Mesh gradient status and plea for help with testing.
Hi,
Mesh gradients are included in SVG 2 but their remaining in the specification is on shaky grounds.[1] We need to generate suitable SVG content using meshes in order to generate enough buzz to get the browser vendors to take notice. With that goal in mind, we would like see meshes enabled in the 0.92.x release but need help with testing.
I've just checked into trunk the final GUI improvements to mesh handling. Please test and give me feedback. (And share any cool content you creat!) I will be merging the changes into 0.92.x in the next couple of days. For more information on how to use meshes see:
http://wiki.inkscape.org/wiki/index.php/Mesh_Gradients
Thanks,
Tav
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
On 9 Nov 2016 1:35 am, <mihaela.jurkovic@...400...> wrote:
Awesome tool! It's been requested for years by many people, I'm glad we finally have it!
Sampling colors behind nodes is so awesome :) lovely when vectorizing a photo.
Swapping Fill and Stroke when one has a mesh makes them both solid. I'm guessing it has to do with having to duplicate them and scale to boundary box? If both Fill and Stroke are meshed nothing happens when I try Swapping.
Applying an existing mesh to an object through Fill and Stroke dialogue can be a bit confusing because you always see a different mesh name after choosing one. This needs some tooltip or maybe some other guideline, maybe in Statusbar?
Selecting corner nodes only works with Shift + click, click-dragging around like with nodes seems to activate a different mesh, overriding the one already applied to an object. It should probably do nothing, and if possible make it select nodes within the area, unless it conflicts with something else.
Smoothing colors (experimental) doesn't make much sense to me, it produced even more sharper color edges, but maybe I'm using it wrong.
For more complex meshes, and even simple ones that you come back to after a while, you forget the logic behind your choices and it's hard to visualize which handle belongs to which node (they can be dragged way out of their original positions so they end up closer to a different node). It would be great if we could visualize their belonging, color them blue as well when their node is selected, or with a different color, and keep the white for those points that aren't related to currently selected one?
Google revealed this for me http://tavmjong.free.fr/SVG/MESH/Mesh.html Is it still relevant? I love the examples, can we post it on the forum? Also http://svgopen.org/2011/papers/18-Advanced_Gradients_for_SVG/
Can you explain tensor handles and how they relate to near-by beziers to a non-techy person? What should I pay attention to when moving them so I can get some intuitive understanding of what they do?
Mihaela aka prkos On 08.11.2016 19:47, Tavmjong Bah wrote:
Hi,
Mesh gradients are included in SVG 2 but their remaining in the specification is on shaky grounds.[1] We need to generate suitable SVG content using meshes in order to generate enough buzz to get the browser vendors to take notice. With that goal in mind, we would like see meshes enabled in the 0.92.x release but need help with testing.
I've just checked into trunk the final GUI improvements to mesh handling. Please test and give me feedback. (And share any cool content you creat!) I will be merging the changes into 0.92.x in the next couple of days. For more information on how to use meshes see:
http://wiki.inkscape.org/wiki/index.php/Mesh_Gradients
Thanks,
Tav
[1] http://tavmjong.free.fr/svg2_status.html
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing listInkscape-devel@...1901...://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...> wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Hi to all!
I try to upload to inkscape.org a resource done time ago -very simple- with mesh, and show as white doc.
https://inkscape.org/en/~jabiertxof/%E2%98%85logo-mesh
Now compiling new changes but in my version of trunk I coulent copy a mesh gradient between documents.
Cheers, Jabier.
On Wed, 2016-11-09 at 16:46 +0000, C R wrote: Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...> wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
------------------------------------------------------------ ------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@...1901...://lists.sourceforge.net/lists/listinfo/inkscape-devel
------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
If we want the website to generate proper inkscape rendered pngs, then we'll need to get inkscape running on a CentOS 6 machine. Which is really where that whole thing falls apart. All the deps are so old that it makes the problem really hard.
But, if it's possible to package inkscape 0.92 into a self contained opt type package, maybe it would be possible?
I'd like to hear the thoughts of the more experenced package people for Linux.
Martin,
On Wed, 2016-11-09 at 18:13 +0100, Jabier Arraiza wrote:
Hi to all!
I try to upload to inkscape.org a resource done time ago -very simple- with mesh, and show as white doc.
https://inkscape.org/en/~jabiertxof/%E2%98%85logo-mesh
Now compiling new changes but in my version of trunk I coulent copy a mesh gradient between documents.
Cheers, Jabier.
On Wed, 2016-11-09 at 16:46 +0000, C R wrote: Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...> wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@...1901...://lists.sourceforge.net/lis ts/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
I think the goal is a js.polifill to inlude and render ok the SVG as Tav says, I search a bit and dont find :(. To process a PNG we have thumb file now in uploads, let the user use.
Cheers, Jabier.
On Wed, 2016-11-09 at 12:55 -0500, Martin Owens wrote:
If we want the website to generate proper inkscape rendered pngs, then we'll need to get inkscape running on a CentOS 6 machine. Which is really where that whole thing falls apart. All the deps are so old that it makes the problem really hard.
But, if it's possible to package inkscape 0.92 into a self contained opt type package, maybe it would be possible?
I'd like to hear the thoughts of the more experenced package people for Linux.
Martin,
On Wed, 2016-11-09 at 18:13 +0100, Jabier Arraiza wrote:
Hi to all!
I try to upload to inkscape.org a resource done time ago -very simple- with mesh, and show as white doc.
https://inkscape.org/en/~jabiertxof/%E2%98%85logo-mesh
Now compiling new changes but in my version of trunk I coulent copy a mesh gradient between documents.
Cheers, Jabier.
On Wed, 2016-11-09 at 16:46 +0000, C R wrote: Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...
wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@...1901...://lists.sourceforge.net/l is ts/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Finally got time today to play with gradient meshes. First, I'd like to thank Tav for the amazing work. It's really awesome to see these make it into Inkscape, regardless of if they actually make it into the SVG 2 spec. I look forward to using them and showing them off to others interested in vector editing.
I agree that click+drag should used for selecting nodes. This was insta-confusing for me as well. And it threw away my gradient on more than one occasion when I forgot, forcing me to undo and try again. I think it used to be alt+m to add a mesh gradient... but it doesn't seem to work anymore. Anyway, that hotkey was fine for a quick-mesh.
There does not seem to be any easy way to select more than one node if they overlap eachother, like they do at the center and left-hand side of the conical gradient. I would definitely try to make click-drag lasso or box select node instead of applying a gradient. Even if we wanted to use it for adding a gradient, I would expect a linear gradient to be added rather than mesh gradient, and to be done in real-time based on the two points (initial click-hold is the first end of the gradient, and then release to indicate the last point of the linear gradient).
For the conical gradient, I'd lock all the points that overlap together, as moving one node from the middle makes it not-so conical anymore. :) Thus, moving the central node would move the ones under it too to move the center/tip of the cone, and preserve the cone shape/effect.
Also, when you apply a mesh gradient fill to a shape, I would auto-switch to the mesh gradient tool, so the user can see all the mesh options at the top of the screen.
There does not seem to be a method of adding a node to an arbitrary position, which forces the user to increase the mesh size to get more nodes. A double-click should add more mesh-node geometry so the user can designate a different colour at the position clicked (following the convention used in linear gradient editing).
That's my first impression anyway. :) Excellent work as always, Tav. I look forward to seeing it evolve into a staple of vector editing in Inkscape.
-C
On Wed, Nov 9, 2016 at 6:43 PM, Jabier Arraiza <jabier.arraiza@...2893...> wrote:
I think the goal is a js.polifill to inlude and render ok the SVG as Tav says, I search a bit and dont find :(. To process a PNG we have thumb file now in uploads, let the user use.
Cheers, Jabier.
On Wed, 2016-11-09 at 12:55 -0500, Martin Owens wrote:
If we want the website to generate proper inkscape rendered pngs, then we'll need to get inkscape running on a CentOS 6 machine. Which is really where that whole thing falls apart. All the deps are so old that it makes the problem really hard.
But, if it's possible to package inkscape 0.92 into a self contained opt type package, maybe it would be possible?
I'd like to hear the thoughts of the more experenced package people for Linux.
Martin,
On Wed, 2016-11-09 at 18:13 +0100, Jabier Arraiza wrote:
Hi to all!
I try to upload to inkscape.org a resource done time ago -very simple- with mesh, and show as white doc.
https://inkscape.org/en/~jabiertxof/%E2%98%85logo-mesh
Now compiling new changes but in my version of trunk I coulent copy a mesh gradient between documents.
Cheers, Jabier.
On Wed, 2016-11-09 at 16:46 +0000, C R wrote: Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...
wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@...1901...://lists.sourceforge.net/l is ts/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
@ C R double-clicking does add another column of nodes in that place. It may be more than what you bargained for, but I guess grid structure is what it's about ATM.
If I double-click on a mesh line it works for me, but clicking in a space that is not on a line in my version of trunk does nothing... it should (probably) add a row and a column to the mesh. It's not more than I bargained for... it's a mesh after all.. I think the grid structure is implied by that. :)
My 2p. -C
Copying an object with a mesh into a new document doesn't copy the mesh with it!
BTW I was able to achieve this non-smooth pattern by moving nodes and handles (attached 2 screenshots). Is this expected or should it never happen?
Mihaela aka prkos
On 13.11.2016 16:20, C R wrote:
Finally got time today to play with gradient meshes. First, I'd like to thank Tav for the amazing work. It's really awesome to see these make it into Inkscape, regardless of if they actually make it into the SVG 2 spec. I look forward to using them and showing them off to others interested in vector editing.
I agree that click+drag should used for selecting nodes. This was insta-confusing for me as well. And it threw away my gradient on more than one occasion when I forgot, forcing me to undo and try again. I think it used to be alt+m to add a mesh gradient... but it doesn't seem to work anymore. Anyway, that hotkey was fine for a quick-mesh.
There does not seem to be any easy way to select more than one node if they overlap eachother, like they do at the center and left-hand side of the conical gradient. I would definitely try to make click-drag lasso or box select node instead of applying a gradient. Even if we wanted to use it for adding a gradient, I would expect a linear gradient to be added rather than mesh gradient, and to be done in real-time based on the two points (initial click-hold is the first end of the gradient, and then release to indicate the last point of the linear gradient).
For the conical gradient, I'd lock all the points that overlap together, as moving one node from the middle makes it not-so conical anymore. :) Thus, moving the central node would move the ones under it too to move the center/tip of the cone, and preserve the cone shape/effect.
Also, when you apply a mesh gradient fill to a shape, I would auto-switch to the mesh gradient tool, so the user can see all the mesh options at the top of the screen.
There does not seem to be a method of adding a node to an arbitrary position, which forces the user to increase the mesh size to get more nodes. A double-click should add more mesh-node geometry so the user can designate a different colour at the position clicked (following the convention used in linear gradient editing).
That's my first impression anyway. :) Excellent work as always, Tav. I look forward to seeing it evolve into a staple of vector editing in Inkscape.
-C
On Wed, Nov 9, 2016 at 6:43 PM, Jabier Arraiza <jabier.arraiza@...2893...> wrote:
I think the goal is a js.polifill to inlude and render ok the SVG as Tav says, I search a bit and dont find :(. To process a PNG we have thumb file now in uploads, let the user use.
Cheers, Jabier.
On Wed, 2016-11-09 at 12:55 -0500, Martin Owens wrote:
If we want the website to generate proper inkscape rendered pngs, then we'll need to get inkscape running on a CentOS 6 machine. Which is really where that whole thing falls apart. All the deps are so old that it makes the problem really hard.
But, if it's possible to package inkscape 0.92 into a self contained opt type package, maybe it would be possible?
I'd like to hear the thoughts of the more experenced package people for Linux.
Martin,
On Wed, 2016-11-09 at 18:13 +0100, Jabier Arraiza wrote:
Hi to all!
I try to upload to inkscape.org a resource done time ago -very simple- with mesh, and show as white doc.
https://inkscape.org/en/~jabiertxof/%E2%98%85logo-mesh
Now compiling new changes but in my version of trunk I coulent copy a mesh gradient between documents.
Cheers, Jabier.
On Wed, 2016-11-09 at 16:46 +0000, C R wrote: Do you think that is enough? We need to show them and the community the advantages of the feature to get enough backing for it, don't you think?
-C
On 9 Nov 2016 4:22 pm, "Victor Westmann" <victor.westmann@...400...
wrote:
Then we all go together to Google Chrome and mozilla firefox and fill the feature request with this featute
On Nov 9, 2016 2:47 AM, "C R" <cajhne@...400...> wrote:
Okay, then we need video of how gradient meshes help make producing vector art faster and better. And show the clear advantages of having it renderable as vector in browser which eliminates the need for processing it into done other format in the first place.
-C
On 9 Nov 2016 9:39 am, "Eduard Braun" <Eduard.Braun2@...173...> wrote:
Am 09.11.2016 um 09:00 schrieb C R:
I think we should leverage our community power and redo some of the more popular Wikipedia graphics with gradient meshes. If the graphics are clearly better, we should be able to make a better case for meshes in svg 2. Maybe hold some sort of contest for producing the best gradient mesh alternative graphics each month (ongoing).
Thoughts?
That won't work... SVGs on Wikipedia are never directly sent to the users browser but instead are always pre-rendered by Wikimedia Servers using librsvg (which to the best of my knowledge does not have any gradient mesh support).
Therefore including gradient meshes in Wikipedia graphics would basically break them and would send the opposite message to people (like with flowRegions from flowed text which renders as black rectangles which is mostly blamed on Inkscape being "not standard conformant").
The case would look completely different if somebody took on the challenge to implement gradient meshes in librsvg. Then we could use them for graphics on Wikipedia and people would start to ask: "Why does the image look much better on Wikipedia than it does when rendered by my browser?"
Regards, Eduard
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@...1901...://lists.sourceforge.net/l is ts/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Wed, 2016-11-09 at 02:33 +0100, mihaela.jurkovic@...400... wrote:
Awesome tool! It's been requested for years by many people, I'm glad we finally have it! Sampling colors behind nodes is so awesome :) lovely when vectorizing a photo.
Thanks for the positive feedback!
> Swapping Fill and Stroke when one has a mesh makes them both solid. I'm guessing it has to do with having to duplicate them and scale to boundary box? If both Fill and Stroke are meshed nothing happens when I try Swapping.
Fixed in trunk. It was due to an oversight.
> Applying an existing mesh to an object through Fill and Stroke dialogue can be a bit confusing because you always see a different mesh name after choosing one. This needs some tooltip or maybe some other guideline, maybe in Statusbar?
Hopefully we will have a visual paint server selector in the near future so this won't be a problem.
> Selecting corner nodes only works with Shift + click, click-dragging around like with nodes seems to activate a different mesh, overriding the one already applied to an object.
It should probably do nothing, and if possible make it select nodes within the area, unless it conflicts with something else.
I'll look at that possibility. Right now, that's the easiest way to generate a new mesh (it follows the Linear/Radial gradient pattern of generating a new gradient but with Linear/Radial gradients the stops are preserved. I can probably disable creating a new mesh if one already exists.
> Smoothing colors (experimental) doesn't make much sense to me, it produced even more sharper color edges, but maybe I'm using it wrong.
It can probably be removed. Bicubic interpolation is a much better solution.
> For more complex meshes, and even simple ones that you come back to after a while, you forget the logic behind your choices and it's hard to visualize which handle belongs to which node (they can be dragged way out of their original positions so they end up closer to a different node). It would be great if we could visualize their belonging, color them blue as well when their node is selected, or with a different color, and keep the white for those points that aren't related to currently selected one?
Jabier was working on something like this. I'll ping him about the idea.
> Google revealed this for me http://tavmjong.free.fr/SVG/MESH/Mesh.html Is it still relevant? I
Is it still relevant? I love the examples, can we post it on the forum? Also http://svgopen.org/2011/papers/18-Advanced_Gradients_for_SVG/
Go ahead.
> Can you explain tensor handles and how they relate to near-by beziers to a non-techy person? What should I pay attention to when moving them so I can get some intuitive understanding of what they do?
Tensor handles control how the color flows inside a patch. The edges of a patch are painted the same way regardless of the position of the tensor handles. Tensor points are included in Type 7(?) PDF gradients and I was interested in them when I was studying smoothing at patch boundaries. It turns out that while they can help with smoothing they are not nearly as effective as using bicubic interpolation so I never pushed them in the SVG working group (they are not part of SVG 2). They also complicate the syntax. Tav
> Mihaela aka prkos > On 08.11.2016 19:47, Tavmjong Bah wrote: > >
> > Hi,
Mesh gradients are included in SVG 2 but their remaining in the specification is on shaky grounds.[1] We need to generate suitable SVG content using meshes in order to generate enough buzz to get the browser vendors to take notice. With that goal in mind, we would like see meshes enabled in the 0.92.x release but need help with testing.
I've just checked into trunk the final GUI improvements to mesh handling. Please test and give me feedback. (And share any cool content you creat!) I will be merging the changes into 0.92.x in the next couple of days. For more information on how to use meshes see:
http://wiki.inkscape.org/wiki/index.php/Mesh_Gradients
Thanks,
Tav
Mesh gradients are really great!
On 8 November 2016 at 19:47, Tavmjong Bah <tavmjong@...8...> wrote:
Hi,
Mesh gradients are included in SVG 2 but their remaining in the specification is on shaky grounds.[1] We need to generate suitable SVG content using meshes in order to generate enough buzz to get the browser vendors to take notice.
For what it's worth, the related Firefox (Gecko) bug is here: https://bugzilla.mozilla.org/show_bug.cgi?id=1238882
With that goal in mind, we would like see meshes enabled in the 0.92.x release but need help with testing.
I've just checked into trunk the final GUI improvements to mesh handling. Please test and give me feedback. (And share any cool content you creat!) I will be merging the changes into 0.92.x in the next couple of days. For more information on how to use meshes see:
I didn't have time yet to try the UI changes out, though I hope to do so the next days to provide some feedback.
Best regards,
Sebastian
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
Though I have several smaller and bigger remarks on it:
- Double-clicking path with gradient tool selected causes program crash If you select the gradient tool and double-click a path to devide a row or column, the program crashes.
- Toggling sides between beziers and lines or making them elliptical deselects them When multiple corner nodes are selected and you toggle the sides between beziers and lines or make them elliptical, the selection of the nodes is removed. I'd expect the selection to be kept.
- Edges display update delayed While dragging around the nodes the display of the mesh is updated immediately, though the display of the mesh sides is updated with a delay. It would be better UX if they were updated instantenously
- Alt+B doesn't work The shortcut Alt+B to toggle the selected sides doesn't seem to work.
- Hard to distinguish nodes and handles With the only difference being the form of the node, corner nodes and bezier handles are hard to distinguish. A different color for corner nodes (like within the vector tool) would improve that.
- Hard to associate handles with nodes When bezier handles are near to each other it is hard to distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
- Dividing rows/columns requires a corner node to be selected To devide a row or column through a double-click requires at least one corner node to be selected. This is an unnecessary restriction and should be lifted.
- Snapping and preserving angle via Ctrl or Ctrl+Alt doesn't work The status message says that Ctrl allows to snap the angle and Ctrl+Alt to preserve the angle while editing the mesh (like within other tools), though they don't work.
- Status message wrong regarding Ctrl+Shift The status message says that Ctrl+Shift allows to scale around a center, though this doesn't apply to mesh gradients, or does it?
- Mesh tool is separate from other gradients The mesh gradient tool is just another type of gradient, so it should be part of the existing gradient tool instead of having a separate button within the toolbar.
- Linear and radial gradient tool allows to change mesh gradients but not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
- Replacing mesh gradients by another gradient type requires to switch to Fill and Stroke Unlike when switching between linear and radial gradients, to replace a mesh gradient by another gradient, you need to switch to the Fill and Stroke dialog. This is another part of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
- Creating a mesh requires dragging To create a mesh you need to drag the mouse a little. It would be better if a simple click would create a default mesh (spanning over the whole bounding box of the object) and dragging would allow to put up a mesh like creating an object.
- Clicking inside object hides mesh The display of the mesh gradient is toggled when clicking inside the related object. This is different to the behavior of linear and radial gradients, which stay visible when clicking into the object. This is also an issue regarding the separation of rows and columns via double click, because the mesh gets hidden afterwards.
- Toggling tensor points only switches the colors of the selected corner nodes to a mixed color (there are no tensor handles) There are no tensor handles displayed on Alt+G. The only thing I see is that the colors of the selected corner nodes get a mixed color applied. (Though I'm not sure how this feature is expected to work.)
- Color sampling hard to understand without example Picking the colors from underneath the mesh is a very cool feature, though it took me a bit to understand how it works. I'm not sure whether this can be removed in the UI, but this should at least be described in more detail within the wiki.
- Show shortcuts in button tooltips Several buttons within the mesh toolbar have shortcuts applied. Those shortcuts should be displayed within their tooltips.
For what it's worth, I tested the 64-bit version on Windows 7.
Thanks again for this great tool! I really hope mesh gradients get traction among the browser vendors!
Sebastian
On 9 November 2016 at 09:50, Sebastian Zartner <sebastianzartner@...400...> wrote:
Mesh gradients are really great!
On 8 November 2016 at 19:47, Tavmjong Bah <tavmjong@...8...> wrote:
Hi,
Mesh gradients are included in SVG 2 but their remaining in the specification is on shaky grounds.[1] We need to generate suitable SVG content using meshes in order to generate enough buzz to get the browser vendors to take notice.
For what it's worth, the related Firefox (Gecko) bug is here: https://bugzilla.mozilla.org/show_bug.cgi?id=1238882
With that goal in mind, we would like see meshes enabled in the 0.92.x release but need help with testing.
I've just checked into trunk the final GUI improvements to mesh handling. Please test and give me feedback. (And share any cool content you creat!) I will be merging the changes into 0.92.x in the next couple of days. For more information on how to use meshes see:
I didn't have time yet to try the UI changes out, though I hope to do so the next days to provide some feedback.
Best regards,
Sebastian
Thanks for your feedback!
On Wed, 2016-11-16 at 09:55 +0100, Sebastian Zartner wrote:
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
Though I have several smaller and bigger remarks on it:
- Double-clicking path with gradient tool selected causes program
crash If you select the gradient tool and double-click a path to devide a row or column, the program crashes.
Will fix.
- Toggling sides between beziers and lines or making them elliptical
deselects them When multiple corner nodes are selected and you toggle the sides between beziers and lines or make them elliptical, the selection of the nodes is removed. I'd expect the selection to be kept.
Will try to fix but not high priority at the moment.
- Edges display update delayed
While dragging around the nodes the display of the mesh is updated immediately, though the display of the mesh sides is updated with a delay. It would be better UX if they were updated instantenously
True. But may not be trivial to fix.
- Alt+B doesn't work
The shortcut Alt+B to toggle the selected sides doesn't seem to work.
Will try to fix.
- Hard to distinguish nodes and handles
With the only difference being the form of the node, corner nodes and bezier handles are hard to distinguish. A different color for corner nodes (like within the vector tool) would improve that.
Do you mean the Node tool? (Gray for nodes. Clear for handles?)
- Hard to associate handles with nodes
When bezier handles are near to each other it is hard to distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
- Dividing rows/columns requires a corner node to be selected
To devide a row or column through a double-click requires at least one corner node to be selected. This is an unnecessary restriction and should be lifted.
This is a bug. It appears that the first click of the double-click often deselects the object and the second click reselects it.
- Snapping and preserving angle via Ctrl or Ctrl+Alt doesn't work
The status message says that Ctrl allows to snap the angle and Ctrl+Alt to preserve the angle while editing the mesh (like within other tools), though they don't work.
These are unimplemented features. The message should be commented out.
- Status message wrong regarding Ctrl+Shift
The status message says that Ctrl+Shift allows to scale around a center, though this doesn't apply to mesh gradients, or does it?
Again, an unimplemented features. The message should be commented out.
- Mesh tool is separate from other gradients
The mesh gradient tool is just another type of gradient, so it should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
- Linear and radial gradient tool allows to change mesh gradients but
not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
This may be possible but may require a lot of work.
- Replacing mesh gradients by another gradient type requires to
switch to Fill and Stroke Unlike when switching between linear and radial gradients, to replace a mesh gradient by another gradient, you need to switch to the Fill and Stroke dialog. This is another part of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
Will study.
- Creating a mesh requires dragging
To create a mesh you need to drag the mouse a little. It would be better if a simple click would create a default mesh (spanning over the whole bounding box of the object) and dragging would allow to put up a mesh like creating an object.
Double clicking also creates a mesh. There was a reason a single click doesn't work... I don't remember at the moment why.
- Clicking inside object hides mesh
The display of the mesh gradient is toggled when clicking inside the related object. This is different to the behavior of linear and radial gradients, which stay visible when clicking into the object. This is also an issue regarding the separation of rows and columns via double click, because the mesh gets hidden afterwards.
Will study.
- Toggling tensor points only switches the colors of the selected
corner nodes to a mixed color (there are no tensor handles) There are no tensor handles displayed on Alt+G. The only thing I see is that the colors of the selected corner nodes get a mixed color applied. (Though I'm not sure how this feature is expected to work.)
Enabling tensor points adds four new handles to each patch that control how color flows inside the patch. They should not be effecting the corner node color.
Tensor points will be disabled. They were an experimental feature.
- Color sampling hard to understand without example
Picking the colors from underneath the mesh is a very cool feature, though it took me a bit to understand how it works. I'm not sure whether this can be removed in the UI, but this should at least be described in more detail within the wiki.
Will update the description (but low priority at moment).
- Show shortcuts in button tooltips
Several buttons within the mesh toolbar have shortcuts applied. Those shortcuts should be displayed within their tooltips.
Will do (but low priority at moment).
For what it's worth, I tested the 64-bit version on Windows 7.
Thanks again for this great tool! I really hope mesh gradients get traction among the browser vendors!
Thanks! Glad you liked it.
Tav
Sebastian
On Thu, 2016-11-17 at 11:07 +0100, Tavmjong Bah wrote:
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
You want me to update the branch to current trunk? Is a bit outdated.
Cheers, Jabier.
On Thu, Nov 17, 2016 at 11:07 AM, Tavmjong Bah <tavmjong@...8...> wrote:
Thanks for your feedback!
On Wed, 2016-11-16 at 09:55 +0100, Sebastian Zartner wrote:
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
- Hard to associate handles with nodes When bezier handles are near to each other it is hard to
distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
Looking forward to that. You could try improving that, Jabier.
- Mesh tool is separate from other gradients The mesh gradient tool is just another type of gradient, so it
should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
Jabier had made a bunch of improvements related to hiding widgets that are conceptually N/A. It works great here! Take a look at Spray tool mode change behavior (Change mode from Spray copies -> Spray objects into single path). It hides what is unrelated to currently selected mode. I'd move Create Mesh Gradient and Create Conical Gradient just behind Create Linear Gradient and Create Radial Gradient. Followed by what it affects (Fill or Stroke). Then change all other fields related to what type of gradient is selected in the first place.
- Linear and radial gradient tool allows to change mesh gradients but
not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
This may be possible but may require a lot of work.
I am for removing possibility to change gradients outside gradient tool. It is easy enough to learn shortcuts and we already have dedicated tool for that job.
Thanks! Glad you liked it.
Tav
Sebastian
This tool improves Inkscape's value as illustrating tool sooo much... Thanks, Tav!
Vlada
I just fix the branch to try Tav.
On Thu, 2016-11-17 at 14:02 +0100, Vladimir Savic wrote:
On Thu, Nov 17, 2016 at 11:07 AM, Tavmjong Bah <tavmjong@...8...> wrote:
Thanks for your feedback!
On Wed, 2016-11-16 at 09:55 +0100, Sebastian Zartner wrote:
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
- Hard to associate handles with nodes
When bezier handles are near to each other it is hard to distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
Looking forward to that. You could try improving that, Jabier.
- Mesh tool is separate from other gradients
The mesh gradient tool is just another type of gradient, so it should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
Jabier had made a bunch of improvements related to hiding widgets that are conceptually N/A. It works great here! Take a look at Spray tool mode change behavior (Change mode from Spray copies -> Spray objects into single path). It hides what is unrelated to currently selected mode. I'd move Create Mesh Gradient and Create Conical Gradient just behind Create Linear Gradient and Create Radial Gradient. Followed by what it affects (Fill or Stroke). Then change all other fields related to what type of gradient is selected in the first place.
- Linear and radial gradient tool allows to change mesh gradients
but not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
This may be possible but may require a lot of work.
I am for removing possibility to change gradients outside gradient tool. It is easy enough to learn shortcuts and we already have dedicated tool for that job.
Thanks! Glad you liked it.
Tav
Sebastian
This tool improves Inkscape's value as illustrating tool sooo much... Thanks, Tav!
Vlada
Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
I just make some changes to my mesh branch, now allow triangle knots on mousedown on a knot.
Cheers, Jabier.
On Thu, 2016-11-17 at 14:02 +0100, Vladimir Savic wrote:
On Thu, Nov 17, 2016 at 11:07 AM, Tavmjong Bah <tavmjong@...8...> wrote:
Thanks for your feedback!
On Wed, 2016-11-16 at 09:55 +0100, Sebastian Zartner wrote:
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
- Hard to associate handles with nodes
When bezier handles are near to each other it is hard to distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
Looking forward to that. You could try improving that, Jabier.
- Mesh tool is separate from other gradients
The mesh gradient tool is just another type of gradient, so it should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
Jabier had made a bunch of improvements related to hiding widgets that are conceptually N/A. It works great here! Take a look at Spray tool mode change behavior (Change mode from Spray copies -> Spray objects into single path). It hides what is unrelated to currently selected mode. I'd move Create Mesh Gradient and Create Conical Gradient just behind Create Linear Gradient and Create Radial Gradient. Followed by what it affects (Fill or Stroke). Then change all other fields related to what type of gradient is selected in the first place.
- Linear and radial gradient tool allows to change mesh gradients
but not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
This may be possible but may require a lot of work.
I am for removing possibility to change gradients outside gradient tool. It is easy enough to learn shortcuts and we already have dedicated tool for that job.
Thanks! Glad you liked it.
Tav
Sebastian
This tool improves Inkscape's value as illustrating tool sooo much... Thanks, Tav!
Vlada
Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Anothes simplify refactor done on my mesh branch. Tav. Are you suscribed to it?
On Thu, 2016-11-17 at 14:02 +0100, Vladimir Savic wrote:
On Thu, Nov 17, 2016 at 11:07 AM, Tavmjong Bah <tavmjong@...8...> wrote:
Thanks for your feedback!
On Wed, 2016-11-16 at 09:55 +0100, Sebastian Zartner wrote:
It took a bit, though here's my feedback. Generally the mesh gradient tool already works very well. So, congratulations and thank you a lot for that!
- Hard to associate handles with nodes
When bezier handles are near to each other it is hard to distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
Looking forward to that. You could try improving that, Jabier.
- Mesh tool is separate from other gradients
The mesh gradient tool is just another type of gradient, so it should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
Jabier had made a bunch of improvements related to hiding widgets that are conceptually N/A. It works great here! Take a look at Spray tool mode change behavior (Change mode from Spray copies -> Spray objects into single path). It hides what is unrelated to currently selected mode. I'd move Create Mesh Gradient and Create Conical Gradient just behind Create Linear Gradient and Create Radial Gradient. Followed by what it affects (Fill or Stroke). Then change all other fields related to what type of gradient is selected in the first place.
- Linear and radial gradient tool allows to change mesh gradients
but not vice versa When the linear and radial gradient tool is selected, you can also change mesh gradients, but when the mesh gradient tool is selected, you cannot edit linear or radial gradients. This inconsistency should be removed, i.e. you should always be able to edit the gradients. This goes hand in hand with the above idea of integrating the mesh tool into the existing gradient tool. (This also applies to conical gradients.)
This may be possible but may require a lot of work.
I am for removing possibility to change gradients outside gradient tool. It is easy enough to learn shortcuts and we already have dedicated tool for that job.
Thanks! Glad you liked it.
Tav
Sebastian
This tool improves Inkscape's value as illustrating tool sooo much... Thanks, Tav!
Vlada
Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Thank you for the quick response, Tav!
On 17 November 2016 at 11:07, Tavmjong Bah <tavmjong@...8...> wrote:
- Hard to distinguish nodes and handles With the only difference being the form of the node, corner nodes
and bezier handles are hard to distinguish. A different color for corner nodes (like within the vector tool) would improve that.
Do you mean the Node tool? (Gray for nodes. Clear for handles?)
Yes, I mean the Node tool. To be clear, mesh nodes and handles are all white and they should rather look like the ones within the Node tool - gray for nodes, clear for handles (or different colors). The important thing is that nodes and handles should have different colors.
- Hard to associate handles with nodes When bezier handles are near to each other it is hard to
distinguish which one belongs to which corner node. Therefore bezier handles and their related corner nodes should be connected through a line like it's done within the vector tool.
I am working on a solution where hovering over a corner node will temporarily highlight the associated handles. Drawing lines doesn't work well when the mesh is dense. Jabier has a branch that changes the color of the handles for selected corner nodes. If I recall correctly (I am having trouble compiling it now) it worked well for sparse meshes but not well for dense meshes.
Sounds good. I can see the issue with lines for the handles and dense meshes.
- Dividing rows/columns requires a corner node to be selected To devide a row or column through a double-click requires at least
one corner node to be selected. This is an unnecessary restriction and should be lifted.
This is a bug. It appears that the first click of the double-click often deselects the object and the second click reselects it.
Yes, that's what I see.
- Snapping and preserving angle via Ctrl or Ctrl+Alt doesn't work The status message says that Ctrl allows to snap the angle and
Ctrl+Alt to preserve the angle while editing the mesh (like within other tools), though they don't work.
These are unimplemented features. The message should be commented out.
- Status message wrong regarding Ctrl+Shift The status message says that Ctrl+Shift allows to scale around a
center, though this doesn't apply to mesh gradients, or does it?
Again, an unimplemented features. The message should be commented out.
Ok, I've filed https://bugs.launchpad.net/inkscape/+bug/1643217 to remove the message and https://bugs.launchpad.net/inkscape/+bug/1643218 to implement those features.
- Mesh tool is separate from other gradients The mesh gradient tool is just another type of gradient, so it
should be part of the existing gradient tool instead of having a separate button within the toolbar.
This will be hard to do. The editing the mesh gradient acts quite differently from the linear and radial gradients.
Sure, most of their options are different. Though they are all gradients, so they should be combined under one tool. I expect it to work like Vladimir wrote.
- Creating a mesh requires dragging To create a mesh you need to drag the mouse a little. It would be
better if a simple click would create a default mesh (spanning over the whole bounding box of the object) and dragging would allow to put up a mesh like creating an object.
Double clicking also creates a mesh.
Ah, thanks for the hint!
There was a reason a single click doesn't work... I don't remember at the moment why.
I guess the reason was that it makes the behavior inconsistent to the other gradient tools. Double clicking to create a mesh having the size of the bounding box seems to be fine for me. Though I still think that dragging should put up a mesh instead of also doing the same as a double click.
- Toggling tensor points only switches the colors of the selected
corner nodes to a mixed color (there are no tensor handles) There are no tensor handles displayed on Alt+G. The only thing I see is that the colors of the selected corner nodes get a mixed color applied. (Though I'm not sure how this feature is expected to work.)
Enabling tensor points adds four new handles to each patch that control how color flows inside the patch. They should not be effecting the corner node color.
Tensor points will be disabled. They were an experimental feature.
Ok.
Please let me know whether I should file bugs for the other issues I've mentioned.
Sebastian
participants (9)
-
unknown@example.com
-
C R
-
Eduard Braun
-
Jabier Arraiza
-
Martin Owens
-
Sebastian Zartner
-
Tavmjong Bah
-
Victor Westmann
-
Vladimir Savic