[Keith]

The right mouse button (RMB) works like this:

Drag right mouse up from lower left to upper right (or lower right to upper left) zooms in to fit the zoom area to the viewport.

Drag right mouse button down from upper left to lower right (or upper right to lower left) zooms out according to the ratio of the zoom area to the viewport.

Any other effect (single clicking / double clicking with the RMB) is ignored.

Using the wheel mouse (if you have one) zooms in if you scroll forwards. If you scroll backwards, it zooms out. You can set tha amount it zooms in/out per mouse wheel scroll wheel increment by the 'View->Pan/Zoom options'. I find a zoom in % of 125 and zoom out % of 80 is a reasonably good amount to zoom.

[Keith]

Hi,

THere was a but in versions 4.1.53 and one or two before that where sometimes paths could be dropped. It was fixed in 4.1.54. Please download the 4.1.55 latest build and let me know if the problem persists.

regards

Admin

[Keith]

Just tested on win32/win64, linux32/linux64 and it all seems to be working normally for me.

[Keith]

What platform/OS are you running on? Are you running in OpenGL mode or not? Posting the log file might help.

[Keith]

Yes, another release is imminent. The LEF/DEF parsers have been enhanced to support NONDEFAULTRULES in LEF/DEF, also there are 2 bugs recently found:

– a recent change meant vias were sometimes not being displayed in LEF/DEF data

– some nets were not being shown in DEF routing.

These are related to support for LEF/DEF 5.7 version and will be in 4.1.54 hopefully in the next few days.

If you find any more LEF/DEF bugs please let me know!

[Keith]

The latest version fixes several bugs, including a problem in the build on the 32 bit linux version which resulted in an error message along the lines of 'undefined symbol stat64'.

As always, let us know if you find any problems!

[Keith]

Version 4.1.48 is a recommended download as it fixes a bug introduced a few weeks ago whereby paths with variable extensions were being converted into truncated paths. This obviously caused problems with data that was subsequently streamed out via GDS2/OASIS.

Various bugfixes have been made to the 'Compare Cells' command and it is running very well now on multi-cpu machines. In fact this is what picked up the bug above, when two (supposedly identical) chips were compared, one which was the stream out of the other.

[Keith]

There was a problem with one of the builds where I'd upgraded to VC2010. The 4.1.44 build should have the correct DLLs though. See the thread under 'News' entitled Version 4.1.42.

Edit: Just checking, it appears thte 32 bit windows version 4.1.44 still has the 64 bit Qt dll's in the distrubution. Sorry, I don't quite know how that happenned.

2 solutions:

1) copy the Qt dll's from your old version 4.1.39 to the new version

or

2) I am uploading a v4.1.44 win32 build with the correct 32 bit dll's now.

[Keith]

Hi fixrouter4400, thanks for this info, it turns out that the Qt DLL's in the 32 bit version were actually the 64 bit ones, hence why it would not run.

I've updated the DLLs to the correct ones now and re-uploaded both the 32 and 64 bit windows builds just to be safe.

[Keith]

Sorry, not sure I understodd that. Did you get an error when you tried to run 4.1.43 initially? Can you say what the error was?

thanks

Keith

[Keith]

There is an updated version 4.1.42 for Windows users; there is no change for Linux/SunOS from 4.1.41.

This is because the Windows version has been updated to using the VS 2010 C++ compiler, and a few minor problems were encountered with this move. In particular the Qt dll's had to be all recompiled for bothe x86 and x64 versions, as Qt do not (yet) support VS 2010 directly. Qt 4.7.2 does compile OK for both 32 and 64 bit versions and (thanks to depends.exe for a bit of debugging some runtime library issues) everything seems fine now. But if you see any problems, please let me know.

Is VS 2010 any better than VS 2008? Well its too early to say, but Intellisense *does* seem better and more responsive. I don't know why they bothered to redesign the GUI though, there was nothing wrong with 2008. As for compilation speed, setting /MP (multiprocessor compilation) is definitely wizzo on 8 cores.

The Qt add-in to VS still seems to have the stupid bug that some moc files get marked as 'excluded from build'. If anyone else has encountered this and knows of a fix, I'd like to know.

[Keith]

I have the following env vars set for Python (this is on CentOS but still)

PYTHONHOME=/opt/ActivePython-2.6
PYTHONPATH=/opt/ActivePython-2.6:/home/keith/glade4_linux64/bin

I tried a few things and things like:

import tempfile
import math

all seemed to work OK.

[Keith]

What version of Python are you using on Linux?

Glade was compiled with ActiveState Python 2.6.4 as per the message in the logfile.

[Keith]

I have fixed this in version 4.1.39, available later today.

[Keith]

Hi FriendFX, there was a small bug recently introduced that caused this issue. I have fixed in in version 4.1.39, which will be available today.

[Author]

Posts