The collected works of irixpgmr

I am Gus's cohort in crime. I did get the source but I can't reveal how. It was under a strict promise to keep it a secret. However, I can tell you about the problems I had with it.

BTW, this is based off of source code I got in December 2004. This may be old news.

The main problem with compiling boinc on a non-linux/mac/sparc machine is that it is using c++ constructs that ANSI has deprecated. ANSI has dictated that 'using namespace std;' is not to be used any more. ANSI now wants you to only include the name spaces that you are using. For example, 'using namespace std::cout;'.

This wouldn't be too much of a problem but boinc has lots of source files and in the source files, they always have a line for using namespace in the deprecated way. This compiles fine with g++ because gnu compiler group decided to let people continue to use the old way. However, compilers based off of strict ANSI guidelines, have major problems. This is a huge problem when porting. Basically, you have to go through every file and figure out which namespace it is using and have only those namespaces included on the line. It is a big pain.

Don't get me wrong. I like einstein@home. I run it on all my machines except my SGI boxes. It just not viable to port at this time.
Does anyone have any ideas on how to get around the password length problem. I would like to increase my usable password length to more than 8 characters. Has anyone figured out how to do this?
foetz wrote:
you could crypt on your own and paste it into shadow...


Actually, I talked to a person at SGI. The problem is not the password program, it is the login program. Login only looks at the first 8 characters regardless of how many are set by passwd. He suggested that pam might help with this, but he wasn't sure. I have next to no exprerience setting up pam.

Perhaps there is another way, which I am unaware.
stuart wrote:
irixpgmr wrote:
Does anyone have any ideas on how to get around the password length problem. I would like to increase my usable password length to more than 8 characters. Has anyone figured out how to do this?


IIRC, the GUI password/secuirty tools limit you to 8 characters, but the good old UNIX "passwd" utility doesn't impose any arbitrary* length restrictions.

So, the best thing to do is probably to run "pwconv" (to setup/synchronise shadow passwords) and then use "passwd" to set the passwords for any accounts you need.

* although I think there is still a maximum of 255 characters

Actually, the login program truncates at 8 characters. I set the password to a password longer than 8 characters. I then put in the password with junk characters after the first 8 and I was logged in.
A while a go, I tried to get the client code ported to irix. What I found out was that the code was not strict ansi c++ compliant (which MIPSPro requires). The biggest problem is the use of the statement:

using namespace std;

When you use this in multiple source files, you can run into name space collisions. Due to the namespace collision problems, ANSI made the change that you only use the namespaces required. What should be done is including only the namespace you are using for that source file. For example, if you are using cout and endl, you should use the following statements:

using std::cout;
using std::endl;

GCC decided to not enforce it by default. Since most programmers are used to 'using namespace std;' method, it still gets used.

Changing each source file to use only the namespaces that are used clears up quite a few problema, but not all. Life got in the way and I never got the port finished.
I am running neko_ntpd on a 3200 running IRIX 6.5.30. I am getting the following error message in SYSLOG:
Mar 11 11:09:38 6D:turing ntpd[758]: adjusting local clock by 5148.060434s
Mar 11 11:09:38 2D:turing ntpd[758]: adjtime failed: Invalid argument

Has anybody seen this before? Does anybody have a fix for this?

Thanks much
nekonoko wrote:
Yes - it uses adjtime to "adjust the system's notion of the current time, as returned by gettimeofday(3C), advancing or retarding it by the amount of time specified in the struct timeval pointed to by delta." It's not designed for huge leaps; it was designed to make minor adjustments to the clock to keep it in sync by slowing/speeding up the clock.

Also keep in mind we're talking about openntpd here (neko_ntpd).

I reset the time to within 5 minutes of the correct time and it worked.

Thanks much!