window placement post 768 has changed by 25 pixels

[scsijon]

I have a user or two that use set placements for certain applications.

The clearest explanation is from jasper who uses precise on a 1024x768 screen (http://www.murga-linux.com/puppy/viewtopic.php?p=706900, fifth post down on page10 of the puppy thread).

After installing jwm 768 and Restarting X both abiword and dillo open their 936 x 702 windows in a new position (some 25 pixels right of their pre-768-installation point).

PS I used 936 x 702 because if I made my dillo or abiword less than full screen windows with other than a 4 x 3 ratio they only opened in the exact position I wanted on exactly

every second opening.

Also, If he uses a slightly smaller pixel size it works (further down the puppy page). Testing with snapshot 775 has not fixed the problem.

[scsijon]

sorry joe, don't know what happened with the font size changes

[joewing]

The window placement has changed somewhat, so I would expect that some windows might open in different locations. Are the applications being started with a "-geometry" option?

[scsijon]

Apparently not with the above, although another member (session) sugested it so at this stage he is trying to do it, however he's having problems with it.

quote" I tried +0+1 and +1+1, but no joy - perhaps as I'm using dillo v 2.2 from Tman."

leave it for the moment, we will see how they go and if they can work it out themselves using "-geometry" I shall close this problem. I don't think anyone understood that a change had altered the window placement, I know I didn't, and I always read the changelog before building a new jwm pet so I know what to try to test for.

Another (contact via pm to me) uses a widescreen and also has a similar problem, I have told them to follow the thread.

[joewing]

The relevant commits that would have changed the behavior are 3ee8ddcf07ac507eeacb74cf56565a11af902cbc and 4dd7c849bd9da4b60457e8314738b3bf0b4d967f. The problem these commits solve relates to the placement of windows that would go off the screen if placed where JWM would normally place a window.

Commit 3ee8ddcf07ac507eeacb74cf56565a11af902cbc makes it so JWM correctly accounts for the window border size.

Commit 4dd7c849bd9da4b60457e8314738b3bf0b4d967f makes it so JWM doesn't revert to the upper left-most position on the screen if the placement algorithm fails (instead, only the direction of the overflow is reset). In addition, 4dd7c849bd9da4b60457e8314738b3bf0b4d967f constrains newly mapped windows to the area of the screen that JWM is willing to use even if a program specifies its own position.

[scsijon]

and I think the problems relating were from me.

I actually was only giving comment, I didn't expect you to document the commits, but thank you, I shall have to read things more carefully rather than just scanning I think.

thank you again.

[scsijon]

closing this as i believe it has been sorted out