Charles Galpin wrote:
> On Fri, 7 Apr 2000, p-thilts wrote:
> >
> > You said you were running 6.5.3-6 on a 6.1 box but it was not clear to me
> > whether you installed rpms downloaded from the Postgresql web site or if
> > they were included in the Red Hat vs 6.1 distribution you installed. I
> > suspect they were included in the 6.1 distribution because a number of people
> > responded saying Postgresql version was 6.5.3-x for Redhat 6.1 and 6.2.
> > Therefore maybe they are the same as those rpms on the Postgresql web site -
> > but not necessarily. Also, you did not indicate if you were running the
> > fully featured Postgresql with perl, tcl, servers and clients, etc.
> > Basically, that's why I left the question standing with Peter.
>
> whew. You are thinking way too hard about all this :)
>
> I happened to intall the 6.5.3-6.src.rpm from the 6.2
> distribution. Why? because I had a need for postgresql the other evening
> and that was handy (and I assumed pretty recent).
>
> Perhaps one thing that may not be clear to you is that *most* newer
> versions of rpms can be run just fine on older versions of Red Hat,
> *especially* if you build them on that platform. I would not be surprised
> if I could build that rpm on a 5.2 box and use them.
>
> So don't get hung up on versions. Especially if it's 6.x, any rpm out
> there should do just fine. Build from the src.rpm if ever unsure! It only
> takes a few minutes on decent hardware.
>
> But stop thinking so much about all this. Roll up your sleeves and give it
> a try!
>
> >
> > I'm relatively new to Linux and even newer to Red Hat distributions, and
> > tend to seek clarification anywhere I sense ambiguity. This comes from
> > years of experience in IT (Information Technology) where I had to justify
> > project recommendations based on research while not given the means to
> > experiment, prototype, and do hands on analysis. So, if the head and paper
> > analysis made a mistake which affected an approved project or it's timeline,
> > one's head was likely to be cut off and for sure one's credibility was
> > lost(forever). This is basically an apology! I'm really just getting used
> > to the 'try it first, worst that can happen is it bombs' approach to
> > computing that a lot of Linux hackers have taken. One has to be able to undo
> > what bombs. I tried that approach several times to my dismay. In one
> > instance I changed a pointer(at that time not even understanding the concept
> > of pointer) to a key library, and the whole system went down. I sweat blood
> > for hours and to this day can't remember how I fixed it. I still shudder
> > about that one. Lots of other disasters with trying things before I
> > understood what was happenning. I don't like to bother people needlessly
> > but I still sometimes miss things or don't look in the right places for
> > information, or don't assimilate properly what I've read because most of
> > everything is still new to me, but I do try.
>
> sounds like you need a test box. An old pentium would do.
>
> >
> > Anyway, YES, I would appreciate help on this, thankyou for offering, and I
> > will let you know when I run into difficulty. Note that I did not say: "IF I
> > run into difficulty". I can see it now, more pain on the way.
>
> it's really simple. do some reading about rpms if you can first though.
>
> > Bye-Thanks_TED
> >
> >
>
> --
> To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
> as the Subject.
Charles:
Just had to write you back and thank you.
So very kind of you to reply so promptly. With that last question answered I do
feel much more comfortable "rolling up my sleeves and doing something".
It made a world of a difference to me when you indicated that, for example, I
could take source rpms from a Red Hat 6.2 distribution and compile them on an
earlier distribution. In fact, I started to do just that (downloading 6.2 source
rpms to compile on 6.0). However, I was sure to read the 6.2 READMEs this time
and I did notice - as far as the distribution itself goes and initialization -
there are some heavy duty changes, specially (for a number of packages) the split
of server and client into 2 rpm packages where before there had been single
packages. But, I suppose that does not impact the compiling process.
Until now, I had thought that the source code was library focussed, that is, if
the right libraries were not included then compilation and/or linking would
fail. So, that problem is probably looked after by the symoblic links (which I
have a habit of calling pointers) and the name of the symbolic link is what is
used by the compile/link processing. I'll have to get inside the code and check
this all out. So (I'm still making an educated guess) the symbolic link name
points to the version or actual library in place on the distribution and it is
those calls for the actual library that get linked into the code. Hope this
assessment is correct. I'm assuming so for now but I'll check it out.
Don't get me wrong. I've done a fair amount of compiling and linking on Linux but
sometimes I've been flying half blind, basically doing the stuff with my fingers
crossed and as luck would have it, most often things worked. I'd make a really
bad airplane pilot - I just HATE flying blind, I want to see all around me as far
as the eye can see. (Have you ever walked into a brink wall in the dark -
Ouch!!!) That makes it hard on other people when they try to help me because to
them it seems I am just a chain of endless questions - maybe I am. I do like
your idea of having one machine for the purpose of just trying things out. I did
that once, and eventually it became my evaluation machine, loaded to the gills
with IDE's, JAVA, etc. So much work went into the stuff on that machine that I
dared not do anything that would damage the software resident on it. But this
time, I'm going to do it differently. I'm going to add a hard drive to one of the
machines and come "hell or high water" I will use it just to test distributions
and code, never letting work progress too far on it so that on another day I can
just scrap what is there and start all over again. That way I will be able to
pragmatically answer a lot of my own questions. Thanks for bringing that up! I
really should have resolved to do that much sooner. I'm fortunate to have the
facilities so I should be making the most of them -specially since I'm no longer
in that IT department any more and working for myself. Bad habits are sometimes
hard to change.
Also, I'll look more closely at the source rpms in a distribution as I think I
recall that most libraries are built from source just like applications. So if a
compile doesn't work there should be some kind of "Can't find..." error message
and I should be able to search rpm library source files for the missing parts and
then add the part to the library or add the library to the include list, or
whatever seems appropriate. This is the sort of thing I could do on that test
drive thing you recommended. Lot's of freedom once one does not have to worry
about damaging the system or application. A real good idea for me Charles -
thanks again.
Do have a very nice day.
Bye-thanks_TED
PS - I still like emacs better that vi even with the new 6.2 imbedded changes.
Peter does too.
Have a good one.
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
