Mail Archives: geda-user/2011/12/17/12:04:13
On Sat, Dec 17, 2011 at 05:37:39PM +0100, Juergen Harms wrote:
> >... solid enough ... take take those that should be in this bug release
>
> That is (necessarily) subjective. Personally, I have recently seen 2
> (for me) show stoppers for pcb (I cannot judge gschem):
> - the misplaced plated-drill layer (if I understand right: fix or
> workaround available for a stable release)
> - the "tests-fail bugs" (probably only on i3(5)86 platforms (#883768
> - fix committed, #860037, very visible since it throws test
> failures, but no important consequences - I do not know the status)
>
I have pushed a fix for the plated-drill issue. The test failure
is not something a user would see, nor is it a "real" problem. So,
worst case we can just disable that test and call it a day.
IMHO, we should just change the test to output a higher-resolution
picture to avoid rounding errors. In fact, if there is no opposition,
I will push such a change in the next couple of days.
> As long as these bugs are around, I would think that the "naive"
> user who gets pcb as a stable item of "his distro" is better off
> with the older releases - hence, I would not publish a new rpm to
> become available with Mageia 2 (official release forseen for the
> month of mai). I have not been sufficiently involved to judge
> whether other bugs should be added to this show-stopper category.
>
Let's not forget that pcb 20100929 is -far- from bug-free, and that's
what we're stepping up from. (As I recall, it was so bad at deciding
what I was trying to click on, it was nearly unusable to me. That,
and there were rounding errors -everywhere-.)
So, to me, a showstopper is: a save/load error (including exporters
and i18n bugs), or a segfault.
> Could this be a reasonable approach: along the feedback to your
> question, and with your insight, you create a candidate for a
> "stable release" that can be suggested for testing - and that you
> explicitely signal as such - (for test by involved users, distro
> maintainers; taking my example, not exhaustive tests: I, for
> instance, could not test any x86 packages, and would try to check
> with 1 or 2 layouts I recently did - mutch more representative if
> there would be a variety of test environments). With the "next mai"
> deadline of the oncoming Ubuntu and Mageia releases, there should be
> no serious timing problems.
>
This sounds fine to me, if DJ has the time and will to put out a
couple of RCs.
--
Andrew Poelstra
Email: asp11 at sfu.ca OR apoelstra at wpsoftware.net
Web: http://www.wpsoftware.net/andrew
"I don't understand. Are you saying dualism is always good, or always bad?"
- Raw text -