` around.
---
### Author:  [@ben.boeckel](https://discourse.cmake.org/u/ben.boeckel)
#### Post date: [September 4, 2021, 5:06pm UTC](https://discourse.cmake.org/t/transitive-dependencies-for-include-directories-break/4035/2 "2021-09-04T17:06:33Z")
They don’t break; you just can’t ask for them during the configure stage. What do you want this information for?
Edit: See also this question:
> [@Getting all PUBLIC & INTERFACE definitions including from dependencies](https://discourse.cmake.org/t/getting-all-public-interface-definitions-including-from-dependencies/3697):
>
> Hi. I have stumbled upon an use case where I think this might come handy, but I am not sure if possible in current CMake. My project specifies a target A which depends on targets B & C. For third party integration, I instruct users to specify the needed include, library paths and definitions. This day I got bitten by not specifying all definitions which created an API/ABI inconsistency (Compiled A target used different struct size than the users integration). To combat this, I was thinking of…
---
### Author:  [@Ronald](https://discourse.cmake.org/u/Ronald)
#### Post date: [September 4, 2021, 5:42pm UTC](https://discourse.cmake.org/t/transitive-dependencies-for-include-directories-break/4035/3 "2021-09-04T17:42:02Z")
> [@ben.boeckel](#):
>
> you just can’t ask for them during the configure stage
No, I have used `file(GENERATE ...)`, it can get genex values in generation stage.
---
### Author:  [@ben.boeckel](https://discourse.cmake.org/u/ben.boeckel)
#### Post date: [September 4, 2021, 7:57pm UTC](https://discourse.cmake.org/t/transitive-dependencies-for-include-directories-break/4035/4 "2021-09-04T19:57:14Z")
OK, that wasn’t clear. The thing is that `b` has a link dependency on `a`. When CMake generates the full tree for use in a compilation line, _that_ is when the full dependency tree is traversed and expanded. CMake cannot, in general, compute this full tree on-demand because there are ways of constructing them that depend on the target being linked. For example, this will make the include directories depend on whether an executable or library is consuming it:
```cmake
target_include_directories(foo
INTERFACE
"$<$,EXECUTABLE>:executable/only/include>")
```
There’s an issue about getting the transitive dependency closure of a target, but I can’t find it right now.
---
### Author:  [@Ronald](https://discourse.cmake.org/u/Ronald)
#### Post date: [September 5, 2021, 4:13am UTC](https://discourse.cmake.org/t/transitive-dependencies-for-include-directories-break/4035/5 "2021-09-05T04:13:58Z")
The `CMakeLists.txt` below for **app c** works as expected,
i.e. INTERFACE\_INCLUDE\_DIRECTORIES of a is in INCLUDE\_DIRECTORIES and `INTERFACE_INCLUDE_DIRECTORIES` of b
```cmake
find_package(a REQUIRED)
find_package(b REQUIRED)
dump_target_includes(b)
```
for comparison, the `CMakeLists.txt` below for **app c** doesn’t work as expected,
i.e. INTERFACE\_INCLUDE\_DIRECTORIES of a is **not** in INCLUDE\_DIRECTORIES or `INTERFACE_INCLUDE_DIRECTORIES` of b
```cmake
find_package(b REQUIRED)
dump_target_includes(b)
```
BTW, the revised version of `dump_target_includes`
```cmake
macro(dump_target_includes TARGET_NAME)
set(OUT_TMP "====dump_target_includes begin (${TARGET_NAME})====\n\n")
set(INCLUDE_DIRS $)
set(INTERFACE_INCLUDE_DIRS $)
string(CONCAT OUT_TMP ${OUT_TMP} "----INCLUDE_DIRECTORIES----\n")
string(CONCAT OUT_TMP ${OUT_TMP} $,\n>\n\n)
string(CONCAT OUT_TMP ${OUT_TMP} "----INTERFACE_INCLUDE_DIRECTORIES----\n")
string(CONCAT OUT_TMP ${OUT_TMP} $,\n>\n\n)
string(CONCAT OUT_TMP ${OUT_TMP} "====dump_target_includes end (${TARGET_NAME})====")
file(GENERATE OUTPUT test_debug_genex.log
CONTENT ${OUT_TMP})
endmacro()
```
---
### Author:  [@ben.boeckel](https://discourse.cmake.org/u/ben.boeckel)
#### Post date: [September 6, 2021, 2:21pm UTC](https://discourse.cmake.org/t/transitive-dependencies-for-include-directories-break/4035/6 "2021-09-06T14:21:14Z")
No, that’s because the include directories are added after analyzing the interface _link_ libraries of `b`. It sees `a` in that list and then traverses it. You just cannot get the full transitive closure using genexes today (and it’s unlikely to be implemented).