Applied, thanks
ekutils: Allow certificate path relative to certificate list.
regtests: Allow setting PREFIX.
regtests: Count errors, don't exit on first error.
ekutils: Allow certificate paths relative only to list file.
regtests: Allow running from arbitrary location.
regtests: Move test data to a separate directory.
regtests: Remove useless WARN variable.
Remove .cvsignore files.
Also the whole stat dance is to not break lists that are relative to working directory. If that is not intended to work the code can be simplified.
I don't think an absolute path is really required with the current implementation. Relative to your working directory would likely work too. My point is that if somebody crafts and maintains a list file like this it should allow paths relative to the list so that it can be distributed easily.
Note: not tested on Windows
misc cleanups
Remove .cvsignore files.
ekutils: Allow certificate path relative to certificate list.
regtests: Allow setting PREFIX.
regtests: Remove useless WARN variable.
regtests: Allow running from arbitrary location.
regtests: Count errors, don't exit on first error.
regtests: Move test data to a separate directory.
regtests: Allow setting PREFIX.
regtests: Remove useless WARN variable.
regtests: Allow running from arbitrary location.
regtests: Move test data to a separate directory.
ekutils: Set freed pointer to NULL.
ekutils: Allow certificate path relative to certificate list.
regtest: Allow setting PREFIX.
regtest: Count errors, don't exit on first error.
regtest: Remove useless WARN variable.
regtests: Move test data to a separate directory.
regtest: Allow running from arbitrary location.
regtest: Allow running from arbitrary location.
regtest: Count errors, don't exit on first error.
regtest: Allow running from arbitrary location.
regtest: Remove useless WARN variable.
regtest: Count errors, don't exit on first error.
regtest: Remove useless WARN variable.
regtest: Allow running from arbitrary location.
regtest: Allow setting PREFIX.
regtest: Count errors, don't exit on first error.
regtest: Remove useless WARN variable.
regtest: Allow running from arbitrary location.
patch
Alternative patch
To make it completely clear: the use is free to set CFLAGS=-O0 or CFLAGS="-O6 -funroll-loops" and the configure script has no business overriding that.
yes, the way to set -O0 is to set CFLAGS=-O0 which is completely independent of --enable-debug
The reply from the gcc folks is that the warning explicitly says "possible" - as in it is a heuristic and not a certainty, and the heuristic evolves over time. In this particular case it makes sense to initialize the variable and it makes the code more readable, or at least not less readable. The compiler will most likely optimize away the assignment anyway so whatever.
yes, -O0 works better with debuggers. But it generates code different from default and from other flags specified by the user. If the compiler does the code generation correctly the semantic should be the same but the disassembly will not. Also the timing will change which metters for some bugs. So it may be desirable to biuld with -O0 at some times and not others. I can understand that you can think --enable-debug as a debug profile similar to what Visual Studio has which applies a complete set...
possibly uninitialized variable warning
patch
ibmtss is compiled with -O0
fix quoting of ncurses flags
We have autotools now for good or bad. Either way the default makefile works.
x86_64 is covered by xmview-os/xmview/defs_x86_64_um.h:#define NRllseek __NR_doesnotexist Attaching patch. The r_llseek is defined even when NRllseek == __NR_doesnotexist so it should compile. Not tested.
use of llseek om 64bit
tool names and man mage names do not match
nevermind It works when using autoreconf and not automake directly.
automake not working
Hello, On Sun, 09 Jul 2017 22:30:49 +0000 "Ken Goldman" kagoldman@users.sf.net wrote: The separate makefile.fedora is because 1. Each distro seems to have different packaging requirements. Indeed. And that's why you need one universal makefile that does not impose policy and allows the distributions to pass in different options. In the near term, I support a lot of users who are not "installing" the library. These are often people who don't understand how to untar, how to run make, can't install...
I would be happy to have proper build system in every release with proper library versioning. I sure can merge the makefile.fedora changes into makefile. The thing is why do you even need a separate makefile.fedora? It either builds equally working tss tools with versioned library so there is no reason to not do that by default or it is broken and then it should not exist. And in particular I do not want to figure out what version the library is supposed to be for every release. The version information...
makefile.fedora is not maintained
Different database support
license file?
ok, I will disregard the man pages which are the only difference in the 'withman'...
So fedora requires thier srpms to be in public place, this is supposed to be their...
man pages
thanks
Files included in release 886 marked as proprietary/NDA