IRC channel logs

2026-05-31.log

back to list of logs

<daviid>tohoyn I'll try to look at this but probably won't be able to within the next few days - here, unless I first left-click on the image, as soon as I hit one of (any) arrow, it crashes, but the error messages varies, sometimes it is a Gsk related error, sometine it is Gsk relate- like - Gsk:ERROR:../../../gsk/gpu/gskvulkanimage.c:395:gsk_vulkan_image_new: assertion failed: (width > 0 && height > 0)\n Bail out!
<daviid>Gsk:ERROR:../../../gsk/gpu/gskvulkanimage.c:395:gsk_vulkan_image_new: assertion failed: (width > 0 && height > 0) - sometimes it is cairo related, like - guile: ../../../src/cairo-surface.c:1822: cairo_surface_set_device_scale: Assertion `status == CAIRO_STATUS_SUCCESS' failed. ... but if I left-click anywhere in the window before hinting any arrow, then it works fine ... if you have any idea ... let me know
<daviid>tohoyn wrt 'need to sudo to run an example ...', not sure, upstream distribute those examples, but none are being istalled ... not sure why would that aprticular example, unlike others (?) would require to be sudo ... I won't be able to help you 'there' ..
<daviid>tohoyn and sometimes, running ../../pre-inst-env ./layout-manager-2 (in the repo) works just fine, even if I do not left-click in the window .., mistery mistery ... I am not gona spend too much time on this, just so you know
<rlb>dsmith: you had freebsd building before?
<rlb>oh, interesting --- it was failing on an iconv/aix header, but then I tried gmake rather than make, and it started building just fine. So we have a gmake dependency.
<rlb>And do you happen to recall which pkg installs you needed?
<rlb>(That got further, but now it fails on GC_unregister_my_thread)
<rlb>Ahh, do we have to pick libgc-*threaded* on freebsd somehow...
<rlb>nvm (rtfm README)
<dsmith>Yes, had fbsd building. In fact, they just recently closed the bug.
<dsmith>rlb, need gmake and gm4
<dsmith>rlb: M4=gm4 ./autogen
<dsmith>rlb: ./configure --with-bdw-gc=bdw-gc-threaded
<dsmith>rlb: gmake
<dsmith>rlb, https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=282534#c5
<rlb>dsmith: thanks --- it actually worked here without M4=
<rlb>dsmith: wonder if the gnulib update could have affected that
<rlb>Did all the tests pass? encoding-escapes.test is halting the tests here atm,
<tohoyn>daviid: /usr/share/doc/g-golf/examples/gtk-4/layout-manager-2 needs sudo because it builds subdirectory demos and writes some files there
<dsmith>rlb, At the time, yes all non-skipped tests passed.
<dsmith>rlb, Added skipping for things like seeking within a hole.
<rlb>I wonder if it might be something about my particular envt --- from a first look it appears that the encoding is *not* failing as expected (attempting I think to encode non-ascii as ascii).
<rlb>Here it's ./check-guile encoding-escapes.test that fails, but it doesn't tell you much because as in other places, we just use pass-if for multiple (and x y z) conditions.
<rlb>(which provides no details about which one failed, or what the failing value was)
<rlb>In utf-8 I've augmented pass-if(-equal) to return the result, so you can (and (pass-if-equal ...) (pass-if-equal ...) ...) and it'll print the expected and actual values for each failure to the log. Much faster to track down what's wrong.
<rlb>i.e. you can still use the and for short circuiting, while also seeing what actually went wrong.
<daviid>tohoyn yes, it needs a locally built libfpt.so ... either remove this example from debian, or make it so it compiles installs the lib ...
<daviid>tohoyn see the (demos layout-manager-2-init) module, lines 82 -> end