Re: Constant array

On Jan 23, 10:40 am, Aleksej Saushev <a...@xxxxxxxx> wrote:
Helmar <hel...@xxxxxxxxx> writes:
On Jan 23, 10:08 am, Aleksej Saushev <a...@xxxxxxxx> wrote:
Helmar <hel...@xxxxxxxxx> writes:
On Jan 23, 8:59 am, Aleksej Saushev <a...@xxxxxxxx> wrote:

But the usual way is to write that as separate program, generate the data
once, and include it into your program.

This "usual way" is usual only for broken underspecified languages, such as C.
Sane languages should provide this.

That has nothing to do with "sane". I can do a "make -J" for separated
files and generators. You do not ever get a language that "sane" that
it compiles it itself from your wordy expression (at least that would
be magic without any binary "interpreter" used).

Sorry? Which way multiple make jobs help with precalculation of data?

IPC (inter process communication), opening a pipe, using temporary
files - that all is solved since years.

"Show me the code."

You dont know how? Hey, what ever drugs you take, they are not good.
What is the problem? First of the half you OS could offer to use
common data spaces. The other one is that using a temporary file
should not be complicated and even a named pipe is a concept since
years. You could also open sockets to communicate. Nothing is new, all
is there for years.

It never was done the way you tell it, and it doesn't seem even possible.

The way you're able to get it done, requires support for separate compilation
to object code, which is a patchwork around the lack of proper module support:
instead of modules and namespaces you have rules about system-dependent linker
("you are to compile to object code and export symbols for these but not these
functions"). Do you have it in Forth? Do you want it?

Again, you propose dependency on external tools when there's absolutely no need
to do so.

Plus, you don't specify those tools. Do you assume POSIX-compatible environment?
How much of POSIX/SUS do you require? I have no access to SUS3-compliant systems,
and so do many other developers, the world isn't all MacOS X and Solaris.





Not to mention that multiple make jobs are wrong reference, since
both MPICH (MPICH2) and OpenMPI don't support building with multiple
make jobs.

If "-J" meant multiple make jobs:

-j - this was a typo (" needs a shift key here).

NetBSD make(1):

"-J private
      This option should not be specified by the user."

FreeBSD make doesn't support "J" flag, so does DragonFly.

In plain words, you CAN'T do "make -J".

OK, that was a typo in front of " that needs a shift key at my
keyboard. The right option is -j without parameter. What BSD does I do
not know, but my "make"s do multiple jobs then. I'd bet it would work
with smake, but I do not have installed it on the current system I
use ;)

Even after typo corrected it still doesn't work in too many cases.

What?! You've to be sure not to use improper temporary file names.
That is all. Then it should work - the error messages need more
investigation, but if everything works it goes faster.

So far,
-Helmar

Again, even MPICH2 and OpenMPI don't support parallel make jobs,
so do many large programs, where it is more critical than in small ones.
Not to mention that your "make" isn't the only out there and you
definitly can't expect "smake" installed if even you don't have it
on your system sometimes.

--
HE CE3OH...

.

Relevant Pages

  • Re: Constant array
    ... Sane languages should provide this. ... Which way multiple make jobs help with precalculation of data? ... The way you're able to get it done, requires support for separate compilation ... to object code, which is a patchwork around the lack of proper module support: ...
    (comp.lang.forth)
  • RE: Lotus Domino crack http hash tools
    ... Lepton's Crack also compiles on Cygwin on Windows, ... Both versions support the old R4 Domino hash, ... There are several different crackers for Lotus Domino, ...
    (Pen-Test)
  • RE: ascx is ambiguous in the namespace ASP
    ... If I had duplicate files in the ... directory it would occur every time the site compiles. ... Microsoft MSDN Online Support Lead ... where an initial response from the community or a Microsoft Support ...
    (microsoft.public.dotnet.framework.aspnet)
  • RE: Is there a way to generate a CodeCompileUnit from source code?
    ... It compiles. ... Microsoft Online Community Support ... You can send feedback directly to my manager at: ...
    (microsoft.public.dotnet.framework.clr)
  • Re: compile error with eVC SP4 auto_xxx.hxx and allocator.hxx
    ... invalid expression as template parameter ... the ones before version 13 didn't support ... compiles with 13.0 or 13.10, but I'm not sure those are installed with the ... SDKs ...
    (microsoft.public.windowsce.embedded.vc)