tends to annoy your users, and will lead to bug reports.
The annoyance is particularly great within organisations running full build
tests overnight, because sooner rather than later the tests will fail.
Nicholas Clark
On Thu, Apr 20, 2006 at 02:49:51PM +0200, Patrik Wallstrom wrote:
> Does anybody here know how to reach [EMAIL PROTECTED]
How have you tried contacting him, and how has it failed?
Nicholas Clark
rn, and marking
that package as only depending on the packages (modules) that it stated
as run-time requirements, basing that dependency information on the
metadata from the modules (instead of what got installed)
Nicholas Clark
t; just not right that they don't work.
What would you consider to be the "right" that should be happening here?
Answering that will make answering your next question easier:
> I'm not a skilled C/XS programmer, or I would consider taking over
> them. Can anybody have advice on this issue?
Nicholas Clark
On Thu, Aug 24, 2006 at 03:16:22PM +0800, imacat wrote:
> On Wed, 23 Aug 2006 14:28:13 +0100
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > On Wed, Aug 23, 2006 at 06:52:26PM +0800, imacat wrote:
> > > But this ain't right. Crypt-Cracklib is critical to se
ered the other choices.
Sadly few people seem to.
But please don't be surprised or disheartened if you receive no response.
Nicholas Clark
me. No, I don't have a great idea on the other part,
aside from sticking the data into a .pm file you generate at build time,
after __DATA__, and reading it in from there.
(don't forget to close the DATA file handle to avoid wasting resources)
Nicholas Clark
most recent release is some time ago.
Nicholas Clark
OK
>
> Greg's version gets optimized as well:
>
> perl -MO=Deparse -e'sub PI () {3.1415} print PI'
> sub PI () { 3.1415 }
> print 3.1415;
> -e syntax OK
As of 5.10, use constant will use less memory than Greg's version.
I'm loathe to make the same changes for 5.8.9, as the implementation
confuses some modules that inspect the symbol table. The necessary
infrastructure will be in 5.8.9 though.
Nicholas Clark
have a
> change be reflected in the data stored in the optree but I suspect it
> is unlikely.
The optree is read only. So the caller implementation has to respect this.
However, for efficiency it is constructing a scalar which points to the
bytes in the optree. So if anything ignores the readonly flag on the SV it
will be changing the bytes in the optree.
How "Safe" this is, I'm not sure.
Nicholas Clark
gt; Thoughts?
I think you're missing what Slaven has been up to:
http://search.cpan.org/~srezic/Tk-804.027_501/
Nicholas Clark
ependency
on something else you weren't previously using, so you can reach the
situation where upgrading to fix a bug will also bring in something new that
you didn't want (for valid local policy reasons).
Nicholas Clark
en we have to get all the automated test smokers to upgrade.
You mean there isn't a backdoor in the smokers code to allow it to upgrade
itself when it detects that it is smoking a newer version of itself? :-)
Nicholas Clark
future
> ones) as a new distribution.
Some people go on holiday for more than 2 weeks.
Nicholas Clark
ject file extension.
You can get that from $Config:
$ perl -V:obj_ext
obj_ext='.o';
Nicholas Clark
= 0x800f18
SV = NULL(0x0) at 0x800f18
REFCNT = 2
FLAGS = ()
It would be dangerous to rely on this reference counting behaviour remaining
the same.
Nicholas Clark
commit to blead, so someone else can deal with
that part.
Nicholas Clark
o please replace the file hints/freebsd.sh in the core
distribution with the attached file, and then try building.
Nicholas Clark
freebsd.sh
Description: Bourne shell script
On Fri, Oct 31, 2008 at 08:03:10PM +, David Cantrell wrote:
> What if my version of GPL2.txt has an extra CRLF at the end because of
> how I cut n pasted it from the GNU website?
Or has the address changed, as they are wont to do.
Nicholas Clark
r idea:
$ /home/nclark/Sandpit/588ish/bin/perl -Mthreads -e0
This Perl not built to support threads
Compilation failed in require.
BEGIN failed--compilation aborted.
$
Nicholas Clark
t top
level to live under. It is, arguably, an encoding module, but it isn't the
interface of Encode, which is what modules under the Encode top level provide.
So I'm not sure what to call it.
Nicholas Clark
* http://en.wikipedia.org/wiki/Ascii85
On Mon, Nov 24, 2008 at 11:49:04AM -0500, Darian Anthony Patrick wrote:
> Darian Anthony Patrick wrote:
> > Nicholas Clark wrote:
> >> I've written a module that implements the base 85 encoding used by the old
> >> btoa program, and by PDFs as their Ascii85 encoding
ter within my own CAM::PDF module.
I infer that you only use it for input, because as best I can tell, the output
filter does not work: http://rt.cpan.org/Public/Bug/Display.html?id=41085
Nicholas Clark
rt-cpan-ad...@bestpractical.com
Nicholas Clark
mall, and the
Hash lookup should be O(1), independent of number of keys. Of course, a hash
with more keys uses more memory, but so does an array with more elements.
Nicholas Clark
On Mon, Mar 02, 2009 at 11:15:53PM -0800, Joshua ben Jore wrote:
> On Mon, Mar 2, 2009 at 12:22 PM, Nicholas Clark wrote:
> > On Mon, Mar 02, 2009 at 10:23:38AM -0800, Bill Ward wrote:
> >
> >> Personally I always use hashes for objects. Hashes are pretty fast in
> &g
res.
(Note, our internal code is not XS, and is laid out in subversion correctly
for the live deployment. The odd bits of XS code are treated as external
CPAN modules)
Nicholas Clark
* Or at least, this will become true soon.
pace transferred to someone else, so that we can
: create a new package and have that one installed via the CPAN shell?
(it's one level of quoting above the innermost message above)
Given that the original correspondent did not say "I've tried contacting the
author", and stated that he wanted to know the general procedure, I strongly
assume not.
Nearly everyone who has contacted the author before mailing here has said so.
Nicholas Clark
olite but sarky note about it, finishing by
suggesting that he used http://jobs.perl.org/
Nicholas Clark
this, setting argv[0] to (also) be the script's name, with nothing passed
in to give the actual name of the interpreter.
Nicholas Clark
On Thu, Feb 18, 2010 at 12:00:23PM +, David Cantrell wrote:
> On Thu, Feb 18, 2010 at 11:24:24AM +0000, Nicholas Clark wrote:
> > On Wed, Feb 17, 2010 at 08:44:57AM -0800, macke...@animalhead.com wrote:
> > > $^X is simpler, and simpler solutions are preferable.
> > &g
e of storage
management is better than mine. But that's not the key question, in a
volunteer organisation. The questions I ask, repeating Jan's comments in
another message, are.
Nicholas Clark
ee
of the server, which, I thought, was functionally equivalent to the behaviour
of the rsync server.
Nicholas Clark
Sort of like Schwern already did?
http://github.com/gitpan
Nicholas Clark
d this subtree off to that other server"
This would allow the (fast, existing, C) rsync server to serve most of
(say) funet.fi, handing off to a stateful server for the CPAN subtree.
Nicholas Clark
en that client is going fail
sooner or later.
Nicholas Clark
ly, and seeing if that gives you ideas on what
sort of name it should be. You *can* easily change the name of something
before it is published, and others start using it.
[and I note that in reply I've made at least 3 style errors, starting
sentences with conjunctions. I guess I'm structuring my e-mail more like
speech than formal writing.]
Nicholas Clark
Because the nice thing about your suggestion is that it doesn't involve
changing any of the server infrastructure, and it's an incremental change
which can be done by each mirror in turn.
Instead of running rsync over the whole tree, it can change to run a top
level script that runs rsync over the parts that have to be copied, and then
run the symlink generation on the parts that can be recreated locally.
Nicholas Clark
lers regularly crashes perl.
> >
> > 3) He also mentioned that he thinks the OO system of Perl is a hack -
> > that the objects are hash refs and there is no privacy.
>
> Out of curiosity, do you know what version of Perl they were running?
Because #2 sounds like it's pre 5.8.1
Nicholas Clark
ues. Which effectively means that you need to take out
a write lock on every read. Which is "spendy".
2: C extensions access Perl's variables by direct structure manipulation,
rather than calling C functions. Which means that you can't (really)
intercept (logical) accesses to variables to indirect them via thread-safe(r)
code.
Nicholas Clark
es from me (but more than 1000) has just written a
disaster, in Java. And I'm curious in a couple of years how the majority of
recently written Rails apps turn out. (Particularly Rails, because it's rapidly
become very trendy, which means that demand for programmers will have
outstripped experience with it)
Nicholas Clark
On Tue, Mar 29, 2011 at 01:27:36PM +1100, Michael G Schwern wrote:
> After 6.58 MakeMaker will seriously consider shooting 5.6 in the head... I
> mean ending long term support. That will make shipping Scalar::Util
> unnecessary.
Excellent.
(release the hounds)
Nicholas Clark
;s happy to accept competent patches from anyone who makes his code
woke on Windows?
Because you're going the right way about it to get either of those two
results.
I appreciate what you want. I'd also want what you ultimately want.
But you're going the *wrong way* about getting it.
Nicholas Clark
.8.5 in production.
Happy Birthday 5.8.5 - 7 years old today!
Nicholas Clark
does seem that the *only* firm that is* spamming CPAN authors is
Betonmarkets, as they've done it *again*.
Life would be quieter if they kindly ceased trading.
(Because that seems a more reliable way for them to stop, than issuing
an apology then repeating the offence every 18 months)
Nicholas Clark
* or fails to stop agents acting on their behalf.
.
Also, Claes attempted to reply to it and met a challenge-response system.
Not sure what else they can do to make themselves look more stupid.
Hence why I'd like them to kindly cease trading.
Nicholas Clark
* Actually, it seems that Uri was also working to help them recruit, and so
was go
also careful enough to know how to add it back. Whereas for
the people *for whom it matters*, they're just going to copy the code as-is,
see that it works (for their simple test case), and move on.
Free as in landmines.
Nicholas Clark
y release using the
Perl source code providing one doesn't pass it off as Perl.
You seem to be basing a chunk of your reasoning on a lot of false
assumptions.
Nicholas Clark
's compiler is available for Linux (just for x86 and x86_64, I think)
I've used lcc on Linux
I've not tried clang on Linux
That's 4 without trying, all of which I believe can be used in some cases
without payment.
Nicholas Clark
Non-Commercial Software Development", which still seems to
be available:
http://software.intel.com/en-us/articles/non-commercial-software-development/
It's limited to "My use of software products is for personal non-commercial
purposes." which I think actually excludes me, based
ously seemed interested in having this happen, but inexplicably
vanished when we said "go for it!"
Nicholas Clark
rt.cpan.org to see if someone has already
submitted a patch that fixes it.
In quite a few cases, there is a patch just waiting for the module's owner
to apply it at make a release.
Nicholas Clark
alent of a meat cleaver, to
threaten the vegetables, er software, into compliance. So this might be a
problem.
Nicholas Clark
* Other than git reset --hard
On Fri, Dec 13, 2013 at 10:11:09AM +, Nicholas Clark wrote:
> balls it up). And as long as you don't throw away the terminal output,
> push anything or run `git gc`, there's always about 30 days to recover from
> any mess* from the various internal reflogs.
> * Other
o write tests that actually test what they think that they are
testing, and testing the wrong number isn't useful.
IIRC one recent patch sent by a third party ended up being totally wrong in
its assumptions, and fixing the problem properly took somewhat over 5 hours
of serious concentra
#x27;s perl? :-(
[and anyone who assumes that perl is the first perl in my $PATH
(eg http://jarl.sourceforge.net/) does not please me]
I think that's going to be "interesting" for the rare case where the script
comes in on stdin.
Nicholas Clark
re flexible.
However, I've not really thought much about counter arguments, apart from
convenience of getting 1 tarball (versus inconvenience of upgrading 1 tarball
every time any part of one of the three changes) and the possible
interdependency of all 3 modules.
Nicholas Clark
ea
quite who *is* on this list, actually) may not be familiar with all of them.
Could you summarise what the requirements you had were, and why the modules
you investigated didn't match up (even if it's just 1 line for each module)?
(as to your other questions, I can't offer any
be able to cram another 5 to 10% on the CD by turning all the tar.gz
files to tar.bz2. You may also make some savings by re-packing zips as tar.gz
files, and even more as tar.bz2
Nicholas Clark
--
Even better than the real thing:http://nms-cgi.sourceforge.net/
disk space. It's less painful than deleting files.
But a lot slower, as bzip2 is somewhat slower than gzip. But putting lots
of CPU once into compressing more seems to be the correct trade for any
write-once, read-many archive.
Nicholas Clark
--
Even better than the real thing:http://nms-cgi.sourceforge.net/
earching for something else at the top level
should be smart enough to figure out that it's not what they are looking for.
Nicholas Clark
--
Even better than the real thing:http://nms-cgi.sourceforge.net/
t found that a problem.
Loading POSIX itself is slow *by default*
$ time perl -e 'use POSIX'
real0m0.060s
user0m0.060s
sys 0m0.000s
One speed (and slight size reduction) trick is to use an explicit import list
$ time perl -e 'use POSIX qw(log10)'
real0m0.043s
user0m0.020s
sys 0m0.020s
Nicholas Clark
ation detail, but I'm told that others don't agree with this,
and that a lot of existing CPAN practice puts things into the Tie namespace.
You can't win. :-)
Nicholas Clark
. The
strtoll() function returns the result of the conversion, unless the value
would underflow or overflow. If an underflow occurs, strtoll() returns
LLONG_MIN. If an overflow occurs, strtoll() returns LLONG_MAX. In all
cases, errno is set to ERANGE.
strtoul probably does what you want. hex will even try to use doubles if
it runs out of space in an unsigned long.
Nicholas Clark
/search?m=module&q=Math&s=1&n=100
http://search.cpan.org/search?m=module&q=Algorithm&s=1&n=100
Nicholas Clark
PS Surely the choice is Maths::NearestNeighbour vs
Math::NearestNeighbor :-)
ve a use for this package? :)
Possibly. IIRC Autrijus Tang was looking for a pure perl way to replace
the resource icon on an exe file generated by PAR.
I may have some of this wrong, but not the "pure perl" and "replace icon"
bits.
Nicholas Clark
day?
I can see this as a really good way to piss people off.
Anyway, it's moot point as you already know that the person assigned to
VERP takes an awfully long time to getting a round tuit, so it's unlikely
to be finished soon ( http://siesta.unixbeard.net/svn/trunk/siesta/TODO )
Nicholas Clark
quent real, serious, messages.
Nicholas Clark
unity seems to be bad at finding suitably descriptive
names for things, given the seeming inability of later programmers to
find them, and save re-inventing wheels.
The first solution that sprang into my mind was a "CPAN cookbook" but who
wants to write that?
Presumably there are better solutions.
Nicholas Clark
mplementation is not the
most important feature of these modules - why is it the most important
part of their names?
Nicholas Clark
believe that Ingy is working on something]
Nicholas Clark
On Thu, Nov 13, 2003 at 10:52:16PM +, Simon Cozens wrote:
> The module's already uploaded, guys; the thread is dead.
The module was uploaded within 3 hours of you posting the original message.
(but that doesn't seem to have stopped anyone for the past 2 days)
Nicholas Clark
do think
that the discussion is appropriate to this list, as getting this list to
work out the the best way making this change will help future authors.
This won't be the last time someone asks the list this question for an
existing module, and the recommendation is to make this sort of rename.
Nicholas Clark
trieve's
> C function BTRCALL() in perl as BTRIEVE::XS::Call().
> Are these namespaces ok?
IT STRIKE ME THAT ALL UPPER CASE DOESN'T SEEM VERY perl LIKE.
Also, does it need to be at the top level? What is Btrieve? Could your
module sensibly live one level down, even if it means creating a new top
level for it?
Nicholas Clark
ot; string and decide what to return (string, or object ref)
Nicholas Clark
kes life easier trying to keep
track of files in dual-life modules. The other important bit is to actually
change the version number if anything in the file changes.
But I'm biased. :-)
Nicholas Clark
m wary about using "Ham" as a top level as its meaning may not be
clear to everyone. Whilst HamRadio seems much nicer because it is
unambiguously descriptive, is there any reason not to use plain "Radio" ?
Presumably non Ham radio users could also be using any of the equipment
that hams use, for much the same purpose?
Nicholas Clark
e passing all bug reports over to its author Lincoln Stein, who is
actively maintaining it on CPAN.
Nicholas Clark
interested to see
> some real examples where #3 has occured.
Core perl managed it. 5.8.1 accidentally was not binary compatible with
5.8.0. 5.8.2 manages to be compatible with both 5.8.0 and 5.8.1
Nicholas Clark
only 2 functions from Compress::Zlib's API actually need
implementing - inflateInit and inflate.
I'm partway through this, and I'm wondering what the name should be.
Autrijus suggested Compress::Zlib::PurePerl, which I think is reasonable.
Are there further suggestions?
Nicholas Clark
On Mon, Feb 16, 2004 at 10:43:27AM +1300, Sam Vilain wrote:
> On Mon, 16 Feb 2004 10:19, Nicholas Clark wrote;
>
> > Autrijus suggested Compress::Zlib::PurePerl, which I think is
> > reasonable.
>
> ...but it doesn't use Zlib! :) Compress::Gzip?
But it does
back somewhat to work out
which module will need its README reading. It would seem that the quick
change of rephrasing to add the name the module would save users' time.
With this approach, how many subsequent releases of the module do you make
before you drop this text?
Nicholas Clark
eports the configuration for your perl you're having problems
with). The bug report will get to the perl5-porters list, who ought to see
it as I think it's almost a bug in perl (and how perl does things).
That list is also more likely to suggest a good way to fix it.
Nicholas Clark
ng else
already.]
No-one is stopping you setting up a ratings system.
Nicholas Clark
whether the source to it is available, and if so, where to get
it. But resolving that seems to be the first step.
Nicholas Clark
";
printf("'%s' is %g\n", foo, atof(foo));
return 0;
}
on a newer system with a C99 libc gives:
'0x3' is 3
whereas an old system with a C89 libc gives:
'0x3' is 0
A competent standard would have defined a function with a new name for the
new behaviour.
Nicholas Clark
ated DATA in memory and then feed it out at the
pace dictated by select() and non-blocking IO?
> Planning to call this cute beastie HTTP::Server::Selecting.
because I'm not convinced that this name would give much insight into
the implications of the implementation, and the trade offs they give.
Nicholas Clark
deferred according to the best practices of Extreme
> Programming.
Presumably generating pipes to helper tasks can be done with a module on
CPAN? And is it really going to achieve faster non-blocking file IO on
anything other than Unix (or Unix-a-likes) where there will be cat?
Nicholas Clark
wn
improved substitute for search.cpan.org. (although it helps if you have
a public facing webserver if you want to show it to others).
Yet no-one does.
Until someone does, nothing will change. No-one on this list is preventing
anyone from trying this.
Nicholas Clark
name.
Doing this stub with clear descriptive text above seems to me to be the
best way to do it.
Nicholas Clark
an_d_foy/journal/8786
Nicholas Clark
g DB)
$dsn .= lc ";dbname=$dbspec->{db_name}" if length $dbspec->{db_name};
$dsn .= ";port=$dbspec->{port}" if defined $dbspec->{port};
} else {
$dsn .= ":port=$dbspec->{port}" if defined $dbspec->{port};
}
Nicholas Clark
On Tue, Nov 30, 2004 at 10:32:12PM +, Tim Bunce wrote:
> On Tue, Nov 30, 2004 at 09:38:47PM +0000, Nicholas Clark wrote:
> > On Tue, Nov 30, 2004 at 08:53:51PM +, Tim Bunce wrote:
> > The one that I've hit - specifying port and host, Pg vs Mysql (and SQlite):
&g
s a kind offer, but does your server reach me when I'm working offline?
Doubtful.
But I have a local CPAN mirror, and run local database servers for
development.
Nicholas Clark
etc are split out, so that it is easy to make a single line change that is
visibly just a server (staging vs. live) while keeping all the other parts
the same, or just a port, so that development machines can run server per
developer.
Nicholas Clark
were they Whitworth, BSF, UNC or something else?
Nicholas Clark
password provided. If PAUSE accepts it, then you know
it's the real thing.
I can't think of a future-proof way of avoiding IDs you allocate conflicting
with PAUSE, unless you and Andreas collaborate
Nicholas Clark
On Fri, Feb 04, 2005 at 02:40:09AM +0200, Gabor Szabo wrote:
> On Wed, 2 Feb 2005, Nicholas Clark wrote:
>
> >The same hack as rt.cpan.org uses - attempt a login on pause.cpan.org
> >using the ID and password provided. If PAUSE accepts it, then you know
> >it's th
e utilities are archivers, would it work to put them in the Archive::
top level namespace, alongside Archive::Tar and Archive::Zip ?
Nicholas Clark
xcept that this CPAN mirror doesn't delete things.
Nicholas Clark
1 - 100 of 104 matches
Mail list logo