Matthew,

I spent a few hours trying to get your makefiles to work with the your STL header software.  No success.

Would you be kind enough to use your own recls to create and publish a
single listing of your installations of stlsoft-1.91-beta10.zip and
recls-1.8.1.zip
of a UNIX/Linux installation?  If it has dates, times, and complete paths, I
think
users can figure out what they should be doing.

Thank you.
Aubrey Soper

Aubrey

[I'm cross posting this to c++.stlsoft. I think we should continue it on there.]

"DIGITAL MARS" <asoper3@cox.net> wrote in message news:e7i9v9$1r2j$1@digitaldaemon.com...
> Matthew,
>
> I spent a few hours trying to get your makefiles to work with the your STL header software.  No success.

Which makefiles?

Can you be specific as to what problems you're having?

> Would you be kind enough to use your own recls to create and publish a
> single listing of your installations of stlsoft-1.91-beta10.zip and
> recls-1.8.1.zip
> of a UNIX/Linux installation?  If it has dates, times, and complete paths,
> I think
> users can figure out what they should be doing.

That would yield just whatever's in those files.

I think what you're after is to see whether there are any "green" issues, i.e. whether I've files on my local systems that are required to build recls 1.8.1 (for DMC++) that I've failed to include in one or the other distribution.

What I can do is attempt to build in a shell that is cleared of all environment variables. I'll have a go at that, but it would still be useful to me to know your precise problem.

> Thank you.
> Aubrey Soper

FYI: I'm away from home at the mo - in NYC, as it happens - so may not be able to do this until I get back home to sunny Australia towards the end of the week. ;-)


Cheers

Matthew

Ok, well I've found a problem alright.

Apologies for this. I was doing things in a rush, in preparation for this trip.

I've now done STLSoft 1.9.1. beta 11, and corrected the recls page.

It now all works again, save that for some reason (that I will discover upon my return) the generated makefile for DMC++ does not cause the delay-load functionality for wininet.lib (which is in recls_wininet_dl.cpp) to be included, which means that the DMC++ build is missing, and asks for, wininet.lib. If you're building with DMC++, you will probably want to add that to the libraries list in the makefile. I've included a hand-modified version of the DMC++ makefile that (I believe) is now correct.

Again, sorry to you and anyone else for the hassles. It was a bit of a punt to push out recls 1.8.1 in the time I had, and I guess I gambled and lost.

1.8.2 will be on its way - with fully fleshed out docs and test programs and correct makefiles - sometime in the first week of July.

Cheers

Matthew



"Matthew" <nowhere@noaddress.co.us> wrote in message news:e7llo4$fqp$1@digitaldaemon.com...
> Aubrey
>
> [I'm cross posting this to c++.stlsoft. I think we should continue it on there.]
>
> "DIGITAL MARS" <asoper3@cox.net> wrote in message news:e7i9v9$1r2j$1@digitaldaemon.com...
>> Matthew,
>>
>> I spent a few hours trying to get your makefiles to work with the your
>> STL
>> header software.  No success.
>
> Which makefiles?
>
> Can you be specific as to what problems you're having?
>
>> Would you be kind enough to use your own recls to create and publish a
>> single listing of your installations of stlsoft-1.91-beta10.zip and
>> recls-1.8.1.zip
>> of a UNIX/Linux installation?  If it has dates, times, and complete
>> paths,
>> I think
>> users can figure out what they should be doing.
>
> That would yield just whatever's in those files.
>
> I think what you're after is to see whether there are any "green" issues,
> i.e. whether I've files on my local systems that are required to build
> recls
> 1.8.1 (for DMC++) that I've failed to include in one or the other
> distribution.
>
> What I can do is attempt to build in a shell that is cleared of all
> environment variables. I'll have a go at that, but it would still be
> useful
> to me to know your precise problem.
>
>> Thank you.
>> Aubrey Soper
>
> FYI: I'm away from home at the mo - in NYC, as it happens - so may not be
> able to do this until I get back home to sunny Australia towards the end
> of
> the week. ;-)
>
>
> Cheers
>
> Matthew
>
> 

Thank you for your courteous, prompt reply.

I would like to suggest that you build a tarball for Linux use that includes all the STLSoft headers, the recls program example, and a makefile.  I would further suggest that this example program use an #ifdef statement to check for the version of the gnu compiler that is not older than what you specify.

The tarball wouldn't need to be accompanied by any instructions, except on other UNIX systems that have defective include files or include file names, which is to say that I think it would conserve your valuable time.

Thank you.
Aubrey Soper