Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-07-29

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
05:18 Meiermann joined #pdl
10:03 lungching joined #pdl
11:23 run4flat Mithaldu, that's because logic in PDL is per-element
11:23 run4flat what you meant to say was, "return if not defined $instance"
11:23 run4flat right?
11:23 run4flat :-)
11:24 run4flat that, or you meant, "return if not all $instance", which checks that all members of $instance are true
11:29 Mithaldu run4flat: there's some discussion on it on the mailing list
11:30 run4flat heh
11:30 Mithaldu personally i'd say that bit of implementation is broken
11:30 run4flat and what should it do?
11:31 Mithaldu if $pdl contains a PDL object, then it should act to truthiness interrogation the same as any other object
11:31 run4flat but what is truth?
11:31 run4flat :-)
11:31 run4flat at least, in the context of a piddle?
11:31 Mithaldu anything that is not:
11:31 Mithaldu no, not in the context of a piddle
11:31 Mithaldu i'm asking a scalar
11:31 Mithaldu a scalar is untrue if it's undef, the literal number 0, or an empty string
11:32 Mithaldu a PDL object is none of those
11:32 run4flat sure
11:32 run4flat so a piddle should always be true
11:32 run4flat what if it's an empty piddle?
11:32 run4flat still true?
11:32 run4flat or should it mimic the behavior of an array?
11:32 Mithaldu is an empty arrayref true?
11:32 run4flat yes
11:32 Mithaldu principle of least surprise applies
11:32 run4flat 'tis, I believe
11:32 Mithaldu it is
11:33 run4flat the reason why it's a bad idea to have "any" piddle true, is this scenario:
11:33 run4flat if ($first_piddle == $second_piddle) ...
11:34 run4flat that's a typo
11:34 Mithaldu ?
11:34 run4flat it should have been either "if (any $first_piddle == $second_piddle) ..." of "if (all $first_piddle == $second_piddle) ..."
11:34 run4flat you're checking for element-by-element equality
11:34 Mithaldu if ($first_piddle == $second_piddle) ... <- that is true if they're the same object
11:34 run4flat PDL overloads the == operator and returns a piddle
11:34 run4flat no
11:34 run4flat that's useless
11:34 run4flat it's true if they have the same values
11:35 run4flat er
11:35 run4flat oops
11:35 run4flat sorry,
11:35 run4flat spoke too quickly
11:35 run4flat the point is that the programmer *almost* *always* typod in that case
11:36 Mithaldu ok, fine, sure, i don't mind overloading there, but that seems to have nothing to do with !$pdl
11:36 run4flat ok
11:36 run4flat then I don't think I've been quite clear
11:36 Mithaldu that's what i'm saying :)
11:36 run4flat $result = $first_piddle == $second_piddle
11:36 run4flat that returns a piddle
11:36 run4flat now
11:37 run4flat what does if($result) do?
11:37 run4flat is it if( any $result) ?
11:37 run4flat or if( all $result) ?
11:37 run4flat because, ultimately, this decision determines the fate of !piddle
11:37 Mithaldu none of those
11:38 Mithaldu equality has nothing to do with truthiness
11:38 run4flat so it should always be true?
11:38 run4flat no no
11:38 run4flat think more fine grained here
11:38 run4flat equality returns a piddle
11:38 run4flat we want comparison operators to be useful
11:39 run4flat that's actually more important than least-surprise
11:39 run4flat because it's incredibly powerful
11:39 Mithaldu you mean clever and convenient
11:39 run4flat I mean expressively powerful
11:39 run4flat my scientific code is cleaner for it
11:39 run4flat and I laugh at other languages that don't support it
11:40 Mithaldu well, in any case
11:40 Mithaldu i don't see where the connection is
11:40 Mithaldu because, if i understand it right, you want to do something like this:
11:40 Mithaldu my $result = $first_piddle == $second_piddle;
11:40 Mithaldu print 1 if all $result;
11:41 Mithaldu correct?
11:41 run4flat well, I can do something like this presently
11:41 run4flat correct
11:41 Mithaldu what does that have to do with:
11:41 Mithaldu print 1 if $result;
11:41 run4flat catching bugs
11:41 Mithaldu currently it dies
11:42 run4flat I see your point, I was a little off track
11:42 run4flat because PDL can have the all/any behavior and have a defined boolean value
11:42 run4flat but, this helps catch bugs
11:42 Mithaldu i mean, i kind of see the point
11:42 Mithaldu if you have a scientist who doesn't know shit about computers it'll yell if he does things wrong
11:42 Mithaldu but on the other hand, you break compatibility with the rest of perl
11:43 run4flat sure
11:43 Mithaldu because there is literally nothing else that dies when you test its truthiness
11:43 run4flat now, bear in mind, that's a cultural/conventional thing
11:43 run4flat not a language thing
11:43 run4flat and I have found myself making that typo more than once
11:43 run4flat and I know lots of CS
11:43 run4flat :-)
11:44 Mithaldu i know zero cs
11:44 Mithaldu in fact
11:44 Mithaldu i failed out of polytech before i could get to that stuff
11:44 run4flat I find it unlikely that you know zero cs. :-)
11:44 Mithaldu the difference here is one of craftmanship
11:45 run4flat hmm, not quite sure what you mean
11:45 Mithaldu i believe it would benefit PDL if there was a way to say strict mode
11:45 run4flat that is a much larger issue, with which I agree
11:45 run4flat but lexically scoped warnings/dieing is a big project to undertake
11:45 Mithaldu i wouldn't make that kind of error, because when i see:
11:45 Mithaldu if $thing
11:46 Mithaldu then my mind has ONLY the default perl truthiness test for a scalar in it
11:46 run4flat of course
11:46 run4flat and, honestly, any Perl programmer with enough experience would think that
11:46 run4flat I think that
11:46 Mithaldu if yours doesn't then that is only because you're less familar with perl
11:46 Mithaldu not unfamiliar, just less
11:46 run4flat yes, we agree
11:47 run4flat but it's so easy to think "if they're all the same" and then write PDL code that says (if $a == $b)
11:47 run4flat er, if ($a == $b)
11:47 Mithaldu personally i'd just love to have a NO MAGIC version of PDL
11:47 run4flat hmm
11:48 Mithaldu the way it exports asin and other such functions surprised me way too much because it does that without asking me
11:48 run4flat I hate that
11:48 * run4flat sighs
11:48 Mithaldu or rather: without having been asked to
11:48 run4flat I hate that it doesn't give you any fine-grained controll over function exports
11:48 run4flat but it's damn near impossible to solve in backwards-compatible way
11:48 run4flat and I just haven't put the time in to find a solution
11:48 run4flat that's still backwards compatible
11:48 Mithaldu PDL2
11:48 run4flat Sure
11:49 run4flat and then there's a whole bunch of other things we'd like to do
11:49 run4flat and I don't have time to implement those
11:49 run4flat it's an issue of manpower in that case
11:49 run4flat but I think we all agree with you
11:49 run4flat Liberal use of @EXPORT is bad
11:49 Mithaldu get PDL on github :)
11:49 run4flat it is
11:49 Mithaldu https://metacpan.org/release/PDL <- doesn't link to it :/
11:49 run4flat but I got tired of being the *one* guy keeping it current with the master on sourcefoge
11:50 run4flat no it doesn't
11:50 run4flat because it's an unofficial mirror
11:50 Mithaldu well when i say "get pdl on github" i mean, deprecate sf, period
11:50 run4flat (the one on github)
11:50 run4flat Mithaldu, if you could do that, I would be happy
11:50 run4flat Joel and I were both pushing for it
11:50 Mithaldu pdl is only hurting itself
11:50 run4flat is it?
11:50 Mithaldu by not being on github? yes
11:50 run4flat all of the developers are happy using sourceforge
11:50 run4flat are yo usure?
11:51 run4flat are you absolutely sure?
11:51 Mithaldu yes
11:51 Mithaldu yes
11:51 run4flat what would you contribute?
11:51 run4flat seriously
11:51 Mithaldu if it were on github i could send you a patch without even needing to check it out to my hdd
11:51 run4flat putting a huge project on github doesn't solve the problem
11:51 run4flat we need to put it on github
11:51 Mithaldu let me put it like this:
11:51 run4flat and then have a clear process for moving ideas and issues through
11:51 Mithaldu http://pdl.git.sourceforge.n​et/git/gitweb.cgi?p=pdl/pdl
11:52 Mithaldu this is so unbelievably shitty that i cannot bring up the care to even look through it
11:52 run4flat so, in other words, the barrier to your contribution is too high
11:52 Mithaldu and that's with me knowing that "tree" means "yo the files are here brah"
11:52 run4flat right
11:52 run4flat initial contributions need to be super-easy
11:52 Mithaldu imagine how it's like for someone who doesn't know git
11:52 run4flat I know
11:52 run4flat I agree
11:53 run4flat so,
11:53 run4flat do you want me to move PDL to github, or figure out the function exporting thing?
11:53 run4flat 'cause I only have time for one, not the other
11:53 run4flat (actually neither, at the moment)
11:53 run4flat or, the logic issue?
11:53 Mithaldu well what's stopping the move to git? :)
11:53 run4flat inertia
11:54 run4flat It's so easy for any one developer to throw a wet blanket on it
11:54 run4flat because it's easy to find minor but good reasons not to move
11:54 Mithaldu ah
11:54 run4flat chm said we needed web hosting, for example
11:54 Mithaldu so you need someone like me who'll happily argue everything to death?
11:54 run4flat to disprove this, jberger rewrote all of pdl.perl.org using jekyl
11:54 run4flat no
11:55 run4flat no, not argue
11:55 run4flat do
11:55 Mithaldu well yes
11:55 run4flat that amazing thing was that, after jberger did that, chm directly change his stance
11:55 run4flat he just kinda went silent on it
11:55 Mithaldu ok
11:55 run4flat there are others who argue that sf's new interface is really nice
11:55 run4flat basically, few of the key PDL developers are active on github
11:55 run4flat that's the problem
11:56 run4flat get them to start working on it
11:56 run4flat then it'll happen
11:56 run4flat at least, I think that's the problem
11:56 run4flat another minor but good point is that sf has a business model that works
11:56 Mithaldu is pdl.perl.org just an alias to a sourceforge site?
11:57 run4flat and one of the devs fears that github hasn't proven itself
11:57 run4flat yeah
11:57 run4flat I think so
11:57 Mithaldu ok, so we can alias it to github hosting just as well
11:57 run4flat I think it's pdl.sourceforge.net
11:57 run4flat we can
11:57 run4flat and jberger went through the effort of porting all the pages
11:57 Mithaldu right
11:57 run4flat cogitate on that for a sec
11:57 run4flat where is jberger ?
11:57 run4flat jberger++
11:57 Mithaldu haha
11:58 Mithaldu and the business model of github is solid, they get away with selling server images for 5k$ a year
11:58 run4flat oh really?
11:58 Mithaldu https://enterprise.github.com/
11:58 Mithaldu minimum price is 5k
11:59 run4flat huh
11:59 run4flat well, the argument isn't that github can't make money
12:00 run4flat it's that sf.net has changed hands a lot, but it's still around
12:00 run4flat github looks like it is doing great with this, but that can be eaten-away by an open-source project
12:00 Mithaldu that's a totschlagargument
12:00 * run4flat looks up what Mithaldu just said
12:00 Mithaldu http://en.wikipedia.org/wik​i/Thought-terminating_cliché
12:01 run4flat hmm
12:01 run4flat It seems like a valid argument to me
12:02 run4flat but, then again, maybe taking the long-view with hosting service is the wrong optimization
12:02 Mithaldu it isn't
12:02 Mithaldu the problem is
12:02 run4flat maybe growing the developer base is a better solution
12:02 Mithaldu he is saying "it could possibly die"
12:02 run4flat than finding a stable hosting platform
12:02 Mithaldu has no indication that that is about to happen
12:03 run4flat well, it's not going to happen in the next couple of years
12:03 run4flat that's for sure
12:03 Mithaldu and nobody can explain that there is no feasible way that it could die
12:03 run4flat it'll take a while for github to die at this point
12:03 Mithaldu i could claim the same about sourceforge
12:03 Mithaldu less and less people are using it
12:03 Mithaldu it could die any minute now
12:03 Mithaldu etc.
12:03 run4flat right
12:04 run4flat actually, if we want stability, we should move to Google code
12:04 run4flat since that's got a nice corporate backing
12:04 run4flat :-)
12:04 run4flat but you see how easy it is to not move to github?
12:04 run4flat one person expresses a concern
12:04 run4flat and it quiets the discussion
12:04 Mithaldu they tell you to use github
12:04 Mithaldu see, that's what i mean
12:04 Mithaldu i'll make a fucking list of opposition points
12:04 Mithaldu and destroy all of them
12:05 run4flat ha
12:05 Mithaldu and then show them that list and DARE them to add more
12:05 Mithaldu (it'll be in a github issue)
12:05 run4flat haha
12:05 run4flat well, you might want to update, or let me update, the github version so it's up-to-date with the sf.net one
12:06 Mithaldu unrelated
12:06 run4flat hehe
12:06 run4flat ok
12:06 run4flat by all means
12:06 Mithaldu is the pdl porters mailinglist a good place to show pdl scripts and ask people to tell me how i should rewrite them for speed and readability?
12:06 Mithaldu i was actually more interested in recent developments
12:06 Mithaldu that's why i asked
12:07 run4flat yes
12:07 run4flat well,
12:07 run4flat sorta
12:07 Mithaldu but apparently recent means last april?
12:07 * run4flat is confused
12:07 Mithaldu let me guess
12:07 run4flat last april?
12:07 Mithaldu PDL3 is a separate repo?
12:07 Mithaldu http://pdl.git.sourceforge.net/gi​t/gitweb.cgi?p=pdl/pdl;a=shortlog
12:08 run4flat no, I think it's just a separate branch
12:08 Mithaldu no newer branch than master: http://pdl.git.sourceforge.net/g​it/gitweb.cgi?p=pdl/pdl;a=heads
12:08 run4flat yes
12:08 run4flat PDL3 is backwards compatible
12:08 run4flat so PDL2 is parked on its own branch
12:08 run4flat and PDL3 is moving forward in main
12:08 run4flat er, master
12:08 Mithaldu ok
12:08 Mithaldu but still
12:09 run4flat sorry, had to dust out the cobwebs
12:09 Mithaldu nobody has worked on things in the past 3 months
12:09 run4flat I haven't been following PDL development for a little while now
12:09 run4flat :-/
12:09 run4flat too busy with a new job
12:09 Mithaldu yeah, stuff happens
12:09 Mithaldu only reason i'm playing with pdl now is that i restarted porting an opengl tutorial
12:09 run4flat oh, cool
12:11 run4flat btw, I'm replying to the email
12:17 Mithaldu cool :)
12:21 Mithaldu run4flat: if you're interested: https://github.com/wchristian/ogldev/tree/perl
12:31 jberger o/
12:31 jberger I've got to leave for the train, but I can hop on from my phone when I'm on it
12:31 jberger github++
12:35 jberger joined #pdl
12:39 run4flat ok, sent the email
12:39 run4flat And now, actually, I must get to work
12:40 run4flat :-(
12:40 run4flat I'm sad I can't spend more time hacking on Perl stuff
12:40 run4flat but I'm happy that I like my work again. :-)
12:48 jberger run4flat++
12:48 run4flat thanks
12:48 jberger \o/
12:48 run4flat how goes it, jberger
12:48 run4flat ?
12:49 jberger Chicago isn't three same without you through
12:49 run4flat if I read that right, then... well thanks
12:49 run4flat but now you have a whole bunch of Python friends to hang out with at BofA! :-D
12:49 jberger (That is a guy with his arms in the air)
12:50 run4flat is the guy with arms in the air trying to type?
12:50 run4flat hehe
12:51 jberger Typing on this thing is TOUGH
12:51 run4flat phone?
12:51 run4flat new phone?
12:52 jberger Yeah
12:52 jberger First smartphone
12:52 run4flat nice
12:52 jberger Swipe typing is great when it works
12:52 run4flat heh
12:52 run4flat yeah...
12:53 run4flat I'll tell you, though, having interactive maps in my pocket is awesome
12:54 jberger The by gist from earlier is just that Mithaldu would like to see a move to github?
12:54 run4flat yes
12:54 run4flat and I pointed out that there is a lot of inertia for sf.net
12:54 run4flat like the web pages
12:54 jberger (<3 maps)
12:55 jberger Yeah
12:55 jberger Well you know my feelings
12:55 run4flat I do
12:55 run4flat and with Mithaldu here, that makes three of us. :-)
12:55 run4flat I guess it's best to say that the PDL developers don't want to move without a consensus
12:56 run4flat that's probably the most diplomatic way to put it
12:56 run4flat so if anybody raises concerns, it puts the breaks on the whole thing
12:56 jberger At work they use CVS, so now I love git and github even more
12:56 run4flat haha
12:56 run4flat seriouslY/
12:56 run4flat ?
12:56 run4flat but mercurial is written in Python
12:56 run4flat it seems like they should fall in love with it
12:57 jberger Yeah, welll :-P
12:57 run4flat yeah, I know
13:00 Mithaldu run4flat: *brakes
13:00 Mithaldu ;)
13:00 run4flat haha thanks
13:00 run4flat :-)
13:00 Mithaldu well, i'm german
13:00 Mithaldu grammar nazi is in my blood
13:01 run4flat haha
13:01 run4flat oh, I should have said "It puts the *coffee breaks on the whole thing"
13:02 run4flat i.e. when somebody raises a concern, we all get up from our PDL work and get coffee
13:02 run4flat thus slowing down the whole process
13:02 run4flat :-)
13:02 run4flat yeah
13:02 run4flat that's what I meant to say
13:02 jberger Hehe
13:02 run4flat of course
13:04 run4flat alright, well, my computer just decided that the HD is read-only. I'll have to restart, and cross my fingers that it didn't destroy everything
13:05 Mithaldu sounds like an ssh
13:05 Mithaldu *d
13:05 Mithaldu and at that point i remember his usb thingy
13:06 jberger Good luck!
13:07 run4flat yeah, I may need to use a live-cd to retrieve my files
13:10 jberger Yuck
13:10 jberger Oh well at least there is that
13:49 Mithaldu agh, what the christ
13:50 Mithaldu C:\Windows\System32>perl -e "use PDL; print pdl() ? 1 : 0;"
13:50 Mithaldu 0
13:50 Mithaldu C:\Windows\System32>perl -e "use PDL; print pdl(1) ? 1 : 0;"
13:50 Mithaldu 1
13:50 Mithaldu C:\Windows\System32>perl -e "use PDL; print pdl(1,2) ? 1 : 0;"
13:50 Mithaldu multielement piddle in conditional expression at -e line 1.
13:51 Mithaldu C:\Windows\System32>perl -e "use PDL; print pdl(0) ? 1 : 0;"
13:51 Mithaldu 0
13:51 Mithaldu C:\Windows\System32>
13:51 Mithaldu that is php levels of inconsistent
14:28 pdurbin heh
15:17 run4flat joined #pdl
15:17 run4flat o/
15:17 run4flat now from a Debian machine. :-)
15:18 Mithaldu welcome back o/
15:19 Mithaldu also sent an email to the list, expecting to be banned from it any moment now :v
15:43 Mithaldu wow, there's so much wrong with that answer, i don't think i can be arsed to try and answer further
15:44 Mithaldu chm: one point of note: i did draw from a quote, specifically something you said, please don't ever assume i just made something up without asking me where i got it
15:45 Mithaldu &
15:50 run4flat Mithaldu, Mithaldu, it's OK
15:50 run4flat :-)
15:52 run4flat Re quotes: if you have a specific message or email, it helps a LOT in these conversations to find a link to the email archives
17:07 Mithaldu run4flat: i didn't think that necessary, because i was replying to the email that contained the sentence "since a pdl object is intrinsically multi-valued, it isn't possible to shoehorn a piddle into T/F options."
17:08 Mithaldu also, no worries, i don't see it as flames, it was more of a "sigh, i really don't care enough to write a lot of words here"
17:17 run4flat haha
17:17 run4flat I know
17:17 run4flat I just like to err on the side of simmering things down
17:18 run4flat as for the quote, now I see what you mean
20:47 chm joined #pdl
20:51 chm Mithaldu: re "one point of note: i did draw from a quote, specifically something you said, please don't ever assume i just made something up without asking me where i got it"
20:53 chm Err, you actually refered to "PDL Porters however seems to think..." which doesn't contain any reference to either to author of the quote (apparently me) or the quote itself.
20:54 chm Also, didn't assume you made anything up, just that helps to keep list conversations tied to specifics: what I said, what you think that means, where you think things are going wrong....
20:55 Mithaldu look, i don't know how to put it more simple
20:56 Mithaldu if i say something that doesn't make sense, send me a message containing a question mark
20:56 Mithaldu right now the paragraph heavily implies that i made that up because of some *opinions* i have
20:56 Mithaldu asking is also less typing :P
20:57 chm That was definitely not the intention.
20:58 Mithaldu good to hear :)
20:58 chm In the case of the "PDL Porters however seems to think..." it would have been totally
20:58 chm clear if you had either disagreed specifically with the quoted text (I may have been using loose terminology but nothing incorrect),
20:59 chm or you could have replied in the list with the quote.  As it was, I was not aware that you might have interpreted my statement differently than I had intended.
21:00 chm In fact, if I hadn't checked the #pdl channel, I would have had no idea that I was in the middle of a flame
21:00 Mithaldu you're not
21:00 chm war and not an animated discussion.
21:01 chm That was hyperbole on my part but with the out-of-band discussion, I was unable to clear things up right away.
21:01 chm Nor am I clear what "wow, there's so much wrong with that answer, i don't think i can be arsed to try and answer further" refers to but
21:01 chm it seems likely that is related to the previous discussion.
21:02 chm Sorry about any misunderstandings that I may have generated.  As I mentioned, the different points of view and feedback from "outside" users
21:02 chm is the only way PDL will continue to evolve in a good way.
21:04 Mithaldu it refers to your last email, which has so many subtle points i would need to adress that i'd need to write a LOT of words, especially when i'm not in a realtime settings
21:04 Mithaldu and well
21:04 Mithaldu i don't care all that much about the boolean stuff at this point
21:05 Mithaldu also i hate mailing lists because they require me to play a mental game of chess with an imaginary chm, so that what i write wouldn't be wasted
21:05 chm Why don't you liked the die always for pdl in a boolean expression?  That allows the interface to be the same and alerts real perlers that something is different about PDL...
21:06 Mithaldu i do like it
21:06 chm Then maybe we could defer some of the fine points until we're in a more interactive environment.
21:07 Mithaldu yep :)
21:07 chm I have problem communicating in #pdl.  Better in list since I can think about the reply but it is clunky and prone to "explode" due to lack of eye contact...
21:08 Mithaldu you've just as much time in irc
21:08 Mithaldu it is asynch
21:08 sivoais wouldn't it be possible to wrap the PDL instance for the times you want interface with other code. Something like PDL::AsObject->new( $dying_pdl )?
21:08 Mithaldu however it's easier to ask a clarifying question
21:09 Mithaldu sivoais: totally possible, but it's a workaround
21:09 Mithaldu (the implication here being that it is only one of a number of long-term suboptimal solutions)
21:10 * sivoais nods
21:10 chm The catch is that PDL has 10 years with the current implementation.
21:11 chm I think it is critical that PDL not break for the users who depend on PDL to "get the job done"
21:12 chm It is clear to me from all this discussion that we *must* improve the documentation regarding this so
21:12 Mithaldu chm: let me summarize it like this:
21:12 chm that, at least, it won't catch any perler's with default assumptions that break with PDL.  :-)
21:12 Mithaldu PDL suffers, as a microcosm, from all the problems that result in Perl getting flak from other languages
21:12 Mithaldu chiefly: it's clever as all get-out
21:16 chm I'm not familiar with the flak from other languages at Perl, but PDL has always had problems with being a perl module (requiring one to know some perl with its arguable flaws and complexity) but it is also a computational tool for researchers and scientists.
21:17 chm That gives lots of inconsistent and incompatible requirements and desires.  This is made worse by the fact that the PDL developers are usually very conversant with Perl and so become blind to complexities because they aren't complex to them any more...
21:17 Mithaldu the two main complaints are inconsistency and implicity
21:19 chm I think it would difficult to find a PDL user or developer who doesn't think PDL is too inconsistent.  The problem is it is easier to point that out than to grind through the code in a clean-up pass---while preserving usability and compatibility.
21:20 chm As for implicity, I think specifics would help here but feedback from new users on where they're having difficulties could help.
21:20 chm It's been nice chatting.  I have to head out now.  o/
21:25 Mithaldu inconsistency: type conversion is done with float, long, etc.
21:25 Mithaldu but those are also methods?
21:25 Mithaldu and crossp is a thing, but is not a method?
21:26 Mithaldu and set is a function as well, instead of a method by default
21:26 Mithaldu printing a piddle stringifies it, but there doesn't seem to be a dump or a stringify method?
21:26 Mithaldu implicity:
21:27 Mithaldu use PDL; is a horror :P
21:27 Mithaldu and the whole override stuff falls under that too
21:28 Mithaldu at the very least there needs to be a clearly labeled document detailing only the overrides, so users are aware of what mundane operations might subtly break
21:31 chm joined #pdl
21:32 Mithaldu haha, wb
21:32 Mithaldu i was just rambling for the log so you'd see it later
21:32 chm type conversion routines are both methods and functions
21:32 chm same for crossp
21:32 chm same for set, although the preferred usage now it the slice or niceslice syntax
21:32 Mithaldu hmmm, maybe i got that crossed because i was expecting ->cross to work
21:33 chm stringifying a pdl is not something you usually want to do, if so "$pdl" will do it.
21:33 Mithaldu actually it is
21:33 Mithaldu look
21:33 chm You'll need to up the element count limit if you're trying to dump a big one.
21:33 Mithaldu i wasn't looking for answers really :P
21:33 Mithaldu i was telling you what i stumbled over
21:34 Mithaldu so you can figure out how the docs can be improved
21:34 chm Yes, that is the kind of stuff that is nice to have in the mailing list or as feature requests.
21:34 chm However, I think most of those can be summarized as your last one: use PDL; is a horror :P
21:34 Mithaldu heh
21:35 chm One goal of PDL3 is to regularize the import/export for PDL.  As usual, things take much longer than one would like.
21:35 Mithaldu well, question here:
21:35 Mithaldu git says there's not been a change in 3 months, true?
21:36 chm Are you sure you've got the correct repo?  The sf.net migrated to a new layout which changed the URL.
21:37 Mithaldu haven't a clue actually, i clicked on the cpan link
21:37 Mithaldu it might be wrong
21:38 Mithaldu yep, looking at the wrong one
21:38 chm Well, that is another example of why it is helpful for users to actually give feedback rather than holding it in and possibly becoming grumpy.
21:38 chm I think the URL needs to be updated in the release meta info.
21:38 Mithaldu hey you noticed i'm happy to talk AND grump :P
21:39 chm Well, you lured me back but I really have to go.  Cheers.  o/
21:39 Mithaldu ciao o/
22:14 run4flat whoah
22:20 run4flat lots of chatting
22:20 run4flat glad things got more-or-less resolved. :-)
22:21 run4flat alright, I'm outa here
22:21 run4flat o/
23:46 jberger_ joined #pdl

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary