Mark, as this does exactly what he says here and he is the primary (only current?) portage tester, then I would say go ahead and commit this unless you see something that gives you pause.

It looks thorough, he's taking advantage of the recent changes to detect command failures and such.

On Aug 3, 2007, at 5:14 PM, Eric Searcy wrote:
On Jul 30, 2007, at 5:32 PM, Jo Rhett wrote:

Ignore this and my missing attachment.  Eric will be sending a
revised one.

Here it is (attached). This revises the Portage package installation, and fixes to small other bugs in BuildCommandLine.

A few remarks:

- Verbose("BuildCommandLine(): Processing package %s at location % d.\n", package->name, &package ); + Verbose("BuildCommandLine(): Processing package %s at location % u.\n", package->name, &package );

This changes the printf placeholder from %d (decimal) to %u (unsigned). This keeps the addresses printed out positive, since memory addresses aren't signed.

+      ++cmd_args;

I put some code in place in my patch ``More buffer overflow fun'' to prevent passing too many arguments to cfpopen in BuildCommandLine, but neglected to increment the argument counter that was keeping track of this.

The rest only modifies code in PortagePackageCheck.

--
Eric Searcy
OSU Open Source Lab


Attachment: portage.patch
Description: Binary data

_______________________________________________
Bug-cfengine mailing list
[email protected]
https://cfengine.org/mailman/listinfo/bug-cfengine

--
Jo Rhett
senior geek

Silicon Valley Colocation
Support Phone: 408-400-0550




_______________________________________________
Bug-cfengine mailing list
[email protected]
https://cfengine.org/mailman/listinfo/bug-cfengine

Reply via email to