-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi all,
thanks to all of those who provided some information on an ooRexx
grammar. Here, I want to share a central finding and my conclusion from it.
Parsing (oo)Rexx is hard. The reason for this is that a symbol can be a
symbol or a keyword,
Parsing (oo)Rexx is hard. The reason for this is that a
symbol can be a symbol or a keyword, depending on the
context. Have a look at the following example: [snip]
Yes, that's correct (NetRexx makes it a bit more regular, in that once a
variable name always a variable name as it were, but
[Aside: I would argue that the reliance of language designers on simple
LALR parsers, etc., has significantly affected the development of
languages. It has made them easier for computers to process but harder
for people to use. I think that's backwards :-).]
Mike
I agree, but even
[Aside: I would argue that the reliance of language designers on
simple LALR parsers, etc., has significantly affected the
development of
languages. It has made them easier for computers to
process but harder
for people to use. I think that's backwards :-).]
I agree, but even
We don't seem to be getting much activity with the 4.0.1 rc1. In my
opinion, we should just go ahead and do a release for 4.0.1.
However, my opinion may be biased a little bit because I'm real busy
right now and I'll be out of town for 2+ weeks starting May 7. I'd
like to start the builds in a
Go for it! The crickets are starting to keep my up at night :-)
Rick
On Mon, Apr 26, 2010 at 11:34 AM, Mark Miesfeld miesf...@gmail.com wrote:
We don't seem to be getting much activity with the 4.0.1 rc1. In my
opinion, we should just go ahead and do a release for 4.0.1.
However, my
Um, didn't think there has been an announce of RC1 yet? (Last I
got/downloaded was 31 March, beta1?)
Mike
-Original Message-
From: Rick McGuire [mailto:object.r...@gmail.com]
Sent: 26 April 2010 16:36
To: Open Object Rexx Developer Mailing List
Subject: Re: [Oorexx-devel] 4.0.1
The announce went out April 17th and was posted to the RexxLA list,
the ooRexx annoucement list, the ooRexx dev list and comp.lang.rexx.
Not sure why you didn't see it.
Rick
On Mon, Apr 26, 2010 at 11:59 AM, Mike Cowlishaw m...@speleotrove.com wrote:
Um, didn't think there has been an announce
I was looking at the llvm compiler, and was wondering if anyone had
compiled oorexx with it.
Thanks,
Bruce
--
___
Oorexx-devel mailing list
I'm not even sure what that compiler *is*, but I'm pretty sure I
haven't tried it.
Rick
On Mon, Apr 26, 2010 at 12:04 PM, CVBruce cvbr...@gmail.com wrote:
I was looking at the llvm compiler, and was wondering if anyone had
compiled oorexx with it.
Thanks,
Bruce
About 10 years ago, I did a lot of performance work on Object Rexx.
Key to that performance works was having access to the performance
profiler that was part of Visual Studio then. It appears that the
performance profiler is now part of a separate and fairly expensive
package that I don't have
llvm is Low Level Virtual Machine compiler. Supposed to be GCC
compliant.
This is some of the underlaying technology behind Grand Central
Dispatch, and OpenCL. Grand Central Dispatch(GCD) is software that
helps move code to multiple processors, while OpenCL is technology
that allows GCD
Rick,
How much are we talking about. Perhaps we could solicit some
donations to cover it.
Bruce
On Apr 26, 2010, at 9:14 AM, Rick McGuire wrote:
About 10 years ago, I did a lot of performance work on Object Rexx.
Key to that performance works was having access to the performance
profiler
I'm not real sure about the Visual Studio tools, but I suspect this
around $1500 or more. I still haven't gotten a very clear picture as
to which product offering that's included in. The vTune package is
around $800. AMD has a similar offering, but full use requires an AMD
processornone of
Well if you don't find a linux tool, I'd be happy to chip in. I know
in SF that there is a donation button. Do you know if that will get
to the right place? Another possibility is to set up a fund under
RexxLA, so that one could recoup the 501C deduction.
Bruce
On Apr 26, 2010, at 9:25
I'm not even sure the SF donation button is even active. I think Mark
Hessling set that up originally...not sure it's gotten a lot of
presses to date :-)
Rick
On Mon, Apr 26, 2010 at 12:39 PM, CVBruce cvbr...@gmail.com wrote:
Well if you don't find a linux tool, I'd be happy to chip in. I
One year ago, I started to investigate Valgrind (under Linux) because I had
poor performances with csvStream, compared with stream + parse.
The datas, scripts and results are available here :
http://jean-louis.faucher.perso.neuf.fr/ooRexx/profiling/csv
If I remember correctly, the numbers you see
This data looks like the sort of information I'm looking for. How
difficult was it to get this set up and running? The shell scripts
for invoking this look pretty simple. Did you need to compile the
code with any special options to gather the data?
Rick
On Mon, Apr 26, 2010 at 1:09 PM,
Nothing to recompile, you just run Valgrind by passing the path to the
executable. But the execution is very slow...
Jean-Louis
--
___
Oorexx-devel mailing list
*Slow* data is still better than *no* data. And not having to do
anything to prepare the executables means quicker iterations between
measurements, which is good.
Rick
On Mon, Apr 26, 2010 at 1:21 PM, Jean-Louis Faucher
jfaucher...@gmail.com wrote:
Nothing to recompile, you just run Valgrind
Strange indeed .. I usually automatically follow the link and download. (I
did for the beta1, but didn't install because traveling the next day.) I
don't recall seeing an RC1 anouncement at all.
Mike
-Original Message-
From: Rick McGuire [mailto:object.r...@gmail.com]
Sent: 26
If using GCC then gprof is another option .. I used it for a while and I
think it was helpful.
-Original Message-
From: Rick McGuire [mailto:object.r...@gmail.com]
Sent: 26 April 2010 18:24
To: Open Object Rexx Developer Mailing List
Subject: Re: [Oorexx-devel] Performance
22 matches
Mail list logo
