headers
Mark A. Flacy | 1 Feb 2003 04:50

[arch-users] 1.0pre17 news and more tutorial chapters available

Tom, please stop reading this until after you finish the up-coming release,
OK?  

>>>>> "Scott" == Scott Parish <sRp <at> aspen.srparish.net> writes:
Scott> 
Scott> On Fri, Jan 31, 2003 at 01:15:13AM -0800, Tom Lord wrote:
>> XML is a conspiracy.  It should be purged from the earth.  It's a
>> scam.  It's a sham.  It's ugly.  It's needlessly complicated.  It has
>> abstractions in all the wrong places.  It's a family of standards sans
>> experience and implementation.  It's sole success is as a cheesy way
>> to give java, C++, python and perl programmers a crude approximation
>> of s-exprs.  It's a stupid design.  It's a 20-ton collection of
>> "standards" where 20 pounds is called for.  It's a *parody*, friends.
>> They thought you'd *get it* by now.  XML is a *joke*, folks -- one
>> that got a little out of hand.  It's straight-faced *satire* of
>> everything that's wrong when capitalists dominate engineering.  It's a
>> con job.  It's proof of an industry driven by buzzwords and naive
>> idiots who think programming is trivial.  It's the emperor's new
>> clothes. 
Scott> 
Scott> Note to self: Self, be _very_ careful what kind of jokes you play with
Scott> earth in the state its in; they could be taken seriously and contaminate
Scott> a whole industry.

Note to self:  Self, Tom has some very strange ideas sometimes.  I'm sure
that we would all be absolutely fascinated by the use of s-exprs as a text
markup language, which would be undoubtedly easier for humans to read.
DocBook XML came originally from DocBook SGML, which in turn traces its
lineage from earlier versions of text markup languages.
(Continue reading)

Permalink | Reply | 8 comments |
headers
Walter Landry | 1 Feb 2003 07:37

[arch-users] N-97% mkpatch speedup

Karel Gardas <kgardas <at> objectsecurity.com> wrote:
> 
> Hello,
> 
> here is my final experimental patch for speeding up mkpatch a bit. It's
> based on two main changes:
> 
> 1) filtering out not changed files and directories, so both cpu consuming
> main loops in mkpatch shell code will work with less data
> 
> 2) on demand directory creation
> 
> The result of second point is that we don't need to use famous
> multi-dirname utility.
> 
> I've done small benchmarks on our MICO source tree. The first test is with
> using the same trees and the second uses trees with the longest path i.e.
> generates the biggest patch.

I've finally gotten around to figuring out how to add rudimentary C++
support to the build system.  So I've added this patch in, and it
makes quite a difference.  Not as dramatic for ordinary commands as
what Karel posted, but impressive nonetheless.  For example, computing
what-changed for the src/arx sub project (formerly src/arch) for the
new version

  boo <at> grue:~/arx/arx/src/arx$ time arx what-changed 
  *** patched regular files

      ./patch-sets/mkpatch.sh
(Continue reading)

Permalink | Reply | 3 comments |
headers
Mark A. Flacy | 1 Feb 2003 21:10

[arch-users] tutorial: "larch tag"

>>>>> "Robert" == Robert Anderson <rwa <at> alumni.princeton.edu> writes:
Robert> 
Robert> After reading the section "Elementary Branches -- Maintaining
Robert> Private Changes", I am wondering why the command is called "larch
Robert> tag" and not "larch branch."
Robert> 
Robert> From reading this section, it seems quite clear to a casual reader
Robert> that the function of "tag" is to branch.  Comments?

Name overload.

We already use the term "branch" as a sub-division of "category" in our
namespace.   We also already have a command "larch make-branch" which does
something not directly related to tagging lines of development.

It may be better to find a new name for the sub-division of "category",
unless the semi-proposed "larch branch" command restricted itself to
tagging within the same category.  I don't see a reason to require that
restriction. 

We could change one of the two arch uses of the word "branch" to the word
"fork".  (I'd choose to change the one associated with "category", myself.) 
There's undoubtedly a better word we can use, but that one popped into my
head at the moment.

--

-- 
Mark A. Flacy       (972) 685-8347
moc.ibtta <at> 1ycalfm  <<- To defeat not very bright spambots.
"There is no virtue without temptation."
(Continue reading)

Permalink | Reply | 0 comments |
headers
Tom Lord | 2 Feb 2003 08:02

[arch-users] 1.0pre17 news and more tutorial chapters available


       Tom, please stop reading this until after you finish the
       up-coming release, OK?

:-) Thanks for the focus shove.  You'll be happy to know I'm writing
this reply waiting for a compiles and the like.  The tutorial is
nearly finished (well, enough for 1.0pre17) -- I'm on track for a lazy
sunday of release engineering and a busy early-next-week handling any
bug reports for stupid things I missed after N+1 months of absense
from making releases.  

Mostly, I ranted about XML as a sort of esoteric joke -- people who
know me know my more considered opinions about XML; readers of this
list know me for expressing myself with somewhat extreme statements
from time to time.... it seemed a harmless enough reinforcement of
certain caricatured images of me. 

        Note to self:  Self, Tom has some very strange ideas
       sometimes.  

When I surround text with `<rant>...</rant>' that's a hint that I'm
exaggerating and, when the rant is about XML, being a bit ironic.
Nevertheless, unless you invoke Occam's razor (simple stupidity
explains a lot), I think you'd have a hard time proving, based on
external evidence, that XML isn't a joke of the sort I described.

In fact I _don't_ much like XML, but as a tool for marking up textual
documents -- well, that's at least a _far_ more reasonable use than
many that it's put to.  At least it stays close to it's roots in such
applications.
(Continue reading)

Permalink | Reply | 7 comments |
headers
Stig Brautaset | 2 Feb 2003 14:00

[arch-users] Labnotes doc bug?

I'm reading a bit about `package-framework', trying to find out if it is
something I would want to apply to my own future programs. It is
tempting. I'm wondering if I've found a little bug in the documentation,
and have included a patch:

--- Labnotes/PackageFramework.d/BuildingLibraries.orig  Sun Feb  2 12:52:24 2003
+++ Labnotes/PackageFramework.d/BuildingLibraries       Sun Feb  2 12:52:45 2003
 <at>  <at>  -40,7 +40,7  <at>  <at> 

 * A Library Spread Over Multiple Directories

-  A very large directory might be spread over multiple
+  A very large library might be spread over multiple
   directories.  For example, suppose we have the directory
   `hackerlab', in which `libhackerlab.a' will be built.
   But we also have `arrays' and `chars' which contain `.c' files

Stig
--

-- 
brautaset.org
Permalink | Reply | 4 comments |
headers
Stig Brautaset | 2 Feb 2003 14:14

[arch-users] Labnotes doc bug?

On Feb 02 2003, Stig wrote:
> I'm reading a bit about `package-framework', trying to find out if it is
> something I would want to apply to my own future programs. It is
> tempting. I'm wondering if I've found a little bug in the documentation,
> and have included a patch:
...

And another one:

--- Labnotes/PackageFramework.d/ConfigVariables.orig    Sun Feb  2 13:10:27 2003
+++ Labnotes/PackageFramework.d/ConfigVariables Sun Feb  2 13:10:45 2003
 <at>  <at>  -144,7 +144,7  <at>  <at> 
   for example, the option `posix-shell' becomes
   `$(cfg__posix_shell)').

-  Command line variables are also exproted as *CPP* macros in
+  Command line variables are also exported as *CPP* macros in
   `$(buildroot)/config-include/config-options.h'.  The directory
   `$(buildroot)' is added to the default search path for include
   files for *C* programs.

Stig
--

-- 
brautaset.org
Permalink | Reply | 0 comments |
headers
Robert Anderson | 2 Feb 2003 21:14

[arch-users] 1.0pre17 news and more tutorial chapters available

On Sat, 2003-02-01 at 23:02, Tom Lord wrote:

> :-) Thanks for the focus shove.  You'll be happy to know I'm writing
> this reply waiting for a compiles and the like.  The tutorial is
> nearly finished (well, enough for 1.0pre17) -- I'm on track for a lazy
> sunday of release engineering and a busy early-next-week handling any
> bug reports for stupid things I missed after N+1 months of absense
> from making releases.

Where can we "advertise" this new release?

Can some of the newer or more infrequent posters tell us how they got on
the arch trail?

Was it the slashdot article, the linux journal article, a different
mailing list?

Thanks,
Bob
Permalink | Reply | 6 comments |
headers
Walter Landry | 2 Feb 2003 21:19

[arch-users] Labnotes doc bug?

Stig Brautaset <arch <at> brautaset.org> wrote:
> I'm reading a bit about `package-framework', trying to find out if it is
> something I would want to apply to my own future programs. It is
> tempting. 

I've been looking at it.  The main problem I have with it is that it
is really only knowledgeable about C.  Scripts work fine because they
don't need to be built, just installed.  In general, it tries to be a
little to clever for my taste.  It also requires PLUGIN directories in
addition to makefiles.  I'm not sure how well it works with parallel
builds.  Finally, it doesn't actually solve the speed problem when
invoking recursive makes.

But for C and script projects that aren't too large (<100000 lines),
it is probably great.  But realize that there are many other options
(SCons, jam, Boost's jam, Cook, Bras, Ant, PyAnt, Odin, Sire).  I
couldn't tell you what the best one is right now.

> I'm wondering if I've found a little bug in the documentation,
> and have included a patch:

Thanks.  I've fixed them both (BuildingLibraries and ConfigVariables).

Walter
Permalink | Reply | 2 comments |
headers
Stig Brautaset | 2 Feb 2003 21:22

[arch-users] 1.0pre17 news and more tutorial chapters available

On Feb 02 2003, Robert wrote:
> On Sat, 2003-02-01 at 23:02, Tom Lord wrote:
> 
> > :-) Thanks for the focus shove.  You'll be happy to know I'm writing
> > this reply waiting for a compiles and the like.  The tutorial is
> > nearly finished (well, enough for 1.0pre17) -- I'm on track for a lazy
> > sunday of release engineering and a busy early-next-week handling any
> > bug reports for stupid things I missed after N+1 months of absense
> > from making releases.
> 
> Where can we "advertise" this new release?
> 
> Can some of the newer or more infrequent posters tell us how they got on
> the arch trail?
> 
> Was it the slashdot article, the linux journal article, a different
> mailing list?

I *believe* it was an article (or at least a mention) in a magazine that
put me on the trail of arch. 

Stig
--

-- 
brautaset.org
Permalink | Reply | 0 comments |
headers
Walter Landry | 2 Feb 2003 21:28

[arch-users] 1.0pre17 news and more tutorial chapters available

Robert Anderson <rwa <at> alumni.princeton.edu> wrote:
> On Sat, 2003-02-01 at 23:02, Tom Lord wrote:
> 
> > :-) Thanks for the focus shove.  You'll be happy to know I'm writing
> > this reply waiting for a compiles and the like.  The tutorial is
> > nearly finished (well, enough for 1.0pre17) -- I'm on track for a lazy
> > sunday of release engineering and a busy early-next-week handling any
> > bug reports for stupid things I missed after N+1 months of absense
> > from making releases.
> 
> Where can we "advertise" this new release?

FYI, I've changed the freshmeat listing over to ArX, so Arch should be
free to anyone who wants it.  For this kind of release, that is
probably appropriate.

Beyond that, I think it is still premature to start shouting to the
heavens.  You're just going to end up with a lot of disappointed
people.  It is not like the speed has gotten reasonable.

Walter
Permalink | Reply | 1 comment |

Gmane