andy-spiro test rendering changed completely?
For some reason the livepatheffect/andy-spiro test is now rendered quite different from how it used to be rendered (it changed between the 22nd and the 25th). I didn't create the test and am quite unfamiliar with the LPEs, so I have no idea how to judge which is right and which is wrong. Perhaps anyone can shed some light on this? Of course any ideas on what caused the change (regression?) would also be welcome.
For reference, the two images (the original output/reference and the new output, respectively): http://home.hccnet.nl/th.v.d.gronde/andy-spiro1.png http://home.hccnet.nl/th.v.d.gronde/andy-spiro2.png
I think the change is not in Spiro but in the pattern-along-path LPE(s) which were recently changed.
As a side note, so cool that the testing framework finds such discrepancies for us automatically! It's invaluable.
bulia byak wrote:
I think the change is not in Spiro but in the pattern-along-path LPE(s) which were recently changed.
But is it expected/desired that the output also changes? I didn't see any discussion on the mailing list about this during the past few days, and I also couldn't find any SVN revision since the 22nd that touched lpe-patternalongpath directly.
As a side note, so cool that the testing framework finds such discrepancies for us automatically! It's invaluable.
I couldn't agree more :) (Although it's still not fully automatic, in the sense that I still have to start the tests manually on my notebook.)
On Mon, Jan 26, 2009 at 1:27 PM, Jasper van de Gronde
But is it expected/desired that the output also changes? I didn't see any discussion on the mailing list about this during the past few days, and I also couldn't find any SVN revision since the 22nd that touched lpe-patternalongpath directly.
"Pattern along path" is a generic term here; there was a proliferation of similar LPEs, I'm not sure which one this file uses :) It may also be caused by a change in 2geom.
Of course such changes should be avoided whenever possible. We must remain backwards compatible. Sometimes, however, when bugs are fixed, changes in LPE rendering are inevitable. Hopefully this one can be reverted, otherwise the early adopters of our LPEs will be rather upset.
-----Original Message----- From: Jasper van de Gronde [mailto:th.v.d.gronde@...528...] Sent: maandag 26 januari 2009 18:27 To: Inkscape Devel List Subject: Re: [Inkscape-devel] andy-spiro test rendering changed completely?
bulia byak wrote:
I think the change is not in Spiro but in the pattern-along-path LPE(s) which were recently changed.
But is it expected/desired that the output also changes? I didn't see any discussion on the mailing list about this during the past few days, and I also couldn't find any SVN revision since the 22nd that touched lpe-patternalongpath directly.
Perhaps it is my bug fix 20565 or 20566? Although I don't see how that change would affect things...
As a side note, so cool that the testing framework finds such discrepancies for us automatically! It's invaluable.
I couldn't agree more :) (Although it's still not fully automatic, in the sense that I still have to start the tests manually on my notebook.)
YES! I hope this shows to people that we really need this testsuite, and that we have got to get this running on a server somewhere, perhaps even after every commit (and only display a new column in the ResultViewer when something changed, together with the SVN rev...)... Invaluable indeed.
Cheers, Johan
On 01/26/2009 03:43 PM, J.B.C.Engelen@...1578... wrote:
YES! I hope this shows to people that we really need this testsuite, and that we have got to get this running on a server somewhere, perhaps even after every commit (and only display a new column in the ResultViewer when something changed, together with the SVN rev...)... Invaluable indeed.
Once I get a little more info back from Jasper (I just emailed him), we can potentially get the testsuite going on our webserver. As for the tests... I've still yet to finish any successful test files because they crash once I get to a certain point. I'd abandon the files I've started, but obviously since there are crashes it's quite valuable. Also, do we know what the way to enable the experimental lpes on Linux is? For all the complaining when linux folk mess with things that affect windows people, I felt it's necessary to reciprocate since it's the other way around. ;)
Johan, would you like for me to email you backtraces?
Cheers, Josh
-----Original Message----- From: Josh Andler [mailto:scislac@...400...] Sent: dinsdag 27 januari 2009 20:14 To: Engelen, J.B.C. (Johan) Cc: th.v.d.gronde@...528...; inkscape-devel@lists.sourceforge.net Subject: Re: [Inkscape-devel] andy-spiro test rendering changed completely?
On 01/26/2009 03:43 PM, J.B.C.Engelen@...1578... wrote:
YES! I hope this shows to people that we really need this testsuite, and that we have got to get this running on a server somewhere, perhaps even after every commit (and only display a new column in the ResultViewer when something changed, together with the SVN
rev...)...
Invaluable indeed.
Once I get a little more info back from Jasper (I just emailed him), we can potentially get the testsuite going on our webserver. As for the tests... I've still yet to finish any successful test files because they crash once I get to a certain point. I'd abandon the files I've started, but obviously since there are crashes it's quite valuable. Also, do we know what the way to enable the experimental lpes on Linux is? For all the complaining when linux folk mess with things that affect windows people, I felt it's necessary to reciprocate since it's the other way around. ;)
Open "/src/live_effects/effect.cpp", and on the first line, write: "#define LPE_ENABLE_TEST_EFFECTS"
(yes, i am not going to figure out how to add it to the configure script)
Johan, would you like for me to email you backtraces?
Always :-)
participants (4)
-
unknown＠example.com -
bulia byak -
Jasper van de Gronde -
Josh Andler