and
> > the composer grid with current svn?
> > I did some prints with the nv_grid shape  this morning (to png, 200
> > dpi) and
> > it seemed to be ok.
> >
> > Regards,
> > Marco
> >
> > Am Dienstag, 20. Oktober 2009 19.27:53 schrieb John C. Tull:
> >> Hi Marco,
> >>
> >> As always, thanks for your great work and progress on the composer!
> >> I'll look forward to testing your changes.
> >>
> >> Cheers,
> >> John
> >>
> >> On Oct 19, 2009, at 11:26 PM, Marco Hugentobler wrote:
> >>> Hi John
> >>>
> >>> Thank you for reporting. I'm currently implementing the possibility
> >>> to rotate
> >>> the composer map and this changes a lot in the composer grid code.
> >>> Therefore,
> >>> I'm going to address this issue after the rotation code is in trunk.
> >>>
> >>> Regards,
> >>> Marco
> >>>
> >>> Am Montag, 19. Oktober 2009 22.44:56 schrieb John C. Tull:
> >>>> Marco,
> >>>>
> >>>> I was excited to see the new map composer grid feature. I tried it
> >>>> out
> >>>> on some field maps last week and found out that the grids were not
> >>>> accurate. I believe you might be able to reproduce an error with
> >>>> the
> >>>> attached test file. It appears to be a dpi issue, for some reason.
> >>>>
> >>>> Here are my steps:
> >>>> 1) Open a new qgis canvas and add the attached shapefile.
> >>>> 2) Verify that the projection (CRS) is set to epsg 26911 for the
> >>>> project.
> >>>> 3) Zoom in to a subset of the grid at approximately 1:100000 scale.
> >>>> 4) Open the map composer and add the map canvas to the composer
> >>>> canvas.
> >>>> 5) Turn on the composer grid with a 5000x5000 spacing.
> >>>> 6) Now, zoom in the map composer 4 or 5 times.
> >>>> 7) With the map layer selected, update the preview and note the
> >>>> offset
> >>>> between the two grids.
> >>>> 8) Output a png image at 300 dpi (should look correct).
> >>>> 9) Output a png image at 200 dpi (the offset should show).
> >>>>
> >>>> On another note, it would be nice to be able to suppress the
> >>>> scientific notation of the grid labels. In the UTM example I have,
> >>>> the
> >>>> scientific notation is actually a larger label than the full
> >>>> integer
> >>>> value.
> >>>>
> >>>> Cheers,
> >>>> John
> >>>
> >>> --
> >>> Dr. Marco Hugentobler
> >>> HUGIS - GIS programming and consulting
> >>> Webereistrasse 66
> >>> CH-8134 Adliswil
> >>> marco at hugis.net
> >>> http://homepage.hispeed.ch/hugis
> 

-- 
Dr. Marco Hugentobler
HUGIS - GIS programming and consulting
Webereistrasse 66
CH-8134 Adliswil
marco at hugis.net
http://homepage.hispeed.ch/hugis