CMake 3.6.3 in MIPSpro
#11
RE: So this data type isn't complete...
I couldn't wait, tested at lunch. Confirmed! Moving that definition, and its dependencies, to the header was enough to remove the related errors! That one change dropped me from 56 errors to 14.
(This post was last modified: 04-23-2020, 03:45 PM by nintendoeats.)
nintendoeats
Octane

Trade Count: (0)
Posts: 85
Threads: 8
Joined: Nov 2019
Location: Canada
Find Reply
04-23-2020, 03:41 PM
#12
RE: So this data type isn't complete...
There is another possibly deeper problem. Cmake seems to compile a version of itself then runs a series of tests to see if it can complete the process. It fails many of these tests. More research required.
(This post was last modified: 04-24-2020, 01:26 AM by nintendoeats.)
nintendoeats
Octane

Trade Count: (0)
Posts: 85
Threads: 8
Joined: Nov 2019
Location: Canada
Find Reply
04-24-2020, 01:26 AM
#13
RE: So this data type isn't complete...
You may need to run it against dbx or cvd (the latter is the recommended one apparently on 6.5)

I'm the system admin of this site. Private security technician, licensed locksmith, hack of a c developer and vintage computer enthusiast. 

https://contrib.irixnet.org/raion/ -- contributions and pieces that I'm working on currently. 

https://codeberg.org/SolusRaion -- Code repos I control

Technical problems should be sent my way.
Raion
Chief IRIX Officer

Trade Count: (9)
Posts: 4,240
Threads: 533
Joined: Nov 2017
Location: Eastern Virginia
Website Find Reply
04-24-2020, 03:26 AM
#14
RE: CMake 3.6.3 in MIPSpro
The way this seems to work is that CMake compiles a limited version of itself (bootstrap CMake), then uses that to properly compile itself. Attempting to invoke the bootstrapped version gives you an angry message about how you shouldn't use it for anything except compiling CMake.

As part of the TRUE compile process, it runs a series of tests about your system. Many of these tests can fail, so long as they fail in the right way. Compilation was being halted by the fact that CHECK_TYPE_SIZE doesn't appear to work (I'm guessing the issue is with MIPSpro, but it could be related to MIPS3). I actually think this is a bug, since only afterwards does it run a test to see if it can determine the sizes of data types.  Really, it should be running those tests first, and then run the test for endianess only if they pass. I can bypass that issue to continue compiling by tweaking the module that generates the error, but it does mean that CMake will not have full functionality.


[Image: CMake_data_type_errors.png]
nintendoeats
Octane

Trade Count: (0)
Posts: 85
Threads: 8
Joined: Nov 2019
Location: Canada
Find Reply
04-24-2020, 11:18 PM


Forum Jump:


Users browsing this thread: 1 Guest(s)