Re: Configure tips
From: John E. Malmberg (wb8tyw_at_qsl.network)
Date: 11/28/05
- Next message: AEF: "Re: HP : Massive strike and protest march today"
- Previous message: AEF: "Re: question about set file/attributes"
- In reply to: Martin Borgman: "Re: Configure tips"
- Next in thread: Dave Froble: "Re: MAKE utility for openVMS?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sun, 27 Nov 2005 22:44:08 -0500
Martin Borgman wrote:
> John,
>
> Thanks you for your comments.
>
> John E. Malmberg wrote:
>
>> Page 30, discussion on fork().
>>
>> If a program has a MS-DOS/Microsoft Windows variant, instead of fork,
>> it will use run-time library calls of spawnXX() with the "XX" being
>> replaced with characters that indicate how the arguments are passed.
>>
>> I have not tested this, but it appears that this might be able to be
>> implemented with either a static procedure or a macro using vfork()
>> and the corresponding execXX() functions, which means less
>> modifications to the original source.
> To be perfectly honest, most of the stuff I port is UNIX open source
> stuff. And I did not come across any spawnXX() calls. But I think your
> assessment of the spawnXX() calls is right.
> When I have some source code with a spawnXX() call, I can test your
> thesis. Do you have something?
I saw this in the Open Office Source code a few years back, specifically
in the dmake program.
> Are you using the simlink kit?
Only on some test systems at work.
> What version of GNV are you using?
At work, I am using 1.06-11 with my patches.
[Code to create a search list for GNV files edited out.]
> This is a very good idea, I would like to add this to the porting guide.
Feel free to.
> I have the very bad habit of correcting something that is wrong in the
> first place. I would include string.h in getopt.c and I would include
> stdlib.c in obstack.c.
Can you get the maintainers of M4 to put the fixes in the next release?
Otherwise you end up making the same edits over and over again.
Right now, M4 is not on a critical path for what I need to get done.
>> I did not link in a lib$initialize program on the build that I did, I
>> was depending on the DECC$ feature logicals.
>
> Yes, this works when you don't use the CRTL command line parsing functions.
As long as you set the DECC$ logicals before the main() program is
called the command line parsing options work.
Currently I am using a modified version of Perl that uses those logicals
to change to a mode that is more compatible with Bash. I am slowly
getting those changes inserted into the main Perl version.
> There already is something in the gcc wrapper, but when I tested this,
> there appeared to be something missing. This was quite some time ago
> though.
In one of the projects I am working on, I have had to enhance the CC
wrapper every few days in one way or another to continue on it.
> So you are finally updating the pipe code in bash. Or will there be a
> completely new version of bash?
Just an update to the version in GNV, it was needed as part of other
tasks that I am doing for work.
> Were you aware of this porting guide?
No.
> Are you interested in contributing to this guide?
If I have time.
I am noticing that there seem to be several open source projects like
GTK and such that are needed for multiple projects, and a number of
smaller ones.
It takes a lot of time to participate in the mailing lists or other
forums for those projects once an OpenVMS port is done, to follow up and
keep the project buildable on VMS.
-John
wb8tyw@qsl.network
malmberg@dskwld.zko.hp.compaq.dec
Personal Opinion Only
- Next message: AEF: "Re: HP : Massive strike and protest march today"
- Previous message: AEF: "Re: question about set file/attributes"
- In reply to: Martin Borgman: "Re: Configure tips"
- Next in thread: Dave Froble: "Re: MAKE utility for openVMS?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
