I'm using the Arch Linux AUR package from http://aur.archlinux.org/packages.php?ID=16851

Up until a day ago they worked fine, but after a system update Shift + F1-F10 doesn't register with vba-m. Pressing F1-F10 to load a save state works fine, as does saving the state manually from the menu.

The packages I updated just prior to this issue were mpfr, xf86-input-evdev, xf86-input-keyboard, xf86-input-mouse, xf86-input-synaptics, xf86-video-intel, farsight2, fceux, glade, and virtualbox. Removing or reinstalling any of these packages does not resolve the issue. Reinstalling vba-m does not resolve the issue either. Saving states does not work in vanilla VBA. Shift+F1 works in every other program.

This is not our problem, take it up with the distro.

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

I am using a completely different distro (Ubuntu 10.04 LTS), and I have the same issue.

Happening for me as well as of 1042 on a freshly installled Lubuntu 12.04, but worked on 11.10. This guy thinks a Xorg update might have broken it:
https://bbs.archlinux.org/viewtopic.php?id=128329

(10-14-2011 10:33 AM)Squall Leonhart Wrote:  This is not our problem, take it up with the distro.

same on gentoo

whos problem is it then if not yours?

the idiot who broke the linux libraries.

vba-m isn't the only one affected.

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

link to upstream bug of that library?

maybe related

https://bbs.archlinux.org/viewtopic.php?...3#p1015593

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

Right, but we can't include a (random) fork of evdev in main repositories. That is pretty much against most policies and not sane to expose the user to testing that.

Telling users to report this downstream is the wrong way to go, because this issue is distribution agnostic.

Why not just use a different method for saving states?
Otherwise plz care to report this bug to freedesktop.org cause it's affecting your code/software.

You could just NOT go out on a tandem and break previously working functionality that more than quite a few applications have used for ever with no issues previously.

VBA doesn't get updated anymore so theres not gunna be a fix there, nor is epsxe expected to get a linux update any time soon.

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

Quote:I think the problem is that vba-m don't consider shift as a modifier key. When I try to set the pad buttons, it registered Shift_L and Shift_R as acceptable keys.

GTK considers all the normal modifiers properly, however the input configuration goes through the input library, which shift can also be used validly as a button

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

It's not limited to GTK. I just tried wxvba-m svn 1102 and it also loads the savestate instead of saving. Like others, I have to go into the menu and select save.

yep, these fuckers broke it by adding useless shit to evdev

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7

(06-21-2012 06:21 PM)Squall Leonhart Wrote:  yep, these fuckers broke it by adding useless shit to evdev

Which means now we have to figure out a fix to this? Great. I can barely understand the code and I'm trying to find out where the input would be read at. So far I've found it's in the SDL sources, but I'm having trouble going any further.

Edit: I'm unsure, but it looks like a hotfix would have to be written in that hard codes shift into the emulator. Putting SDLK_RSHIFT and SDLK_LSHIFT directly into the code.

(06-23-2012 06:52 AM)Universal Wrote:  

(06-21-2012 06:21 PM)Squall Leonhart Wrote:  yep, these fuckers broke it by adding useless shit to evdev

Which means now we have to figure out a fix to this? Great. I can barely understand the code and I'm trying to find out where the input would be read at. So far I've found it's in the SDL sources, but I'm having trouble going any further.

Edit: I'm unsure, but it looks like a hotfix would have to be written in that hard codes shift into the emulator. Putting SDLK_RSHIFT and SDLK_LSHIFT directly into the code.

Since upstream does not care about it... mail me if you got something (hasufell@gentoo.org) or create a fork on github or anywhere else. I might also take a look at it when I have time.

Posting in this thread any further seems useless.

(06-23-2012 06:52 AM)Universal Wrote:  

(06-21-2012 06:21 PM)Squall Leonhart Wrote:  yep, these fuckers broke it by adding useless shit to evdev

Which means now we have to figure out a fix to this? Great. I can barely understand the code and I'm trying to find out where the input would be read at. So far I've found it's in the SDL sources, but I'm having trouble going any further.

Edit: I'm unsure, but it looks like a hotfix would have to be written in that hard codes shift into the emulator. Putting SDLK_RSHIFT and SDLK_LSHIFT directly into the code.

No, a fix would be not fucking with things that work in the first place

Midnight commander was also affected apparently by these retarded changes.

VBA-M | NGOHQ | Post Impact Productions | TNHW | XBCD 0.2.7