discuss-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: GNUstep app fails on Ubuntu 16


From: Andreas Höschler
Subject: Re: GNUstep app fails on Ubuntu 16
Date: Thu, 7 May 2020 00:10:10 +0200

Hi Riccardo,

It actually does happen with a few form of my app only as I have just figured out. Small forms seem to work.

2020-05-06 18:04:06.080 SOObjectBrowser[2483:2483] <FBMarkupController: 0x276b8b0> <SOFormContext: 0x5c9f430> setAccesslevel 2
2020-05-06 18:04:06.087 SOObjectBrowser[2483:2483] Cairo status 'invalid value (typically too big) for the size of the input (surface, pattern, etc.)' in DPSinitgraphics
2020-05-06 18:04:06.087 SOObjectBrowser[2483:2483] returning controller <FBMarkupController: 0x276b8b0>
2020-05-06 18:04:06.092 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x5ef4b50)
2020-05-06 18:04:06.102 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x6251360)
2020-05-06 18:04:06.103 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x13ec870)
2020-05-06 18:04:06.114 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x1097a40)
2020-05-06 18:04:06.116 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x1097a40)
2020-05-06 18:04:06.117 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x1097a40)

When this error occurs the app is no longer usable. Quitting the app via the menu still works though.

Could it be that this happens if the form is too large and does not fit on the screen? I have resized my display to 1600 x 1200 pixels but the problem still occurs if I try to open some forms, it doe snot occur with others!? I couldn't pinpoint this to a specific GUI class. But it indeed seems to be cairo/back related!

This combo is so far unusable. I will try libart now ...

I have one Ubuntu test computer, but I don't remember which version it is, perhaps the same as yours, I could check.

What app are you running? the errors are not nice, something is not properly initialized I think and could be to "max int" for example a -1 that overflowed or so.

What do you mean with "forms" ? windows? views?

It's a rather complex and highly customized database application that connects to a RDBMS in which hundreds of forms (this is GUI descriptions) are stored. The app gets the description of these forms in propertyList format from the database, builds the GUI from it (resizable) and creates the applications menu with a menu entries for all the forms. When the user clicks on one of the menu entries the corresponding form (window) is created and presented to the user. This includes connecting to the database, grabbing data from it and presenting the data in the form (tableviews, textfields, tabviews, etc). 

I have attached screenshots of two of those forms to give you an idea (from the app running on MacOSX).



The app used to be source code compatible with GNUstep, meaning that I just had to compile the complete software stack (a dozen frameworks, some bundles and the app) on GNUstep and were ready to go. However, I have only really used the app on Solaris 10 with some ancient GNUstep tree so far (worked perfectly). This is my first (serious) attempt to get it to run on Linux with latest GNUstep sources.

The cairo backend produces 

2020-05-06 18:04:06.087 SOObjectBrowser[2483:2483] Cairo status 'invalid value (typically too big) for the size of the input (surface, pattern, etc.)' in DPSinitgraphics
2020-05-06 18:04:06.087 SOObjectBrowser[2483:2483] returning controller <FBMarkupController: 0x276b8b0>
2020-05-06 18:04:06.092 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x5ef4b50)
2020-05-06 18:04:06.102 SOObjectBrowser[2483:2483] WARNING: -drawGState called with a NULL target context ((null)) or source context (0x6251360)
...

and the libart backend

2020-05-06 18:43:11.295 SOObjectBrowser[8177:8177] File NSView.m: 1183. In -[NSView setFrame:] given negative width
2020-05-06 18:43:11.295 SOObjectBrowser[8177:8177] File NSView.m: 1188. In -[NSView setFrame:] given negative height
2020-05-06 18:43:11.384 SOObjectBrowser[8177:8177] Warning: shmget() failed: Invalid argument.
2020-05-06 18:43:11.384 SOObjectBrowser[8177:8177] Falling back to normal XImage (will be slower).
2020-05-06 18:43:11.384 SOObjectBrowser[8177:8177] Warning: failed to create image for window!
2020-05-06 18:43:11.387 SOObjectBrowser[8177:8177] Warning: shmget() failed: Invalid argument.
2020-05-06 18:43:11.387 SOObjectBrowser[8177:8177] Falling back to normal XImage (will be slower).
2020-05-06 18:43:11.387 SOObjectBrowser[8177:8177] Warning: failed to create image for window!
2020-05-06 18:43:12.696 SOObjectBrowser[8177:8177] X-Windows error - BadAlloc (insufficient resources for operation)
          on display: :0
                type: 0
       serial number: 21442
        request code: 130
2020-05-06 18:43:12.697 SOObjectBrowser[8177:8177] X-Windows error - BadPixmap (invalid Pixmap parameter)
          on display: :0
                type: 0
       serial number: 21443
        request code: 2


It would be great to find out what causes this problem to get this running on GNUstep (on a recent OS) again. 

Please test with standard gnustep applications: Ink as a minimum example, a more complex one would be GWorkspace or any of my apps from GAP for example.

I will try that tomorrow. But since the problem occurs only in some of the more complex forms of my app chances are high that the error won't be triggered by Ink or GSWorkspace. I am wondering where in the back code I should place a break point (or raise an exception) to find the cause of the problem!?

Thanks,

 Andreas


reply via email to

[Prev in Thread] Current Thread [Next in Thread]