Hi, I just discovered your netradiant branch. As a radiant contributor I try to reduce the most possible the diff between living radiant branch. That's why for example I sometime backport stuff from NetRadiant to GtkRadiant or I import stuff from GtkRadiant to NetRadiant. I also do that with other trees, for example I recently imported in NetRadiant the iqm model support from AARadiant.

I maintain a tool to help me in comparing various radiant forks. Also, take a look at that massive list of various radiant I referenced! I hope one day seeing people being lost in all these trees will be a thing of the past. So every time I can make radiant trees at least partially merging together I'm an happy man, and that's good for everyone of us. 😀

So I was very surprised to discover your heavily squashed history… It's very very hard to read the change, not only because of the commit titles but also because it appeared they all are squashed commits from a the merged branch of a very recent (like 3 month ago) master tree and another custom tree, making original commits from this custom tree impossible to identify and review. Also these squashed commits mix different unrelated changes, making again review harder.

Just as an example, this commit just merge one of mine plus one of yours (renaming an option and activating it by default), making the read of your diff above mine almost impossible to recover, and I found this one totally randomly since even the commit message was rewritten with a different wording…

I just noticed some problems in your tree that are now fixed upstream (like silent bugs which does not break build and does not crash software but produces garbage in files), but it's now very hard to backport the fixes due to the rewritten history. 😯

I noticed the master-backup (commit) branch it looks like it's just upstream history up to 3 month ago but in fact squashing happened after a commit from 2014. Do you have somewhere a backup of the branch resulting from the merge of the master one and the custom one before the massive three-year commit squash? Can you upload it somewhere? 😃

Using my radiant comparison tool the diff noise between upstream and your history is already 24000 line long (probably also because of all the heavy changes that were made upstream since july). Without that merged tree before the commit squash It looks very hard to backport stuff across each other trees (and I'm very afraid about fixes), also it now makes automatic merging very hard… So, do you have a backup of this merged but unsquashed tree you can push in a dedicated branch or upload somewhere? Or just the custom branch before the merge?

Thanks in advance! 😃

I have tried to compile this on macOS 10.15.3 and am encountering multiple errors;

tools/quake3/common/vfs.h:39:10: fatal error: 'linux/limits.h' file not found
This one is easily fixable, by changing linux/limits.h to just limits.h

After fixing that, it will build but fail to link with the following error:

Undefined symbols for architecture x86_64:
  "_stricmp", referenced from:
      _PC_PushScript in l_precomp.o
      _PC_Directive_include in l_precomp.o
      _LoadCfgFile in aas_cfg.o
      _OutputFile in bspc.o
      _CreateAASFilesForAllBSPFiles in bspc.o
      _main in bspc.o
      _QuakeFileExtensionType in l_qfiles.o
      ...
     (maybe you meant: _Q_stricmp)
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [install/mbspc.x86_64] Error 1

I am using brew to install my dependencies, and they are all up to date.

Sometimes, when working on a map, it irkes me a bit not being able to hide both common/botclip and common/botdonotenter brushes. I can use Filter -> Botclips to hide the former, but not the latter.

It would be really cool if Filter -> Botclips could also hide these brushes. Or if an additional Filter -> Botdonotenters command could be added.

When i trying to compile the teminal show me this error:

In file included from tools/mbspc/botlib/../qcommon/q_shared.h:104,
from tools/mbspc/botlib/be_aas_bspq3.c:32:
tools/mbspc/botlib/../qcommon/q_platform.h:184:3: error: #error ARCH_STRING should be defined by the Makefile

error ARCH_STRING should be defined by the Makefile

^~~~~
tools/mbspc/botlib/../qcommon/q_platform.h:309:2: error: #error "Architecture not supported"
#error "Architecture not supported"
^~~~~
make: *** [Makefile:517: tools/mbspc/botlib/be_aas_bspq3.o] Error 1

Hello im compiling this on an arm device and get this error message:

tools/mbspc/botlib/../qcommon/q_platform.h:311:2: error: #error "Architecture not supported" #error "Architecture not supported"

im guessing its possible to fix this as i have compiled netradiant on arm before, any idea? :D

SIncerally
Ztealmax

*** UPDATE ***
Commit: aa44982 worked thanx you

Hey,
I was going through the feature list and I saw a feature called autocaulk.

What is it?

When changing from object mode to face/edge/vertex mode the former mode selection remains selected. It is a bit annoying when you forget to deselect before changing because you can end up with both the brush and a face/etc selected so you have to deselect all. It's a minor thing, but it would feel better if changed.

I don't think there is any use for having hybrid selections? I've never had a need for it, at least.

It probably shouldn't deselect right at changing, but at selecting the first thing in the new mode.

Essentially what the title says. In Gtk it pops up with a drop down list of configured games to select - I switch between Q2, Q3 and QL often so it's cumbersome to change game from the settings menu.

That feature is necesary to know how the lights will looks before compile de maps.

I have downloaded a mod, this one: https://www.moddb.com/mods/uber-arena/ and I want to create some maps for it, however, the mod won't load the entities.ent or entities.def file, even if both are well formatted, so I cannot use the mod-exclusive entities.

These are the mod's entity files, for testing:
https://www.dropbox.com/s/ffrac91xegxx36n/entities.def?dl=0
https://www.dropbox.com/s/s3p5hawn7959sux/entities.ent?dl=0

Hi. We already mentioned this on Discord, I just open this "issue" to have it in a more handy place.

One of the nice features of NetRadiant Custom is the ability to easily customize keyboard shortcuts.
But considering it's also easy to make a mess doing it, it may be nice to have the ability to restore things to default.
While it is possible to relatively easily restore all binding defaults by manually deleting shortcuts.ini file, not everybody may think about that, and it's an "all-or-nothing" thing.

So, I think it may be useful to have a button which would allow to reset only the highlighted command to its default key. In case that key (or key combination) is already used by some other command, it should show the dialog window to choose what to do, just like you changed it normally.

Maybe a further evolution may be, in case one clicks the button with no command selected, ask for confirm to reset ALL keys at once, but that's not as important I think.

The text in the 2D and camera windows (brush/entity selection size, class and target names) is so small, it's barely readable on my resolution. It used to be fine, but then I reinstalled Windows 10 and some text scale setting it previously had must have been reset (or different, no idea how I did it before). I use 1920x1080 and apparently a custom scale of 115%. In Netradiant, the menus are fine, but the custom scale does not affect the text inside those windows. Could be another case of outdated SDK stuff? It's not a super critical issue, but it would nice if there was a way to increase the text size. I would have expected "Misc/Colors/OpenGL font" did that, but it seems it only affects the grid numbers.

Good morning.

Noting that NetRadiant supports graphic themes (including my favorite dark ones ...) I figured I could create my own theme for myself and share with others too. Of course, we are talking about a dark theme. It is true that there are a few already on the program, but I would like to do my own. (My eyes refused to obey many times when I was sitting at the monitors for a long time).

And hence my question, Is there any way to create a theme file that is in the "share / themes / ..." directory?

What kind of file is it, is it just bitmaps in certain colors, or maybe compiled (configuration) files with variable RGB values something like an Android application? Or something completely different?

I am not sure if it is related to the current release or possibly older:

"Correct"

"Wrong"

Map files:
"etf_desire_backup.map" is the fallback that still works when unsaved.
Upon using save as to desire2.map and reloading that file, it looks like the bottom screenshot.
desire.zip

Hi, this is a feature request as cod1\2\4\waw\BO3 does not have OBJ export functionality and I feel that this version of Radiant is far superior to these older versions in many ways. I'd rather make a level in this version then export to the other radiants for compiling. I can supply brushes\meshes\patches etc if you want to accept this as an incorportated feature in a future release.
thanks :)

Netradiant can already render maps in wireframe mode. Can this be extended to filling in the faces with flat colours or textures to give the effect of panels with outlines?

Then allow this to be compiled to bsp without the need to manually create face specific textures in a paint editor. Creating the outlines for each face texture and then aligning them is painful and inaccurate.

The result in the game will be a total (or partially) tiled map in the shape of the faces, emphasizing the brush geometry.

I hope this idea appeals to you.

So yeah, I hit the 0.1% use case lottery, it's just with bin, it will only trigger the issue by naming the root directory bin.

Naming the root directory where radiant.exe anything else than bin solves this issue.

Seems to be related to #17

The libgdk_pixbuf bundled with the latest release 20210105-win64 is 2.40.0, while newer, it doesn't seems to work producing the "no icon graphics" bug.

A workaround that seems to work is to take the libgdk_pixbuf bundled with 20200514-win64 that is 2.32.1.0 and replace the other with this.

I propose a new hotfix package (with the older libgdk_pixbuf) to be released as a temporal solution to this issue before the next release, just to prevent confusion.

It would be good to look into why a newer version of libgdk_pixbuf is not working properly, then as a real fix for this.

This fork works great for editing Wolfenstein: Enemy Territory maps, too. I've used the Xonotic Netradiant before (https://github.com/xonotic/netradiant) but this one here runs more smooth and stable IMHO. But I'm missing a very important feature which is: regrouping/ungrouping of Entities.

For example: you have a func_group of brushes and patches and want to add some brushes from another func_group. In the other Netradiant, I select the whole source entity with SHIFT+Brush and CTRL+ALT+E and then the Brush(es) I want to add holding SHIFT, clicking them and just select Entity > Regroup which adds them to the source func_group.

In this version, there are no such options. I know there is Entity > Move primitives to Entity but it works different. The selection process is the same, but I can't add a single brush from another group as the whole group gets added when I selected single brushes. The only workaround is to "worldspawn" these brushes first, creating a separate func_group for them which I then add to the other func_group.

Perhaps you can add such an option as a new feature which is very useful.

I'm working on some gargantuan terrain maps, and would like to be able to set the grid to larger sizes than 256, which is the current maximum. Perhaps adding grid sizes 512, 1024, 2048 and 4096 would be adequate.

Got this error trying to run the latest release build (downloaded from here, not built myself):

[netradiant-custom-20200320]$ ./radiant.x86_64
./radiant.x86_64: symbol lookup error: ./radiant.x86_64: undefined symbol: gdk_gl_font_use_pango_font

Pretty sure I have all the needed libraries installed, is this maybe about a deprecated function?

Problem: loading a .pts file only works if the pts file was written to the folder containing the .map file. Compilers however, tend to write the .pts file to the same folder where they write the .bsp file. This is often different to the folder containing the map file.

Solution: The .pts loading should look for the pointfile in the folder where the .bsp was written to first.

"surfaceparm origin" had to be added to the "system.shader" for the origin texture to get it to work properly. This only had "q3map_origin" which wasn't sufficient.

In GTK Radiant, you can drag non-brush entities around without directly dragging from the entity itself by dragging surrounding areas in the grid or 3d space. This allows you to drag entities without having them directly visible after moving your camera out of range in the view port, which in some ways can speed up level design by allowing you to be less precise and not needing to use the translation widget or directly hitting the entity in order to drag it around in the grid.

https://user-images.githubusercontent.com/3003653/110164392-edad2780-7db6-11eb-9c28-5e290eea5808.mp4 (GTK Radiant dragging behavior)

https://user-images.githubusercontent.com/3003653/110164815-85127a80-7db7-11eb-8e56-093057c42c64.mp4 (Netradiant behavior)

I'm so happy to discover this custom radiant (2018-05-16 v3.1.0 build 9.17.10.4459). Solved major texture mapping issues and bug/crashes from Ingars radiant, but was so disappointed to find all in vain since I need more than f.xx Arbitrary Rotations precision for my precision geometrical maps. This was never an issue before. Please recompile a version that allows greater precision (f.xxx at the VERY least). I would if do it myself if I could. This would then be the perfect mapping program. Thanks for the good work!

Using the release build of netradiant-custom-ubuntu-18-04-x86_64. (BTW what is ubuntu specific about this build??) netradiant-custom-20181213 on Manjaro Linux KDE with Nvidia proprietary driver.

When looking at q3dm17sample.map these giant white squares appear over parts of the grid view and in 3d view as you move close to a displayed model in view making it a bad experience. The corner of the squares seems to start from the model's origin. Here is a small video I made demonstrating the issue.

Let me know if there is any further info you need. I will reply when I can.

In the camera/3D view, light radiuses tend to get in the way (especially bad if you are trying to work with many together). It would be nice to either have the radius more transparent so you can still see the environment or an option to adjust opacity in preferences.

This is technically two feature requests, but under the same related concept so I am lumping them together:

1. A short term memory for the last selection of multiple items (replaced every time you select multiple items). So for instance you select a bunch of lights, then deselect them to move a single a brush. With this feature you could then hit a key to reselect the lights without having to do so manually.
2. Persist selection of items along with undo/redo history. For example if I'm adjusting a layout with a large mix of brushes and entities, only to find out I made a mistake and want to undo, I will lose the prior selection. With this feature that would be retained and you would be able to continue moving the items after undoing without wasting time reselecting everything.
3. A hybrid of 1 and 2: Keep a selection history (separate but parallel to regular undo/redo) and make it possible to step back/forward through it.

With 2 I know you can use func_groups to make related brush selection faster, but that is not always viable, especially if working with entities.

If you setup a light with a negative light value (to darken an area), the value will be lost if the entity is moved. On first move it is changed to NaN, then it is reset to the default of 300.

shader example:

dirt-rock { qer_editorImage textures/e1m2/e1m2_drt1_1_l { } }

Hey, i am trying to compose an rtcw pak for netradiant, but seems like netradiant can't read entity definition (.def) file, i am not sure what's the issue but nor netradiant-custom or original netradiant can load the .def, but only the .ent files, despite the fact the code to load the .def files is in place.
However, original radiant 1.5 can do that without any problems.
The game definition file has entityclasstype="def" set, but for some reason it doesn't do anything.

Here is the entity and game definition files i can't make to work, i hope you can figure this out.

games.zip

Hi, after 20+ years using Q3/Gtk Radiant I'm trying to make the jump to this. The problem is that after so many years, I'm very much used to Gtk's controls and find myself stumbling about in Net trying to do what I normally can in Gtk.

I think it would attract a lot of attention from the many folks like me still stuck in Gtk to make the move over to Net if there was a config/preset option to make it behave more similar to Gtk. Working with func_groups and viewport navigation seem to be the two biggest barriers for me at the moment.

Should be same as CONTENTS_SLAG

Useful for Enemy Territory Fortress mod which uses basically Q3 maps anyway whether it be map file recompile or bsp to bsp ports.

Windows 10 just followed all instructions for Win32 - MSYS2 64 bit.
The result is this:

The splash screen is also a missing icon, I checked and all the icon files are correctly placed under bitmaps/ folder but they don't display in the program.

Using Linux Mint 17.3 Cinnamon. Have mingw32 installed.

I wanted to use the new, cool features in your version of NetRadiant, but cross-compiling to Win32 via mingw outputted the following error:

`i586-mingw32msvc-gcc tools/quake2/qdata/images.c  -MMD -W -Wall -Wcast-align -Wcast-qual -Wno-unused-parameter -fno-strict-aliasing -g -O -mms-bitfields -I/$HOME/netradiant-dependencies-mingw32/radiantdeps/include/libxml2   -Itools/quake2/common -Ilibs -Iinclude   -DWIN32 -D_WIN32 -D_inline=inline -I/$HOME/netradiant-dependencies-mingw32//radiantdeps/include -I/$HOME/netradiant-dependencies-mingw32//gtk/include -DRADIANT_VERSION="\"1.5.0n-git-e3d6c0a\"" -DRADIANT_MAJOR_VERSION="\"5\"" -DRADIANT_MINOR_VERSION="\"0\"" -DRADIANT_ABOUTMSG="\"Custom build\"" -DQ3MAP_VERSION="\"2.5.17n-git-e3d6c0a\"" -DRADIANT_EXECUTABLE="\"exe\""  -c -o tools/quake2/qdata/images.o`
`tools/quake2/qdata/images.c:1: error: stray ‘\357’ in program`
`tools/quake2/qdata/images.c:1: error: stray ‘\273’ in program`
`tools/quake2/qdata/images.c:1: error: stray ‘\277’ in program`
`In file included from tools/quake2/qdata/qdata.h:28,`
            `from tools/quake2/qdata/images.c:22:`
`tools/quake2/common/cmdlib.h:44: warning: ignoring #pragma intrinsic `
`make: *** [tools/quake2/qdata/images.o] Error 1`
`rm icons/q2map.rc`

I'm using a custom cross-compile file in order not to mess up with the one from the repo, which has this content:

`# Usage:`
`# copy this file to Makefile.conf`
`# edit the paths in here`

`# how to call the mingw tools`
`MINGW_EXEC_PREFIX = i586-mingw32msvc-`

`# use mingw32`
`include mingw-Makefile.inc`
`include mingw-radiantdeps-Makefile.inc`

How do I fix this?

Its very confusing.

I like the rest of your hard work so far.

Both jk2 and jka use system/trigger for triggers and system/caulk for caulk, this should be added to the gamepacks.

Also nodraw could be added but I don't know any use case to test it.

games.zip (fixed .game fields)

Immediately after deactivating Edges mode after having modified a brush, Edges mode stays active until after a second click, resulting in brushes continuing to change shape and sometimes deforming into bad shapes. If this is an intended feature, an option to turn it off would be appreciated.

https://user-images.githubusercontent.com/3003653/110163077-37950e00-7db5-11eb-8b61-4c0f43ee5572.mp4 (Netradiant brush deformation error/behavior)

https://user-images.githubusercontent.com/3003653/110163084-3a8ffe80-7db5-11eb-9382-abd56cd7b7e5.mp4 (GTK Radiant 1.6 behavior)

https://user-images.githubusercontent.com/3003653/110163090-3d8aef00-7db5-11eb-8253-d515d51081dd.mp4 (Requiring a second click to drag both faces as normal)

Hello.
Is it possible to turn off texture transparency (common) in the 3D window and why only some textures are transparent?
I can hardly see the borders of the brushes.

A simple feature that I miss from using GTK Radiant 1.6.X.
Makes it easier to find leaks and structs where they should not be, as well as get a general layout of the space at a glance.

This issue is not present on legacy nvidia cards + win7 or AMD cards + win7.

I'm unable to test current nvidia cards with win10 at the moment.

This is something that bothers me in all radiants, and this one is the one I like the most and seems the most active so I'm creating the suggestion here.

When I use radiant, I often find myself using W and R to toggle translate and rotate tools, most of the time I just need to do one action with this tool and then return back to normal, right now I need to go back to normal manually after every time and I often forget and end up doing wrong actions with the wrong tool.

I suggest an option that would make it work kinda like blender, where the tool mode automatically resets back to normal after performing an action, so for example I can hit R, do the rotation, and it returns back to normal immediately after I've done it. For me this would make it much smoother and faster to do basic mapping especially when placing lots of misc_models that I use often in my work.

Would it ever be possible to build a config for the cod series (CoD4, WaW, BO1, etc...)? If so i would really love to edit my maps with this radiant.
Or is there a tutorial of how to build a config for a specific game out there?

Thanks in advance!

Introduced in 23d2497:

Those types are obviously not the sames.

It produces that compilation error:

/home/illwieckz/dev/netradiant/radiant/texwindow.cpp: In function ‘gboolean TextureBrowser_button_press(ui::Widget, GdkEventButton*, TextureBrowser*)’:
/home/illwieckz/dev/netradiant/radiant/texwindow.cpp:1532:22: error: invalid conversion from ‘const char*’ to ‘char*’ [-fpermissive]
   char* dir = strrchr( sh, '/' );
               ~~~~~~~^~~~~~~~~~~
/home/illwieckz/dev/netradiant/radiant/texwindow.cpp:1535:16: error: invalid conversion from ‘const char*’ to ‘char*’ [-fpermissive]
    dir = strchr( sh, '/' );
          ~~~~~~^~~~~~~~~~~

Current algorithm is technically wrong, but covers good amount of cases more or less right.
In short it is: find avg axes of patch out of its points, find longest column and row, find uv projection space out of triangle, including these col and row, placed on manipulator's plane along avg axes.
Apparently this doesn't always turn out well, examples:
unwanted skew:

skewed and sizing is odd:

is fine, when texture is placed in 'natural' style, previous was projected

odd skew and sizing ratio in the very classic case:

'natural' may be odd too:

Wished qualities of unwrapping are:

• having intuitively alright angle between texture axes
• having right texture axes sizes (this is also important for expected results after rotation)
• preserving manipulator representation after affine transformations of uv space (this is good to keep transform pivot consistent relative to manipulator in uv space and to have expected result during and after transformation) (representation change after moving some UVs separately is acceptable imo)
• consistent representation w/o saving some state in the tool; this is good to have same look after reselecting a patch

Current implementation fulfills 3 and 4, 1 and 2 aren't always like wished.

Error when building against the latest gtkglext, it would appear support for gdk_gl_font_use_pango_font has been dropped.

libs/gtkutil/glfont.cpp: In function ‘PangoFont* tryFont(const char*, PangoFontDescription*&, GLuint)’:
libs/gtkutil/glfont.cpp:678:9: error: ‘gdk_gl_font_use_pango_font’ was not declared in this scope
  678 |  return gdk_gl_font_use_pango_font( font_desc, 0, 256, font_list_base );
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~

Here is the discussion on the Arch mailing list.
https://lists.archlinux.org/pipermail/arch-dev-public/2020-April/029956.html

When packing external lightmaps, q3map2 creates the lightmap UVs and surface lightstyles properly but does not save the created lightmap UV coordinates into the BSP. q3map_lightmapsize 1024 is used to force external lightmaps.

Is -extlmhacksize needed on top of -styles / -style or is there anything else missing to get this feature to work?

Writing hacky shader lightstyles as q3map2 would do seems broken too.

When writing a shader with q3map_stylemarker the autogenerated mapshader does not include any of the styled lightmap stages with the appropriate alphaGen and rgbGen waveforms from _styleNrgbGen and _styleNAlphaGen in the worldspawn or any other entity.

Maybe there is just a little bug in how the UVs are stored and they are discarded by accident. However the resulting bsp cannot use tcGen lightmap as the coordinatesystem for them has not been exported along creating artifacts and bugs with the renderer in Raven Games.

Some games (e.g. Jedi Academy) do not use textures/common for core editor shaders and instead use textures/system. This is properly defined in the .game configuration, but netradiant ignores this and always assumes the common path for certain operations (automatic caulking, applying trigger when creating a trigger, possibly others but those are what I am aware of).

Secondly, it would be nice to be able to turn off the automatic application of the trigger shader on entity creation (for cases where the mapper wants to use a custom one), just like there is an option to turn off automatic caulking.

Hello,

Some Tremulous mappers started to use your netradiant fork relatively recently with good success. One of those mappers, @BillyRazOr , gave me a .zip of his setup for Tremulous mapping, here is the link with his modifications for your review: https://www.dropbox.com/s/8bnpocvahs6kc08/netradiant-custom-win64.zip?dl=0

Would you include support for Tremulous mapping out of the box in the releases? If that would be the case, we have some mappers who would be eager in helping out with testing and offer feedback.

Hi! Compiled the latest Github version of NetRadiant. How can I enable the grid numbers in the 2D views? I need them to do some adjustments, but I can't find the option.

If those numbers were removed, can they be brought back?