[PC-BSD Testing] PCBSD 7.1-RC1 and Runports

Kris Moore kris at pcbsd.com
Mon Mar 30 07:47:13 PDT 2009

Ian Robinson wrote:
> Testing PCBSD 7.1-RC1 (i386)
> 1.  Runports failed.
> "The runports command is just a small shell script which unsets some
> variables, and changes the PATH in order to ensure that your port builds
> don't conflict with apps in /PCBSD/local incorrectly. "
> Although I updated both source and ports tree, the make command in runports
> failed at compiling orbit-idl-2 2.14.17:
> ==============Begin Error Messages============================
> orbit-idl-2 2.14.17 compiling mode, show preprocessor errors, passes:
> common
> Processing file
> ../../../include/orbit/orb-core/../../../src/orb/orb-core/orbit-interface.idl
> ../../../include/orbit/orb-core/../../../src/orb/orb-core/orbit-interface.idl:15:
> Error: `TTypeCode' undeclared identifier
> ** (orbit-idl-2:5778): WARNING **:
> ../../../include/orbit/orb-core/../../../src/orb/orb-core/orbit-interface.idl
> compilation failed
> gmake[4]: *** [../../../include/orbit/orb-core/orbit-interface.h] Error 1

This error looks like a bug in the orbit port, if it doesn't get fixed 
soon, you may want to contact the maintainer about it. Ports tend to 
break often, which is a pain I know.

> I'll keep updating ports each day and try again until I get it to
> complete.    Even though I was on a fast computer (dual Xeon 3.4 MHz w/ 4 Gb
> RAM & fast scsi disk) I spent 1 hour running "make clean" and six hours
> running "make" to the point where it failed.   Older computers will compile
> slower.
> While I am not daunted by this (I am a runports-breaks-survivor), and know
> that whatever is out of sync in the ports tree will get fixed, the whole
> runports process may be off-putting to users if the installation changes
> from a 20-30 minute install to a more than six hour process.  This will be
> especially true if they spend six-plus hours waiting to get to failure.
> Hopefully, PCBSD could offer an installation option that includes loading up
> the PCBSD ports.

The runports option is not intended for "regular" users, I.E casual 
desktop users. I wouldn't expect a desktop user to figure out how to 
compile firefox, or wait 8 hours for it to finish. Users like that 
should stick to PBI's or packages. The more advanced users who like 
using ports, will have to expect to run into bugs like this, since thats 
the nature of the FreeBSD ports tree, especially when you are using a 
current version of it.

> 2.  System Manager displays negative amount of memory.  This is a trivial
> bug.  After entering the System Manager, the display window with the PCBSD
> Logo and the option tabs displays the system memory as -805351424B.
> dmesg reports real memory  = 3489615872 (3327 MB), avail memory = 3401822208
> (3244 MB)

Thats weird!

Can you run this:

# sysctl hw.realmem

Thats what the system manager is using to pull the available memory on 
the system. Maybe our function is failing to read that int properly.


Kris Moore
PC-BSD Software

More information about the Testing mailing list