Great! Thank you. 🙂

This issue seems related to T67111. Perhaps they could be merged.

So, this is not a bug therefore this ticket could be closed.

Hi @Brendon Murphy (meta-androcto), I went through the samples_scene.py, trying to fix the errors and commented the original lines, and now it seems to work in the latest RC2. Here is the modified script, hope it helps.

Thank you Federico for testing it on Windows too. So it's confirmed. Do you think it can be fixed? This is a crucial add-on but in this state it's not only useless but also dangerous if your are not aware of how it messes up with the paths in the edited linked file.

This seems an easy thing to fix, replacing the scene.render.layers with scene.render.views as follows:

Unfortunately, with such scenario, it seems that even with the previous version 2.79b, when you edit the linked object using the Edit Linked Object add-on (ver. 0.8.1), and then return to the start file, it doesn't apply the correct relative path to the just edited file breaking links to external assets.

I can confirm the above issue on a daily build of version 2.80 (Hash: 65168825e0b0) on my laptop (Win10 64bit, GeForce 1070). The add-on seems to loose the correct relative path to the linked material, after editing the linked object, apart from other minor issues (see below):

Tested on my laptop (Win 10 Pro 64bit, Nvidia GTX 1070) with the RC1 (but the issue below was there even before the release candidate) and, despite the add-on activates and works apparently, the following is the error message in the console just after add-on activation and continuously running on the console:

I tested that scene file on my laptop and it renders fine. The memory consumption is about 20GB of RAM and 10GB of VRAM during the scene's rendering with tiles at 64x64, on a Windows 10 Pro 64bit OS, with Nvidia GeForce GTX 1070 (8GB VRAM), and Blender 2.80 (Hash: 12ceea04344a, 10/07/2019).

I've just tested the same exact build on my Win10 notebook and I am not able to reproduce that problem.

Yes, in fact it can be found inside ..\2.79\python\lib\site-packages\requests\packages but you cannot import directly the urllib3 module. Instead it can be imported as a variable:

After further research I found this article: Preventing Cycles Crash after CUDA timeout

I started experiencing such an issue after I updated one week ago my Win10 Pro 64bit OS from Build 1511 to 1803 on an ASUS G752VS notebook (NVIDIA GeForce 1070), and then I updated the Nvidia Graphics drivers to the latest version. Reinstalling the graphics drivers (398.11) with a clean install didn't fix the issue. The error appears randomly when switching the 3D View to Rendered mode; closing and restarting blender and re-opening the same .blend project file sometimes it allows the GPU to render without errors.
The Blender version is 2.79b 64bit official portable release; the same issue appears on the latest buildbot releases 2.79.4.
The following are screenshots of the issue appearing after switching the 3D View to Rendered mode inboth Blender 2.79b and 2.79.4:

Did you try to run Blender with its default settings? If that way it works, then you might try enabling one add-on at a time to discover which one could be the cause of the issue. You may also run Blender in debug mode and watch the Blender's console messages for clues. The last warning message is easy to fix in the preferences.

I confirm that the source of the issue is in the Clip Start value, which was set at 1mm. I will take it in consideration for the future, and from my side you can close this ticket and better spend developers' valuable time. Thank you for addressing me to the source of the problem.

Ooops, you're perfectly right, sorry. I will pay more attention in the future. Lesson learned. Thanks. :-)

Sybren, I simply replied to the questions, perhaps in a little too much verbose way. Anyway, the currently available developmental 64bit versions (hash: 1802d14) both experimental (VC14) and not-experimental work fine. Thanks.

Sybren, I usually download the development builds from the official repository (https://builder.blender.org/download/) and only Experimental Build Branch by VS 2015 (Windows Vista/7/8/10) version. I do not install Blender and instead I create a close controlled environment for each portable version I use.

Here on Win10 64bit + Blender 2.79.1 (Hash: a466d7a) 64bit portable the Amaranth Toolset add-on is working fine. Thank you for the fix.

I can confirm the same issue here on Win10Pro 64bit and Blender 2.79 official portable 64bit.

The information about this add-on on the Wiki page https://wiki.blender.org/index.php/Extensions:2.6/Py/Scripts/Paint/Palettes is outdated and still referencing version 0.9 and Blender 2.58 while current version is 0.9.2 and Blender 2.63. Thank you.

As per the R(0°) formula the "sqr" term should stand for or should be read as "to the power of 2" or "squared" not as "square root of":

I can confirm that in 2.79 official release 64bit that's is the behaviour. In Edit Mode you cannot hide the object currently being edited but you can hide other objects in the scene while you're in Edit Mode. It was the same in 2.78 and previous versions.

Gotcha! Thank you very much Brecht.

Sorry, I missed those changes because that is the release notes for 2.8, but what about 2.79 (https://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.79/Cycles)? Will these options still be available? Thanks.

Perhaps the "issue" raised by the user is due to the Glossy Indirect pass which if the Filter Glossy parameter is set higher that 1 will blur the glossy reflections, giving the impression that they are missing. Here are some examples and the .blend file created for this tests.

Here on Win10 Pro 64bit (build 1511) portable, and Blender 2.79RC2, the reflections work correctly with zero roughness as you can see below:

Thanks for your efforts. From several years I use this script named Link_Sky_to_Sun_Lamp.py by Oscurart and Greg Zaal: https://www.dropbox.com/s/wap2jimkud5dntv/link_sky_with_sun.py

Thank you Bastien for the clue about where the issue might be coming from. Regards.

Hello. I did some tests and found that it got broken with the 2.7x series while it was working with previous 2.6x versions. Hope it helps.

Hi Andrea, the hair rendering settings have been moved inside the Geometry options in the Render panel:

The same here with the official 2.78b portable version (see below)

The Open Street Map add-on is available only in the development version of Blender, inside the folder ...\2.78\scripts\addons_contrib\io_scene_open_street_map.py and you can get it by simply downloading a development version from here.

I tested your sample file with official versions 2.77, 2.76 and 2.62 and they all show that behaviour, right or wrong (or by design) it might be.

The OpenGL 2.1 is currently required to run the most recent builds of Blender.

Here on Win7 64bit and Blender 64bit portable versions I can confirm the issue in 2.78 while it works fine in 2.77. I've attached a folder with source and destination files plus screenshots on what happens in both Blender versions after appending the Group "APPEND Cubes" and then assigning the particle system with the same name to a plane, in case it helps.

The extended support date for Windows 7 is January 14, 2020 as per the following Microsoft document:

Thank you Brecht for the explanation.

I checked on my Win7 64bit system with Blender 2.76b 64bit official stable, and it recognizes those network paths mapped to a drive letter as well as shared folders on the network server as you can see below:

Hello. Here on a Windows 7 Pro 64bits OS, with Blender 2.76b 64bit official release I get the same behavior when inserting the RGB Curve node, despite having set the HDR image to be read as non-color data (I though it might be the source of the problem, but it wasn't).

That issue seems to be present also in the 2.76b official release 64bit on Win7 Pro 64bit.

I'm really sorry Bastien, and I apologize for the evidently wrong tone of my comment, totally avoidable and useless, I admit it.

Hi Alexey, my comment wasn't directed at you but at the developers.

After almost a month this as well as bug report T47025 have not been triaged yet, and nobody seems to have anything to comment. Please, could you take proper actions? Thank you.

Hi Brendon. Yes, I have very good reasons to still use \addons_extern, which are to keep in sync the \addons and \addons_contrib with the GIT repositories, and because from day one I discovered Blender four years ago I've NEVER installed it but always used the zipped version, which works perfectly, and it's out of control by Windows OS and its user rights management.

Aaron (Abpy) to let your addon live together the official one, and considering that I store all non-official addons like your inside the ...\scripts\addons_extern folder (as it was available in the past, patching the file addon_utils.py), I modified the __init__.py file of your addon at line 93 and 792 respectively as follows:

I wish to let you know that Blender 2.76 64bit stable release, zipped version, works fine and using the -r option it registers properly the application and its path as the default one to open .blend files, as well as I can now see the thumbnail preview of the blend file on my Windows desktop as well as in the file manager. Well done!

Peter, that's the reason why I don't install Blender but instead I use the zipped version so that I can set the rights I wish on the unzipped folders.

Hello, here with the zipped versions of both 2.75a and 2.76RC3 the add-on Sapling Generator works as expected when saving and loading tree presets in an old .blend project (made with v2.69) as well as in a brand new file. My current environment is i7-720Q, Windows 7 Pro 64bit, 8 GB RAM, Nvidia GTS-360M 1GB VRAM. Does the user have the rights to read/write on the \presets folder inside the add-on ...\2.75\scripts\addons\add_curve_sapling?

I can confirm the same issue here too on Win7 Pro 64bit, Blender 2.76 RC1 64bit (Hash: c586e30). Version 2.75a official (Hash: c27589e) 64bit works as expected.

Hi Sergey, I'm not sure this could be of any help but, just for curiosity I tested this bug in Blender 2.76 RC1, Windows 64bit, zipped, and it's still there but I found that subdividing once the huge plane it seems to fix the problem, as you can see below.

@Thomas Dinges (dingto) and @Sergey Sharybin (sergey), thank you for the explanations about the BVH; in fact it wasn't totally clear to me that "Persistent Images" was related to the texture images, since there is no mention of it as well as other options in the Render | Performance panel inside the online Manual (eg. the Persistent Images option was reported only in the Blender 2.65 Release Notes (Cycles)) . Anyway, that's another story, so go ahead and do your "magic" stuff in Blender.

If you plan to remove the BVH cache I do hope that the re-rendering process won't be affected in terms of time needed for every subsequent re-render of the same scene. Thank you.

Thank you Sergey. As far as I can see in my laptop's system-info.txt my graphics card is missing ARB_gpu_shader5 at least. Time to consider buying a new, more modern laptop.

OK, I understand and thank you Bastien for your reply, now that's clear.

Hi guys, sorry if I jump in but I not sure I understood the above issue. I've never seen the .blend scene "preview" thumbnail inside any Windows 7 explorer file manager, while I can see the saved preview of the .blend file inside the Blender's File Manager. I use only the zipped version of Blender and I do register manually the application every time I install a major "stable" release: blender.exe -R.

Hi Sergey, I am not sure if my GPU can support OpenSubDiv and since you mentioned above GL_EXT_geometry_shader4 which apparently seems supported by my graphics card, would you mind checking my system-info.txt report below and let me know about it? Thank you.

Just a brief update having downloaded the second testbuild and I still have NONE only option in the OpenSubdiv compute's dropdown list, being on CPU only mode (Compute Device is None).

Thank you Sergey for your kind reply. FYI, the currently installed NVIDIA driver is version 341.81 and if there is any further information I can provide to help you find a fix, please let it me know.

Just for your information, today I installed the latest 64bit zipped version of Blender (Hash: 6a80c2c) and differently from my previous post (see above), now none of the previously listed options is available in the System Preferences, and therefore I cannot enable Use OpenSubdiv in the Subsurf Modifier.

Here, on Windows 7 Pro 64bit, these are the options available in the OpenSubdiv Compute Type's drop-down menu in Preferences | System for the current Windows 64bit build Hash:44384c6 (by default it's set to None):

Hi Nick, is this what you were waiting for to upgrade the Cloud Generator add-on?

Hi Greg, I can confirm the same issue on my environment (Win7 Pro 64bit, 8GB RAM, CPU rendering only) with Blender 2.75a official release for Windows 64bit (ZIP). Only one time I was able to bake successfully AO without crashing Blender, after unwrapping again your model, but that was an exception because it continued to crash Blender as per your description. No special messages in the console.

Hi Gaia,

Oh, I didn't know about that minimum number of branching levels. Good to know and thank you Bastien.

I can confirm the reported bug under Windows 7 Pro 64bit.

I can confirm that the reported issue with number of leaves not changing is affecting Blender 2.73a (Hash: bbf09d9) Windows 64bit, ZIP version, as well as at the same time it's fixed (meaning that the number of leaves changes properly) in the recently published Blender 2.74 RC3, Windows 64bit, Zipped.

bat3a, for completeness of information I should mention also the following change after line 158 "force alla contrib addons to be 'TESTING'":

Hello bat3a and mont29,

Thank you Julian for your reply and I appreciate it very much the explanation. You're doing a very good job.

I possibly found the papers that refers to this problem and, despite not being a matematician, I now probably understand why an alternate solution hasn't been implemented: it is very complex and requires an high degree of knowledge in matematics, not to speak about its implementation. It seems to be more appropriate for a GSOC submission than a To Do that nobody would do. Sorry for my ignorance and for opening my mouth without knowing what I was talking about.

Just a simple question Julian. Shouldn't this (or T32022) be classified as a "To Do" up for grab instead of simply closing it, considering that you stated that a later version of the algorithm exists that solves the problem, but it has not been implemented yet? Thank you.

I understand, so I hope you don't mind if I will keep modding locally that file (addon_utils.py) with future releases of Blender, because this works perfectly for my needs as it did in the last four years. I've never used that directory for SVN but only to keep order in the addons management, consistently with the three categories shown in the scripts catalog (Release, Contrib, and External).

Thank you Campbell.

In the meantime, I found out what have been deleted for some reason, and I modified the file named ..\2.73\scripts\modules\addon_utils.py reinstating the followings:

There should be a strategy to manage addons "without" messing up the \addons and \addons_contrib folders that I use to keep in sync with the ones on the GIT repository. Thank you and I look forward to read Campbell's reply too.

Here on Windows 7 Pro w/SP1, 64bit, Blender 2.73 official stable release (Hash: b4d8fb5) (zipped version) the Ivy Generator seems to work fine:

Thank you.