C++ header files basically guarantee a slow build due to reparsing. Precompiled headers only help if you don't change a prototype.
If you could have C++ but pull class/function prototypes out of the object files like with Java or C# it would fix a lot of build time problems as you would never need to reparse a header.
What about the compile server? Is that an alternative to PCH?
IMO the problem with PCH is that you end up with all your.cpp's dependant on all your headers - thus stuffing up any attempt at doing correct minimal includes and auto dependency tracking from make.
I know J2EE developers that could just barely manage to install J2SE+J2EE properly on their local machine on a good day. You can not expect these people to adopt XDoclet/Ant/CVS/ArgoUML/etc and make it all function productively. You can expect them to write perfectly good business applications from within an integrated environment.
True enough. The JBoss IDE effort for eclipse might be off use for those people. I think eclipse also has remote debugger support, CVS integration et al. Being more of an emacs/command line type person I haven't really looked into it that much.
JBoss is not threat to the big J2EE vendors. Implementing a single server side class in J2EE requires writing at least three separate bits of Java code for the home, remote and bean interfaces/classes. There may also be local variants of these to overcome marshalling overhead. XML metadata must also be maintained. This is for a single EJB.
XDoclet makes this really easy - you only have one source file and the rest are generated.
The main problem with redhat upgrades is that there online tools (e.g. up2date) are pay-for and only support patching, not release upgrades.
I've been using apt-rpm from Fresh RPMS for a while now, which doesn't have these problems.
Apt-rpm provides a free solution, with an open back end (so that you can publish your own packages - for example gstreamer use apt-rpm in this way) and works really well. The alternatives, up2date and red-carpet, have closed back ends, you have to get redhat or ximian to publish your package for you.
Perhap we should go to Red Hat Bugzilla and raise enhancment requests for then to support this?
The Linux release is a version of the headless Rational Realtime Purify. You still need a Windows or Unix version of Purify to actually use it.
After using both purify and valgrind, I think that valgrind is much better. IMO Rational have already lost the Linux market to valgrind. This is too little too late.
Does CLX still require the borlandc C++ extensions?
One problem with VCL is that it has "properties" which are a borland only extension which means that you can't build code that uses VCL headers with any other compiler.
I use CVS daily and think its excellent. However there a number useful of patches with haven't made it into the main distribution literally for years. e.g.
* alternate port number patch * proxy tunneling patch * LOCAL_BRANCH patch
Is anyone reading this involved in the development of CVS itself and able to comment on when these patches will make it in?
Stallman is right... GPL encourages sharing
on
Thus Spake Stallman
·
· Score: 2
His clear and principaled stance on free software is needed in this muddled world.
I have published software under the GPL, and had to remind those porting it to Amiga and Archimedes (hey it was a while ago!) to make source available.
The whole concept of sharing had never occured them before. I felt that the GPL had accomplished something then.
Using the standard GPL/LGPL saves reinvention of the wheel. All these companies that feel a pressing need to invent their own license are in fact often restricting the utility of their code as it prevents integration into a GPL'd whole.
E.g. you would find it difficult to put an MPL'd html widget inside a GPL'd application without a license conflict.
The (L)GPL and BSD licenses are standard, so use them.
I hope Office (aka Steam) Power stays on the DRS6000:) I remember that you could only send email to an internet address if you set up about three screens (per address!) of settings first.
Slashdot Mirror
C++ header files basically guarantee a slow build due to reparsing. Precompiled headers only help if you don't change a prototype.
If you could have C++ but pull class/function prototypes out of the object files like with Java or C# it would fix a lot of build time problems as you would never need to reparse a header.
http://gcc.gnu.org/wiki/IncrementalCompiler may be of use if you still hold out hope of faster C+ builds...
Er, the thames already does that ;^)
Any chance of seeing a comparison of Solaris (x86 presumably) on the same hardware.
An HP-UX comparison would also be useful although you couldn't be sure the hardware was equivalent.
The existing VCL seems to use Pascal strings (aka AnsiString) and some Dephisms like the "property" keyword.
To use non-borland C++ compilers this would need a new genuinely C++ based VCL equivalent. If there is - how hard will be upgrade be?!
Er, thats not right. SWT is ported to GTK, Motif and OSX. Eclipse works on all those environments.
Oh, and SWT is open source whereas swing isn't.
What about the compile server? Is that an alternative to PCH?
.cpp's dependant on all your headers - thus stuffing up any attempt at doing correct minimal includes and auto dependency tracking from make.
IMO the problem with PCH is that you end up with all your
I'd like SWT to be released separately from eclipse so that I can develop/distribute SWT based apps easily.
Also an SWT gui builder wouldn't go amiss....
I first saw this story on the inquirer - See:
The article and
my letter in response
I've been using apt-rpm from Fresh RPMS for a while now, which doesn't have these problems. Apt-rpm provides a free solution, with an open back end (so that you can publish your own packages - for example gstreamer use apt-rpm in this way) and works really well. The alternatives, up2date and red-carpet, have closed back ends, you have to get redhat or ximian to publish your package for you.
Perhap we should go to Red Hat Bugzilla and raise enhancment requests for then to support this?
Linux won't be able to ship with Java/Gtk by default until Sun open source the jdk.
.NET environment (from mono), but not a Java environment.
If they don't do anything you will have the weird situation of Red Hat 9.x shipping with a
I know about gcj etc, but to be able to run Apache Tomcat you really need a Sun derived JDK.
Its a really handy tool ported from debian. See http://apt.freshrpms.net
Now if only Red Hat would adopt it instead of up2date...
Redhat should include it by default and make their updates available over it.
The Linux release is a version of the headless Rational Realtime Purify. You still need a Windows or Unix version of Purify to actually use it.
After using both purify and valgrind, I think that valgrind is much better. IMO Rational have already lost the Linux market to valgrind. This is too little too late.
Does CLX still require the borlandc C++ extensions?
One problem with VCL is that it has "properties" which are a borland only extension which means that you can't build code that uses VCL headers with any other compiler.
With exchange the calendar is shared between users, so you can do things like schedule meetings without having to phone round first.
The Evolution calendar stuff has all the features needed except that your calendar isn't accessible to other users.
Hopefully someone will write a free iCal server and an evolution backend to it.
indeed :) If only std::string had been common place when the spec was written
I use CVS daily and think its excellent. However there a number useful of patches with haven't made it into the main distribution literally for years.
e.g.
* alternate port number patch
* proxy tunneling patch
* LOCAL_BRANCH patch
Is anyone reading this involved in the development of CVS itself and able to comment on when these patches will make it in?
His clear and principaled stance on free software is needed in this muddled world.
I have published software under the GPL, and had to remind those porting it to Amiga and Archimedes (hey it was a while ago!) to make source available.
The whole concept of sharing had never occured them before. I felt that the GPL had accomplished something then.
And to think I recommended Nvidia cards to two people who just bought PCs.
I'll be recommending someone else's cards next time.
See http://www.lunatech.com/research/corba/cope/
Using the standard GPL/LGPL saves reinvention of the wheel. All these companies that feel a pressing need to invent their own license are in fact often restricting the utility of their code as it prevents integration into a GPL'd whole.
E.g. you would find it difficult to put an MPL'd html widget inside a GPL'd application without a license conflict.
The (L)GPL and BSD licenses are standard, so use them.
I hope Office (aka Steam) Power stays on the DRS6000 :) I remember that you could only send email to an internet address if you set up about three screens (per address!) of settings first.
An ex-ICLer
Hopefully they'll put some guys on to precompiled headers and a working -frepo switch for C++.
Long compiles are a real headache.
Alex.