Thick lines are longer when printed

[phil_frost]

I've noticed when I export schematics with thick lines to postscript, those thick lines are longer than they would be if they were thin. See the attached screenshots for example. The line in that symbol representing the gate is thickness = 40, while the three shorter lines for the source, drain, and body are thickness = 20. Viewed in gschem, the top and bottom of the gate line align horizontally with the other lines. Viewed in postscript, the gate line is longer.

It's especially problematic with the MOSFET symbol I'm trying to make here. I'd like to make the drain, source, and body lines thicker. If I do that, they get long enough to merge together. If I make them shorter so they look right in postscript, the symbol looks pretty goofy in the GUI.

[phil_frost (phil-frost)]

Ah yes, and version is 1.6.1.20100214 on Mac OS, built from source. Have the same problem on 1.6.0 on Ubuntu.

[Peter Clifton (pcjc2)]

Has this behaviour changed between 1.4.x and 1.6.x ?

I'm vaguely of the mind that it is the GUI which needs fixing, since I don't believe the postscript code has changed - I presume Phil, that you're running the official source version with no patches - which still uses the old postscript printing code.

[Krzysztof Kościuszkiewicz (k-kosciuszkiewicz)]

I can still reproduce this with 1.7.0.
This occurs when line end types do not match with the gschemrc setting for "output-capstyle" setting.
Current postscript output code overrides line end settings from gschem with the type specified with the config option.

[Peter Clifton (pcjc2)]

This will be resolved when gschem moves to using the same cairo rendering code for printing as it does for rendering. Assigning Peter Brett, as he is working on that code.

[Peter TB Brett (peter-b)]

BTW, that's under development in my 'libgedacairo' branch <http://repo.or.cz/w/geda-gaf/peter-b.git>. Take a look, if you want. Warning: not currently feature-complete, likely to asplode, use with caution.

[Peter TB Brett (peter-b)]

This isn't going to make it in time for 1.8.0, so I've removed it from that milestone.

[Vladimir Zhbanov (vzhbanov)]

I've made a patchset for current master git head, which solves this bug. See attachments. Last of them is a schematic to test objects.

[Vladimir Zhbanov (vzhbanov)]

Patch 1

[Vladimir Zhbanov (vzhbanov)]

Test schematic ;)

[Peter TB Brett (peter-b)]

Do you have a branch I can pull from?

[Vladimir Zhbanov (vzhbanov)]

No public branches, only local.
And I have some more related fixes (a couple). They are not for this bug, however.

[Vladimir Zhbanov (vzhbanov)]

First two patches are also related to bug #789427.

[Peter TB Brett (peter-b)]

I've committed this patch to master.