After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 723811 - Smart Playlist Editor has greater width than screen
Smart Playlist Editor has greater width than screen
Status: RESOLVED FIXED
Product: banshee
Classification: Other
Component: User Interface
git master
Other Linux
: Normal major
: ---
Assigned To: Banshee Maintainers
Banshee Maintainers
Depends on:
Blocks:
 
 
Reported: 2014-02-07 02:10 UTC by Nicholas Little
Modified: 2014-03-29 15:21 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Swaps Double.MaxValue for int.MaxValue on a Spinner Control (991 bytes, patch)
2014-02-07 02:10 UTC, Nicholas Little
committed Details | Review
Before (29.37 KB, image/png)
2014-02-19 01:26 UTC, Nicholas Little
  Details
After (22.30 KB, image/png)
2014-02-19 01:27 UTC, Nicholas Little
  Details
Before Screen (156.34 KB, image/png)
2014-03-29 13:06 UTC, Nicholas Little
  Details

Description Nicholas Little 2014-02-07 02:10:45 UTC
Created attachment 268369 [details] [review]
Swaps Double.MaxValue for int.MaxValue on a Spinner Control

While using Gnome 3, I've noticed that opening the smart playlist editor results in a dialog that is complete and normal height, but is perhaps three times the screen width. Obviously this makes it very difficult to use.

I've attached a patch which corrects the problem on my system, turned out the problem was in Hyena, is this still the correct place to file?

Thanks.
Comment 1 Andrés G. Aragoneses (IRC: knocte) 2014-02-18 20:27:15 UTC
Hey Nicholas, can you post screenshots for the before and after? Sorry it took so long to reply to this...
Comment 2 Nicholas Little 2014-02-19 01:26:42 UTC
Created attachment 269647 [details]
Before
Comment 3 Nicholas Little 2014-02-19 01:27:42 UTC
Created attachment 269648 [details]
After

nicholas@shambles ~/Pictures $ file *.png
Screenshot from 2014-02-19 01:18:10.png: PNG image data, 3321 x 304, 8-bit/color RGBA, non-interlaced
Screenshot from 2014-02-19 01:23:27.png: PNG image data, 630 x 304, 8-bit/color RGBA, non-interlaced

Here you go :)
Comment 4 Andrés G. Aragoneses (IRC: knocte) 2014-03-26 18:35:57 UTC
Thanks for the screenshots, Nicholas. Can you attach fullscreen shots instead? I think the issue would be clearer with them.

Sorry that I haven't replied earlier!
Comment 5 Nicholas Little 2014-03-26 19:32:48 UTC
Unfortunately not at the minute! I'm unable to successfully compile banshee's GIO module :(

I'll make the necessary screen shots when I get round to sorting it.

Is the image size not representative enough? Width of 3321 vs 630 :)
Comment 6 Andrés G. Aragoneses (IRC: knocte) 2014-03-26 20:58:33 UTC
(In reply to comment #5)
> Is the image size not representative enough? Width of 3321 vs 630 :)

Well, that's the thing, I bet that you didn't actually see that image, you just were able to capture it via ALT+PrntScreen, right? Best way would be to see the image going outside the boundaries of the screen.
Comment 7 Nicholas Little 2014-03-29 13:06:59 UTC
Created attachment 273229 [details]
Before Screen

An updated before screenshot
Comment 8 Andrés G. Aragoneses (IRC: knocte) 2014-03-29 13:57:55 UTC
Ok!

BTW I have to apologize for asking the last screenshots, because I can actually reproduce the bug!

The thing is, this bug seems to be a gtk3 regression (because I can't reproduce it in the stable branch), therefore it's not really a hyena bug.

So, can you please open a new bug report about https://developer.gnome.org/gtk3/stable/GtkSpinButton.html#gtk-spin-button-new-with-range? Ideally it would include a very simple testcase (i.e.: a single-file python app) that reproduces the problem.

After you have reported it, I will commit your patch to hyena (but adding a comment that explicitly says that the usage of int.MaxValue instead of Double.MaxValue is actually a workaround, and referencing the URL of the new bug report against gtk+).

Thanks!
Comment 9 Nicholas Little 2014-03-29 14:45:45 UTC
Sounds Good :) Don't worry about it, my gtk-sharp build was broken so gudev-sharp was building essentially a blank dll. Ack, back to the cairo memory leak warnings for me!

After committing the workaround perhaps we should keep this one open and link it to 727294?

So when the regression is remedied and we get notification, we can revert the workaround?
Comment 10 Andrés G. Aragoneses (IRC: knocte) 2014-03-29 15:16:31 UTC
Comment on attachment 268369 [details] [review]
Swaps Double.MaxValue for int.MaxValue on a Spinner Control

(In reply to bug 727294 comment #2)
> What's a double in python?

No idea, but FYI int.MaxValue in C# is 2147483647, and Double.MaxValue  is 1.79769313486232E+308 , so you could modify your testcase there to include the latter (or doing a binary search, the lowest gdouble that starts showing the bug).


(In reply to comment #9)
> Sounds Good :) Don't worry about it, my gtk-sharp build was broken so
> gudev-sharp was building essentially a blank dll. Ack, back to the cairo memory
> leak warnings for me!
> 
> After committing the workaround perhaps we should keep this one open and link
> it to 727294?
> 
> So when the regression is remedied and we get notification, we can revert the
> workaround?

No need, as we're already both subscribed to the gtk+ bug. And in theory by adding the workaround, the banshee bug is really fixed (because lowering the max 
to 2147483647 is honestly not a big deal).

I've committed your patch now, with a bigger commit msg.
Comment 11 Andrés G. Aragoneses (IRC: knocte) 2014-03-29 15:17:24 UTC
This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report, and patch!