pcb

make check always builds and uses the gtkhid

Bug #1804564 reported by Chad Parker on 2018-11-22
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
pcb
Medium
Chad Parker

Bug Description

Even if you've selected to build the lesstif hid, the checking is always performed with the gtkhid. On the one hand, this leads to consistency of the testing environment, always gtk. On the other, if gtk isn't present, things will fail.

I think that the make check probably ought to use the batch hid, since we don't have any gui tests.

Changed in pcb:
status: New → Confirmed
importance: Undecided → Medium
Changed in pcb:
milestone: none → pcb-4.2.0
assignee: nobody → Chad Parker (parker-charles)
Chad Parker (parker-charles) wrote :

Branch LP1804564 sets "make distcheck" up to use the batch HID instead of the gtk hid.

Bert Timmerman (bert-timmerman) wrote :

Hi Charles,

I have one question:

Do you see the same spew as below:

<snippet>

Test: RouteStyles
(cd outputs/RouteStyles && ../../../src/pcbtest.sh --action-script routestyles.script default.pcb)
unknown action `PointCursor'
no action PointCursor()

</snippet>

I would have merged this commit into master if mot for the PointCursor action artefact, which is new to me.

Kind regards,

Bert Timmerman.

Changed in pcb:
status: Confirmed → In Progress
Chad Parker (parker-charles) wrote :

The PointCursor action doesn't exist in the batch HID, because there isn't a cursor to point!

What I'm wondering, however, is, what's calling that action? It's not coming directly from the routestyles.script. The only place that calls it is FitCrosshairIntoGrid.

My current working theory is that FitCrosshairIntoGrid gets called during the main initialization sequence.

Do you see in the output of any of the other tests?

Chad Parker (parker-charles) wrote :

One point of clarification:

"make check" should use whatever HID was configured. So, if the user built with lesstif, it should use the lesstif HID for the tests.

"make distcheck" should use the batch HID, since this is the lowest common denominator, and since tests should be GUI independent.

I think it is safe to ignore the missing PointCursor action.

Changed in pcb:
status: In Progress → Fix Committed
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers