Hi guys, Bit of a conundrum, which way is the end marker on a curve meant to point? if I've got a curve thus
----------------
and I apply the same arrow marker to each end should I end up with
a) <------------->
or
b) <-------------<
?
I've read the spec, and cant work it out. I looked at the test cases, however they use triangles aligned like this V------V which look the same with both options a + b, so its not very helpful. batik renders b, so thats how I've done it currently, however I just wanted to be sure. (inkscape was rendering them as per option a before.)
Let me know what your thoughts are
Cheers
John
__________________________________ Do you Yahoo!? Yahoo! Small Business $15K Web Design Giveaway http://promotions.yahoo.com/design_giveaway/
On Mon, 12 Apr 2004, John Cliff wrote:
Hi guys, Bit of a conundrum, which way is the end marker on a curve meant to point? if I've got a curve thus
and I apply the same arrow marker to each end should I end up with
a) <------------->
or
b) <-------------<
?
It's my understanding that (a) is correct. I believe that there's a bug in the marker rendering code that causes the end marker to not behave correctly, but I couldn't pinpoint where it crops up.
There's also a bug that was causing the mid-markers to not work at all; I don't know if the latest code fixed that, but if not, that also requires further investigation. I figure if we can get end and start markers done, that's a great enhancement aside from the mid markers, but there's a whole range of cool things that we could do if we can get mid-markers working right too.
batik renders b, so thats how I've done it currently, however I just wanted to be sure. (inkscape was rendering them as per option a before.)
Hmm, well batik is kind of the de facto authority, so we'd do worse than to copy them. Can anyone see how the document would render with Illustrator?
Bryce
--- Bryce Harrington <bryce@...260...> wrote:
It's my understanding that (a) is correct. I believe that there's a bug in the marker rendering code that causes the end marker to not behave correctly, but I couldn't pinpoint where it crops up.
Thats been fixed, Carl had done it, for some reason it hadn't patched properly on my system the other day so it wasnt in that stuff you commited, but it was in the changes I commited last night.
There's also a bug that was causing the mid-markers to not work at all;
Mid markers are functioning properly now, I sorted that yesterday. (also nailed the bug where turning off the start marker shifted all the mid markers along)
batik renders b, so thats how I've done it
currently,
however I just wanted to be sure. (inkscape was rendering them as per option a before.)
Hmm, well batik is kind of the de facto authority, so we'd do worse than to copy them. Can anyone see how the document would render with Illustrator?
I've put it back to how inkscapes always done them before, as otherwise it breaks any old files people had done. (and it just feels more logical that way) Wouldnt be too hard to drive it from a preference tho, (ie batik-compatible-markers) and switch modes dependant on that.
If you want to see a fun use of markers, cvs up, create a line, set the start marker to INK_mTorso and the end marker to INK_mLegs using the drop downs and enjoy. :)
btw, I've been calling all the markers that I've been putting in markers.svg INK_mWhateverName to try and minimize the chances of confilcting with peoples own markers, Seem reasonable to y'all? If you add any of your own, can you create a line with it applied to one end called markerid_prev in the doc, I'm planning on using those to generate the previews.
cheers
John
__________________________________ Do you Yahoo!? Yahoo! Small Business $15K Web Design Giveaway http://promotions.yahoo.com/design_giveaway/
On Tue, 13 Apr 2004, John Cliff wrote:
--- Bryce Harrington <bryce@...260...> wrote:
It's my understanding that (a) is correct. I believe that there's a bug in the marker rendering code that causes the end marker to not behave correctly, but I couldn't pinpoint where it crops up.
Thats been fixed, Carl had done it, for some reason it hadn't patched properly on my system the other day so it wasnt in that stuff you commited, but it was in the changes I commited last night.
Mid markers are functioning properly now, I sorted that yesterday. (also nailed the bug where turning off the start marker shifted all the mid markers along)
Ah, good deal! Boy, the marker issues are getting solved quick. :-)
btw, I've been calling all the markers that I've been putting in markers.svg INK_mWhateverName to try and minimize the chances of confilcting with peoples own markers, Seem reasonable to y'all?
Sounds good to me. I wonder if we'd want to use this as a convention for other things like gradients?
If you add any of your own, can you create a line with it applied to one end called markerid_prev in the doc, I'm planning on using those to generate the previews.
Make sure to document this somewhere (perhaps in markers.svg itself, or in a README in that dir?)
Bryce
On Tue, 13 Apr 2004, John Cliff wrote:
I've put it back to how inkscapes always done them before, as otherwise it breaks any old files people had done. (and it just feels more logical that way) Wouldnt be too hard to drive it from a preference tho, (ie batik-compatible-markers) and switch modes dependant on that.
http://www.w3.org/TR/SVG11/painting.html#OrientAttribute
Please change the default back to Batik's behavior, and implement the orient attribute. Inkscape's old behavior violated the SVG specification.
Also, in the future, please refer to the specification FIRST if you are in doubt.
-mental
On Mon, 12 Apr 2004, Bryce Harrington wrote:
Hmm, well batik is kind of the de facto authority, so we'd do worse than to copy them. Can anyone see how the document would render with Illustrator?
Quick note: the proper marker orientation depends on the marker's "orient" attribute. See http://www.w3.org/TR/SVG11/painting.html#OrientAttribute
Batik's behavior (b) is indeed the correct one when the attribute is unspecified (defaults to "0"). A value of "auto" yields behavior (a).
-mental
--- MenTaLguY <mental@...3...> wrote:
Quick note: the proper marker orientation depends on the marker's "orient" attribute. See
http://www.w3.org/TR/SVG11/painting.html#OrientAttribute
Batik's behavior (b) is indeed the correct one when the attribute is unspecified (defaults to "0"). A value of "auto" yields behavior (a).
reading the spec is what confused me in the first place, the markers I have are all orient = auto yet batik renders behaviour (b). the spec talks about the positive x axis of the marker being aligned to the tangent vector for the path segment going in/out of the vertex. thing is my vector maths sucks, I'm assuming that the tangent vector is the one heading out of the point (ie if your travelling from start->end the one that wouldn't be going back on yourself) in which case I can't see how the answer is ever behaviour (a).
I'll switch it back tomorrow, too late tonight.
cheers
John
__________________________________ Do you Yahoo!? Yahoo! Tax Center - File online by April 15th http://taxes.yahoo.com/filing.html
On Tue, 2004-04-13 at 19:09, John Cliff wrote:
reading the spec is what confused me in the first place, the markers I have are all orient = auto yet batik renders behaviour (b). the spec talks about the positive x axis of the marker being aligned to the tangent vector for the path segment going in/out of the vertex. thing is my vector maths sucks, I'm assuming that the tangent vector is the one heading out of the point (ie if your travelling from start->end the one that wouldn't be going back on yourself) in which case I can't see how the answer is ever behaviour (a).
Ahh, no, I misunderstood you and was was wrong (and so is Batik, seemingly) -- for orient="auto", behavior (a) is correct.
Sorry about that.
-mental
participants (3)
-
Bryce Harrington -
John Cliff -
MenTaLguY