[comp.lang.c++] ET++, strange files in distribution

euaeny@eua.ericsson.se (Erik Nyquist) (11/09/90)

Hallo world! The current version of ET++ that has been distributed
seems not to work. Without any success at all I have tried to build
it.  Many people that has sent mail to me seems to have had the same
problems as I have had.

I would be grateful if Erich Gamma or Andre Weinand why it is so difficult
to build ET++ this time. 

I also found some rather strange files in the distribution:
et2/src/PathLookup.h contained this:
---------------
...
class PathLdsfsdfsdfsdhfkjsdfkjsdhfksdhfjsf`
while true do echo blabla
ich tkdkd sadjflf gehdfs fddsfsdfkdf sdfklsjdfjfjfdfkjsaflsjdf
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
jkjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj
...
ookup: public Object {
friend class PathIter;
    char *path;
...
--------------------
Why does this file contains this stuff? The file is OK to use because
macros prevents that stuff from being include, but isn't it a rather
strange programming style to have this junk in the file. ;-)

Erik Nyquist
Ellemtel Utvecklings AB
Box 1505
S-125 25 Stockholm
Email: Erik.Nyquist@eua.ericsson.se

curt@meaddata.com (Curt Kohler) (11/09/90)

In article <1990Nov8.181319.14703@eua.ericsson.se>,
euaeny@eua.ericsson.se (Erik Nyquist) writes:

...

|> I would be grateful if Erich Gamma or Andre Weinand why it is so difficult
|> to build ET++ this time. 
|> 
|> I also found some rather strange files in the distribution:
|> et2/src/PathLookup.h contained this:
|> ---------------
|> ...
|> class PathLdsfsdfsdfsdhfkjsdfkjsdhfksdhfjsf`
|> while true do echo blabla
|> ich tkdkd sadjflf gehdfs fddsfsdfkdf sdfklsjdfjfjfdfkjsaflsjdf
|> kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
|> kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
|> jkjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj
|> ...
|> ookup: public Object {
|> friend class PathIter;
|>     char *path;
|> ...
|> --------------------
|> Why does this file contains this stuff? The file is OK to use because
|> macros prevents that stuff from being include, but isn't it a rather
|> strange programming style to have this junk in the file. ;-)
|> 

...

It looks like you may have corrupted version of the source.  I recently
obtained version 2.0 from iamsun.unibe.ch and the garbage you refer
to in PathLookup.h isn't there.  I only had to add one header file
(Don't remember offhand which one right now) and set the Makefiles
to the Sun versions, modify the etCC driver and everything but a few
sample applications compiled without any problems.


Curt Kohler                                             (513) 865-1106   
Mead Data Central                          DTS Fabrication Development
P.O. Box 933                                         curt@meaddata.com
Dayton, Ohio  45401                            ...!uunet!meaddata!curt

gamma@ifi.unizh.ch (erich gamma) (11/12/90)

Mr. Nyquist writes:

>> The current version of ET++ that has been distributed
>> seems not to work. Without any success at all I have tried to build
>> it.  
>> I would be grateful if Erich Gamma or Andre Weinand why it is so difficult
>> to build ET++ this time.

Flame on

May I suggest to Mr. Nyquist, that instead of wasting net bandwidth
by posting flames to contact us directly by mail... may be
we have solution to his problem!

Flame off

From people which are willing to communicate with us we
got some feedback concerning installation problems. 
 
Problem 1:
Linking an ET++ application under SUNOS4.1 results in:
 
	    ld: calrelco symb lookup botch
	    *** Error code 4
	    make: Fatal error: Command failed for target `micky'
 
This is due to a bug of ld in SUNOS4.1. The workaround is to change
the definition of ETO in ET_DIR/src/makefile from:
  ETO=  $(PE_OFILES) ofiles3.o Init.o $(WS_OFILES) $(PR_OFILES) \
	 ofiles2.o ofiles1.o
to:
  ETO=  $(PE_OFILES) $(OFILES3) Init.o $(WS_OFILES) $(PR_OFILES) \
	$(OFILES2) $(OFILES1)
 
Problem 2:
Applications crash when compiled with the original AT&T C++2.0 
compiler on SPARCstations. This problem has to do with the handling of variable
argument lists on the SPARC. 
Change the sed command in ET_DIR/bin/cpp from:
  sed -e "s/#line/#/g" -e "s/_au0___builtin_va_alist/__builtin_va_alist/g" $file
to:
  sed -e "s/#line/#/g" -e "s/__0__builtin_va_alist/__builtin_va_alist/g" $file
 
Mr. Nyquist continues:
>> I also found some rather strange files in the distribution...
					^
The ONLY strange file he found is indeed strange, but this file
does not cause any problems for the installation (there is of course
a funny story behind this file...). 

A new version of ET++(2.1) and patches to upgrade from 2.0 are available
by anonymous ftp from iam.unibe.ch [130.92.64.10] in "C++/ET++" or "pub"

--erich gamma