#foswiki 2012-02-24,Fri

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)

WhoWhatWhen
flexibeastOkay, so now i declare myself to be stuck: i can't work out how to save form data edited via the FlexFormPlugin %RENDERFOREDIT% macro. [00:33]
gac410I have not used that one flexibeast. [00:36]
flexibeastgac410: Fair enough. - But when one saves a topic, isn't it usually the case that form data associated with it automatically gets saved also? [00:37]
gac410yes. does the topic already have a form attached? Or are you trying to add a new form to the topic. [00:38]
flexibeastThe topic already has form data attached.
Which i can edit and save in the usual way.
[00:40]
pharveyflexibeast: did you wrap it like so: <form action="%SCRIPTURLPATH{"save"}%" method="post">ENDERFOREDIT{...}</form>
<form action="%SCRIPTURLPATH{"save"}%" method="post">%RENDERFOREDIT{...}</form>
[00:40]
flexibeastpharvey: Pretty much; i'll pastebin it.
http://pastebin.com/RBrw1CCS
The form is being filled with the correct data.
[00:41]
pharveyflexibeast: why is URLPARAM{"taskid"}% used as the topic-to-edit, but then you POST to %WEB%.%TOPIC%? [00:45]
flexibeastBut then if i add text to one of the not-yet-filled fields, and select the "Save changes" button, it appears to go through the save process - it certainly returns me to the TaskList page - but when i reopen that topic, it's not there. [00:45]
pharveynormally the action web.topic should be the same as the RENDERFOREDIT topic [00:45]
flexibeastWell, in this case they're the same thing.
i have a task list, and want to create a custom page for editing task details.
(i hope to make this a Plugin or Contrib once i get this working.)
[00:46]
pharveyflexibeast: can you see that the save'd topic has an updated time modified-datestamp? [00:47]
flexibeasti'll check now.
Hm, well, it doesn't appear at the top of the relevant WebChanges page ....
[00:47]
Okay, modified-timestamp is definitely not getting updated. [00:56]
pharveyflexibeast: check in working/logs/events.log to see *which* topic is getting updated [01:00]
flexibeastActually, it appears you've solved the issue. :-)
i changed the POST URL.
[01:00]
pharveyOH
of course
%WEB% %TOPIC% will be the topic in which you've written the <form markup
[01:01]
flexibeastTo:
%SCRIPTURLPATH{save}%/%WEB%/%URLPARAM{"taskid"}%
That worked. :-) Thank you!
[01:01]
pharveycool
np
[01:01]
................... (idle for 1h32mn)
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/wiRySw
[foswiki/Release01x01] Item11555: unit test for rename failure - GeorgeClark
[02:33]
***GithubBot has left [02:33]
FoswikiBothttp://foswiki.org/Tasks/Item11555 [ Item11555: Square bracket backlinks in format Web/Topic don't get renamed ] [02:33]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/ZwvGNw
[foswiki/master] Item11555: unit test for rename failure - GeorgeClark
[02:33]
***GithubBot has left [02:33]
....... (idle for 33mn)
flexibeast has left "PING 1330052787" [03:06]
.... (idle for 16mn)
SvenDowideitok, can someone please beat me with a stick? [03:22]
gac410to far away [03:22]
SvenDowideitI just noticed (as i do every 2 years or so) that Tasks still uses regex searches [03:22]
gac410Probably a lot of that left around. [03:23]
SvenDowideitbut but but :)
SvenDowideit resolves to write an entirely new task management app
'one day'
[03:24]
gac410right after world peace [03:24]
SvenDowideitfirst i need to work out how to make youTubVideos
world peace? no way!! think of the economy!!
[03:24]
gac410whirled peas? [03:25]
SvenDowideitoooooo, nice idea!
get a lettuce spinner, put in peas and some tasty flavourings
and taadaaa
SvenDowideit likes the name
[03:25]
gac410Old cartoon I saw somewhere that had a guy with peas in a blender saying ..Oh... I thought they said whirled peas [03:26]
SvenDowideithehe
nice!
wikipedia says they were a surf music band
who wrote the docco!
ind all topics that have form ColourForm where the form field 'Shades' is 'green' or 'yellow' but not 'brown'
%SEARCH{"(lc(Shades)='green' OR lc(Shades)='yellow') AND NOT(lc(Shades) ~ 'brown')" type="query"}%
pharvey, is there a proposal to add .web and .topic to the TOM?
[03:26]
pharveySvenDowideit, don't think so [03:28]
SvenDowideitie, form.name.web and form.name.topic
cos it occurs to me taht we are as stupid as a wrok
[03:28]
pharveypharvey nods [03:29]
SvenDowideit"form.name ~ '*.UserForm'" is silly
"form.name.topic = 'UserForm'" at least is precise-er-er
mmm, really can put that into the F::Address proposal >:}
that way the RM won't mind :p
[03:29]
gac410I'm still counting on Lavr coming back to RM 1.2
:P
[03:32]
SvenDowideitha
its my turn really its been a very long time since i did it
but i have trouble feeling confident that i can :/
[03:32]
.......... (idle for 46mn)
gac410SvenDowideit bleh - ToPDFPlugin I think is dead. Developer withdrew it. http://foswiki.org/Extensions/Archived/ToPDFPlugin [04:18]
SvenDowideitso we can close all tasks and support issues related? [04:19]
gac410I assume so.
EugenMayer was the one to say use something else. I moved it to Archived so people wouldn't fight with it.
[04:20]
SvenDowideitoh ffs, i can never remember what the STARSECTION syntax is, why the **** do we have multiples that are not synonyms for each other :(
how can a search work fine in one topic and not the next :(
oh argh :)
need mroe sleep
[04:21]
giggle.
how do you escape a $n that is in a form field, so that FORMFIELD doesn't parse it
thats really almost funny
[04:35]
gac410$dollarn
just guessing.
[04:36]
SvenDowideitthat'll be too confusing [04:37]
gac410I ran into something like that the other day where it was getting parsed. [04:37]
SvenDowideitwow. so I'm reporting a bug: [04:37]
gac410Ah ... It was a task about $dollar and the dollar went away. [04:37]
SvenDowideit$n newline format token doesn't seem to be expanded consistently in SEARCH formats
but there's no way for me to write that without the task's view tempalte tripping
unless i write it so no-one can read the summary field text
foswiki is weird.
[04:37]
gac410Item11523. $dollarcomma in the search summary, $comma in the task. [04:39]
FoswikiBothttp://foswiki.org/Tasks/Item11523 [ Item11523: $comma token not used in SUBSTITUTE and other Spreadsheet macros ] [04:39]
SvenDowideityup, ended up doing the same :/ [04:39]
gac410Search summary shows $dollarcomma token not used...
In these days do companies use totally private internal email domains. SomeUser@company (ie no valid tld?)
[04:39]
SvenDowideitmmm, i'm not sure how many email servers still would support that
but then, ew!
[04:42]
gac410The email regex for auto-link matching is specific for TLDs and its not end of the world if a new tld doesn't auto-link. But I just noticed that registration validates against that regex. So the new .porn tld for example probably couldn't register. [04:43]
SvenDowideitmmm, a tad naive innit
mmm, i need to go eat lunch - almost 3pm :)
[04:43]
gac410have a good lunch [04:44]
.... (idle for 15mn)
pharveySvenDowideit +1 on the content of SimplifyDataFormDevelopment proposal, but, damnit, why do we keep having proposals that state a goal instead of what they're actually doing (like ImprovePageLoadPerformance = ZonePlugin ?) [04:59]
SvenDowideitin my case, because stuffed if i know what it can & can't do
until i implement it
its a feature proposal, not a post development document
if you want big design up front waterfall, then 1. you get to pay for it, and 2. you get lectures from me about how that was a big mistake
[05:00]
pharveyYou're making a DataForm creator? Can't it just be called that? Anyway, I don't want to divert any energy into a pointless argument about naming (or is it too late) [05:01]
SvenDowideitname? you can name it what you like :) [05:01]
gac410WOW... a pointless argument that I'm not involved in :D [05:02]
pharveylol [05:02]
SvenDowideitoh, your complaining that i don't think up topic names the same as you do?
sorry, i refuse to have the same brain as everyone else!
[05:02]
pharveyNo, I hate proposals that are named in a totally opaque way [05:02]
SvenDowideitthen rename it
i personally hate naming shit
[05:02]
pharveyI feel stupid for even bringing it up, I'm just crazy. [05:02]
gac410I have to admit, digging through the proposals, sometimes the name and actual proposal are rather difficult to connect. [05:03]
SvenDowideitmy ideal would be to spout shit here on irc, and if its a good idea, someone else creates a feat req for it
and i'm willing to add text, but naming is like choosing colours, not something i want to take time over
cos my names always get argued over - see FORMAT
[05:03]
pharveypharvey would've named it something like AddDataFormCreatorWizard or something like that [05:04]
SvenDowideitdoit! [05:04]
pharveyta :) [05:04]
gac410The value in the name is really high for me ... when looking at a list of 100's of proposals, that's the *first* opportunity to entice me to look at it or move on. [05:04]
SvenDowideiti really really don't mind
for me the max value is in developing the idea
spending time on the topic name is a dissincentive and distraction from my core need
and i take it as given that most people will want a different name
ah, kids awake, feels like a good 2 hours producing zero code ;)
pharvey, gac410 and anyone else, please, encourage people to rename things
especially anything I named
[05:04]
pharveySvenDowideit: it's now named http://foswiki.org/Development/AddDataFormCreatorWizard [05:06]
SvenDowideitmakes me think i should use codenames
(i'll ignore that it isn't a wizard, because foswiki is a tad (not much) too dumb to do wizards
and because i friggen want wizard functionality alot :)
SvenDowideit goes - see you er, dunno, up too late making yum install foswiki-fcgienginecontrib just magically work
[05:06]
......... (idle for 41mn)
flexibeastgac410: Re. http://foswiki.org/Tasks/Item11449, since it involves intended behaviour, is that Item able to closed? [05:48]
gac410I don't think we've resolved whether or not we want to keep that restriction. So I'll probably leave it pending for now. [05:49]
flexibeastOkay, fair enough. [05:51]
gac410If we don't address it for 1.2, I suspect it will get another re-hash when we tackle UTF8 core for 2.0. [05:53]
flexibeast*nod* [05:54]
...... (idle for 26mn)
***gac410 has left [06:20]
....................... (idle for 1h50mn)
ModAcOstSo, is it possible to change my WikiName on foswiki.org or do I have to reregister? [08:10]
PrsndIs it possible to hide Topics and Webs from different UserGroups depending on the workflow status? [08:12]
.... (idle for 16mn)
***Prsnd has left [08:28]
.... (idle for 18mn)
CDotModAcOst: you would have to re-register (well, actually it would be possible, for an admin, but a right PITA)
Prsnd: do you mean the workflow status of the topic, or the workflow status of some other topic? (It doesn't actually matter, because the answer is still 'no')
[08:46]
.......... (idle for 47mn)
nick_Hey there, Hi there, Ho there
SvenDowideit: are you awake?
I just got the new ubuntu package update.
It has solved some problems it seems.
For example I no longer have the bug where EditChapterPlugin screws up the table of contents.
However, I still have a problem with EditChapterPlugin.
When I click on the little edit icon I get an error box pop up.
The title of the error box says "The page at localho..." and the text just says "error"
[09:34]
.............. (idle for 1h6mn)
***ChanServ sets mode: +o MichaelDaum [10:43]
........... (idle for 50mn)
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/0Cu6tA
[foswiki/master] Item11558: remove bogus reference to %BATCHUPLOADCHECKBOX% - FlorianSchlichting
[11:33]
***GithubBot has left [11:33]
FoswikiBothttp://foswiki.org/Tasks/Item11558 [ Item11558: BatchUploadPlugin: remove dysfunctional references to private extension ] [11:33]
pharveyCDot: http://foswiki.org/Development/LoadDifferentTopicVersions - WTF [11:42]
CDothow can it be coded? It was "Under Investigation", and had an unaddressed concern against it.... [11:43]
pharveyfor the love of GOD [11:44]
CDotCDot is poking things to see if they squeal [11:44]
pharveyCDot: whose concern was unaddressed? [11:45]
CDotmine
the clock cannot start until the concerns field is clear, and it has never been
CDot is happy to raise the concern, now that he has highlighted the process issue
[11:45]
pharveythe concerns field *was* clear? [11:46]
CDotgeorge and I are having one f*** of a time working out what state proposals are in [11:46]
pharveyyour last feedback was in May 2011 concerning a suggestion on speaking parameters that was never pursued [11:47]
CDotthe concerns field was *not* clear when you made the statement "starting the clock ticking" [11:47]
pharveypharvey still digging through to find what the concern *is* [11:47]
CDothttp://foswiki.org/Development/LoadDifferentTopicVersions?rev=26 [11:47]
pharveyCDot: please put me out of my misery, and let me know what part you have a concern with [11:48]
CDotmy concern was raised 3 May 2011 and was never explicitly addressed, AFAICT, though I'm sure you'll tell me now I can raise it.
My main issue with "speaking" parameters is that they duplicate an existing mechanism in most people's minds. For example,http://wiki.co.za/Web/Topic(rev=123)versushttp://wiki.co.za/Web/Topic?rev=123When do I, a simple, inexperienced, user, use the first, and when do I use the second? And what doeshttp://wiki.co.za/Web/Topic(rev=321)?rev=123mean?(Of course this applies to other syntaxes such as @123 as well, but the idea of extending to support bu
i.e. it related to the proposed bracket syntax for referring to versions.
[11:49]
pharveythe current code and the proposal as written does not use the bracketed notation you write about there
I was broaching alternatives, but that syntax was never implemented or proposed
[11:50]
CDotthe proposal isn't explicit as to what syntax *is* proposed, which makes it hard to accept [11:51]
pharveyOkay [11:52]
CDotmy remark about the %QUERY stuff related to your remarks of 2 May 2011 "Initially when we developed the versions[] syntax in QueryAcrossTopicRevisions, I was trying to keep this stuff ("address notation") consistent with the query syntax, but conflating the two problems didn't seem to be helping."
but that was left hanging; I'm not sure where it went from there
.... which is why it pinged my radar when i re-read it this morning
[11:53]
pharveyI am just a little surprised because I *thought* you had read the extensive POD I wrote (or at least, skimmed it) and knew what I was up to [11:54]
CDotno, I haven't read it in quite a while. I am doing so now; is the POD can be pasted into the proposal topic, so much the better
CDot often pastes already-written doc into proposals for clarification (e.g. for mtempest a couple of days ago)
[11:55]
pharveythe Foswiki::Address POD is miles long, I just worry about spewing more details into an already hard-to-follow topic [11:57]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Address [11:57]
CDotok, from the examples I can reverse-imply that @5 is the selected syntax for referring to a revision of a topic in a URL - right? [12:00]
pharveyyes
and I feel like an idiot, or a failure, or both, for having failed to communicate this properly, despite miles of conversation and foswiki-discuss
and in the end I don't care if you throw it all away and prefer Web.Topic[version], I just want something consistent
CDot, so obviously there is a concern that the proposal is unclear (does FoswikI::Address equal LoadDifferentTopicVersions? even I don't know, but is definitely at least a part of it)
but I need to know if you intend to abandon the @n syntax
my thought was that both would be supported
[12:00]
CDotright. There is no adequate description of a Foswiki URL which you could have extended [12:04]
pharvey'Web.Topic@2'/FooField [12:04]
CDotto me that indicates a lack of documentation
BTW I have an ongoing issue with 'AnotherTopic@4'/versions[5]
[12:04]
pharveyI enumerate all the places in which we need to support the address notation
and then I give examples, including an exmaple URL
[12:05]
CDotATM 'AnotherTopic@4' is *not* supported, but to reconcile the notation it maybe ought to be [12:05]
pharveyYes, I began Foswiki::Address as the backbone to it all
I began coding work, but got bogged down trying to unify all the link handling re-work required, http://foswiki.org/Development/ImplementingLinkProposals
[12:06]
CDotso how do I reconcile 'AnotherTopic@4'/versions[5]? I don;t see how I can. On the other hand, I don;t see why we need to support @4 in that context [12:07]
pharveyThis is where I don't know if I'm normal or crazy, but users paste topic names, web.topic names, into formfields which I follow in nested searches and queries dereference them
when a user wants to link to a specific rev, because that's the rev they've vetted, or identified as being important in some other way, the syntax they use should be consistent
or at least, shouldn't be gratuitously inconsistent
and I was wondering things like, perhaps could support QuerySearch expressions in %INCLUDE one day
[12:08]
CDotI don't have a problem with that; but 'AnotherWeb.AnotherTopic' is the address of a topic *with its complete history* whereas 'AnotherWeb.AnotherTopic@4' is the address of a single revision of a topic - which presumably has no versions array. if it has a versions array, then does that array reflect all known versions of the topic (making the @4 redundant) or does it only reflect the versions *before* [4]? [12:10]
pharvey%INCLUDE{"Topic@5"}% might just be a fun way to do rev="5" [12:10]
MichaelDaumbenchmarks available on the new page cache: http://foswiki.org/Tasks/Item11549 [12:12]
CDotCDot prefers to avoid the problems posed by @4 by ignoring it in the query language
that's not to say it isn't useful elsewhere - such as %INCLUDE - but I just can't make it fit in Foswiki::Query :-(
[12:12]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Query [12:13]
CDotit's kinda like the XQuery/XPath problem [12:13]
pharveyCDot: I see in the comments you touched on this indirectly, but I didn't identify this exact problem directly. That's a *great* concern. Please add 'Foo@2'/versions[4] example. [12:13]
CDotin fact, it *is* the XQuery/XPath problem
add it where?
[12:14]
pharveyJust as a comment, is this the crux of your concern (apart from the clarity issue)
until now in the discussion I feel a bit exasperated as to what exactly your concern was
[12:14]
CDotclarity is the biggest issue [12:15]
pharveyI feel I've documented, compared, and invited enormous amounts of feedback, compared to other proposals
ok
[12:15]
CDotthis is a critically important proposal, and needs to be very, very carefully thought out [12:16]
pharveyyes, hence my efforts getting it right [12:16]
CDotbecause otherwise we end up with a legacy [12:16]
pharveyyes, that has been my concern, loudly, all along [12:16]
CDote.g. the .versions - if we don't want to do it that way, I need to know *now* [12:16]
pharveyand I disagree that the syntax proposal is unclear, but I'll look at it tomorrow [12:16]
CDotcos otherwise I will implement and it'll go in 1.2 -> legacy [12:16]
pharveyI guess my problem is that I *really* want to be able to do queries like ApprovedVersion/info.author [12:17]
CDotthe query language has to be able to express ranges, sets of revisions
and the @ syntax is not powerful enough for that
[12:17]
pharveySomehow I thought you'd just decided we could do both [12:18]
CDota versions query has to - for example - "tell me all revisions *before the current one* that have the word 'whoops' in them" [12:18]
pharveyyes, I appreciate that versions queries demand a more notation which supports more expression [12:19]
CDotso do you agree that we reconcile by not supporting the @ syntax in topic names in queries? [12:19]
pharveyI could be persuaded to agree
I thought the benefit of OP_ref being able to follow 'Web.Topic@2' is compelling and cheap to support
[12:21]
CDotfine, but what does it end up at?
AFAICT it ends up at 'Web.Topic' - making the @2 redundant
*or* it imports the entire revision histiry. Hmm, that *is* a possibility
[12:22]
pharveybut it should follow version two of Web.Topic
'Web.Topic@2'/info.author - the author of rev 2
[12:23]
CDotthere's nothing to say the =versions= array needs to be locked to the version being viewd
in which case 'AnotherTopic@4'/versions[5] actually makes sense again
[12:23]
pharveyok [12:24]
CDotbit of a bugger to explain, but it *is* logical [12:24]
pharveywe could say that the versions array is always the array of all versions [12:25]
CDoty [12:25]
pharveyand still allow my pet OP_ref usage? [12:25]
CDotyes
CDot has no intention of implementing 'Web.Topic@2' for 1.2, but it is a logical step from where I'm going
[12:25]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/U0GMag
[foswiki/master] Item11549: minor fixes - MichaelDaum
[12:33]
***GithubBot has left [12:33]
FoswikiBothttp://foswiki.org/Tasks/Item11549 [ Item11549: rewrite of page cache ] [12:33]
pharveyCDot: thank you, I feel much better now
I just bashed out a use-case for the last few minutes, but my WiFi had died
pharvey has low sugar, g'night
[12:34]
CDotI *think* we are all consistent now. All I have to do is implement the version array :-(
sleep well
[12:35]
SvenDowideithey CDot so you want me to add a concern toy you versions feature? [12:35]
CDotSvenDowideit: yeah, probly. I'm not 100% happy with it yet [12:35]
SvenDowideitcos i still think its a bad idea to do it the way you are [12:35]
CDot'still'? [12:36]
SvenDowideitI _don't_ agree that Main.WebHome is a reference to an array of all versions [12:36]
CDotit isn't; and never has been [12:36]
SvenDowideitSvenDowideit goes to quote you above in your discussion with ph [12:36]
CDotMain.WebHome is a reference to *the latest* rev of that topic
'Main.WebHome'/versions is all versions of the topic, though
so if I write 'Main.WebHome'/info.author I get the latest author
[12:37]
SvenDowideit<CDot> I don't have a problem with that; but 'AnotherWeb.AnotherTopic' is the address of a topic *with its complete history* whereas [12:37]
CDotwith its complete history *in the versions array* [12:38]
SvenDowideitbasically
that is the root of the porblem
[12:38]
CDotyes [12:38]
SvenDowideityou're needing to qualify technically what you're refering to [12:38]
CDotthe problem pharvey and I just resolved [12:38]
SvenDowideitie, the syntax is vague and unhelplful
'Main.WebHome'/versions should be _less_ than 'Main.WebHome'
[12:38]
CDotit is. 'Main.WebHome' is a string. [12:39]
SvenDowideitand suddenly, you're adding an accessor to a string address that promotes it to be more [12:39]
CDot'Main.WebHome'/ is a *ref* to a topic [12:39]
SvenDowideiti'm trying to explain my concern [12:40]
CDotthat's what the ref operator is (has always been)
CDot shuts up and listens
[12:40]
SvenDowideiti know what you are _trying_ to say, and i think that makes the syntax harder to grok
'Main.WebHome'/ is a ref to the latest revision of a topic
[12:40]
CDotyes [12:40]
SvenDowideitup to now [12:40]
CDotstill is [12:40]
SvenDowideitadding 'Main.WebHome'/version
confuses that
[12:41]
CDotyou mean versions? [12:41]
SvenDowideityes, ok [12:41]
PeterSvendsenHi all, I have made a search to look for all *.ppt in the web.. I would like to present the result by links to the file... so far I have only been able to link to the topic where the file are attached... how is the correct way to do that? %SEARCH{ "attachments[name~'*.ppt']" type="query" nonoise="on" header="| *Topic* | " format="| [[$topic]] |" }% [12:41]
SvenDowideitthats even worse
if adding a plural makes a scalar into a massive array of complex objects
then we're really toasting users
[12:41]
CDotit isn't a scalar. It's a massive array of complex objects.
it never was a scalar.
[12:42]
SvenDowideiter, no
'Main.WebHome'/version
is a scalar
a simple number
adding versions _there_ means a user mistypes
and they get... this massive array of _not what they wanted_
[12:42]
CDotoh, you are thinking of the shortcut to info.versions.... I had forgotten about that
info.version
[12:42]
SvenDowideity [12:43]
CDotf***ing shortcuts [12:43]
SvenDowideitnot really
what i'm trying to say
[12:43]
CDotok, so would 'history' be happier? [12:43]
SvenDowideitno [12:43]
CDotso propose something [12:43]
SvenDowideitthe root issue is that it should not be coming from the ref to a single topic
my concern is that the existing syntax is bad
i don't need to have a good answer replacement to ahve that concern :(
but.
i have a feeling
[12:43]
CDotCDot isn't going to change the existing syntax, but is open to rational proposals for extensions to that syntax. [12:44]
SvenDowideitthat putting the history operation off ref is a bad thing, and that it might be better on the left side
so i should just raise a concern and block it as a bad idea?
[12:45]
CDotno, I don't want you to block it, I want you to propose a sensible alternative
because I have an existing implementation that is currently in flux, and now is a good time
[12:45]
SvenDowideitok, so i have to block it until __we__ work out a sensible alternative
my basic thought process is this
$meta
is a concrete reference to a particular version
we don't de-reference it and ask _it_ for all revisions
[12:46]
CDotyes. And that version has a history.
every version has a history.
the versions of a topic are part of the meta-data of that topic
[12:47]
SvenDowideityes, but you don't add the history to the object as an attribute [12:47]
CDotthat's what I was trying to express [12:47]
SvenDowideity, and that is wrong [12:47]
CDotwhy not? we add other meta-data as attributes... [12:48]
SvenDowideitat _best_
oh, here's a thing
'Main.WebHome'/history[1] == what?
[12:48]
CDotwe even add the history of *other* objects (FILEATTACHMENT
you might as well ask fields == what?
or META:PREFERENCE == what?
[12:48]
SvenDowideitok, i give up [12:49]
CDotthe concept of an array exists [12:49]
SvenDowideitif you can't help me work though this
i can't do anything
[12:49]
CDotI'm trying to get you to understand where I'm coming from [12:49]
SvenDowideitother than to say, the proposal is not yet good enough
i know where you are coming from
[12:49]
CDotcos at the moment, all I'm seeing is a vague disquiet
with no concrete alternative
[12:50]
SvenDowideitno, the syntax is essentially backwards
and confuses what is otherwise a simplifying language
each step towards the right reduces the breath of what the user is asking for
[12:50]
CDottrue, and that has bothered me [12:51]
SvenDowideitits a core problem for helping the users use it [12:51]
CDotOP_ref already reverse that, however [12:51]
SvenDowideitand i suspect it also screws up the impl [12:51]
CDotstring/ -> $meta [12:51]
SvenDowideityes
and if i'd have been awake when you did that, i'd have made you think harder too
but thats too late
so there's no reason to make things even more complicated
[12:51]
pharveyI just quickly came back online (sugar is good again :) to say, it might be worth reviewing a solid list of example versions queries, to ensure the syntax looks ok to express them.. [12:52]
CDotok, let's say you had the chance to travel in time; what would your alternative to / be? [12:52]
SvenDowideitpharvey imo thats less an issue
as CDot can reasonably easily make the examples and explain them
the problem has more todo with normal users using it
[12:52]
CDotyes, and I'm listening; but I need some flesh on these bones. [12:53]
SvenDowideitCDot i'm not solutioning yet, i'm trying to qualify the problem sufficiently so that i / we have an idea of the boundries [12:53]
CDotI understand your issue, and *in principle* agree with you that "hoisting" the abstraction is a big step, and the current syntax isn't good at it
CDot would *love* an opportunity/excuse to deprecate /, which je has always regretted
[12:54]
SvenDowideitkey to my thinking is that left to right order of complexity reduction is a useful touchstone [12:55]
pharveyI remember we talked about making QuerySearch do things outside just topics, like logs[foo] and that type="query" web="Foo" might also be expressed as topics[web='Foo'] [12:55]
SvenDowideitimo deprecation / wil just hurt more
hehe
[12:55]
CDotCDot isn't serious, just wants to expand the box [12:55]
SvenDowideitquery already does logs
see the ContributorsPlugin
[12:55]
pharveyCDot: you don't like OP_ref? But it's so awesome! :) [12:56]
CDotso, l2r order means to raise an abstraction I have to write things functionally? [12:56]
pharveywe have normalized the bejeesus out of some of our data, without OP_Ref I wouldn't have bothered.. [12:56]
CDotref('AnotherTopic') [12:56]
SvenDowideithell, I actually hate 'Main.WebHome@1..5'/ less than i worry about 'Main.WebHome'/history[1..5]
CDot that would be the first place to look i guess
s/ref/get/ perhaps
[12:56]
CDotso you are simply asking ref - which is an postfix op - to become prefix? [12:57]
SvenDowideitthinking like a user - the word ref is pretty weird [12:57]
CDotyeah, yeah, worry about that later [12:57]
pharveymy colleagues think of it as "kinda like a join"... :} [12:58]
SvenDowideitevery time i worry later, it becomes too late [12:58]
CDotso, say I added a prefix operator "smoogle" that does what postix / does [12:58]
SvenDowideitpharvey omg [12:58]
CDotwould you be happy then? [12:58]
pharveySvenDowideit: only in the sense that it's slow :P [12:58]
CDotand anoth prefix operator "allothem" [12:58]
SvenDowideitget('Main.WebHome').author
er
and thus i kick at query('Main.WebHome').date
[12:58]
CDotso the current expr 'AnotherTopic'/versions[5] becomes allothem(smoogle('AnotherTopic'))[5] ? [12:59]
SvenDowideitno
we do not want to make a horrid thing like that
[12:59]
CDotgood :-) [12:59]
SvenDowideitif you're smoogleing
then use 'Anothertopic@5'
the faster you reduce the working set
the better
that was one really bad thing about versions
[13:00]
CDotwhat is I want revs 1,3 and 5? [13:00]
SvenDowideitthen use a simple F::Address
'Anothertopic@(1,3,5)
[13:00]
CDotmy test example is "give me the dates of all versions that were written by me, not including the most recent"
now you are extending the F::Address syntax quite far.....
[13:01]
SvenDowideityes, am
thats a good eg
[13:02]
CDotCDot can't see how the F::Address can deliver selected versions based on a query [13:02]
SvenDowideitit shouldn't
we were not talking about that far until you just mentioned it :)
[13:02]
CDotso i still need a way to refer to the complete history of a topic, don't I? [13:03]
SvenDowideitthe basic premis i have, is that you need to avoid talking about allversions
until its desperatly necessary
the i don't need all versions syntax should therefor _not_ the the same as the i need all versions syntax
whereas your versions impl previously always blew out memory
even if i was asking for v1
or for v1,3,5
[13:03]
CDotwell, the compromise the pharvey and I worked out delivers that, cos 'ABC@3'/ is more terse than 'ABC'/versions[3] [13:04]
SvenDowideitits not just more terse, its much faster [13:04]
CDotbut you still need some way to use a query to refine down a set of revisions, don't you? [13:04]
SvenDowideitunless you start re-writing the query node
which i did start, but was problematic
i agree, there's a need
[13:04]
pharveypharvey is wary of making @ more magical; wanted to keep address-notation & querying kinda separate-ish... [13:05]
SvenDowideitbut imo it should be a separable syntax to normal single item access
pharvey deffinitaly
[13:05]
pharveyoh damn, I borked http://trunk.foswiki.org/Development/QueryAcrossTopicRevisions [13:06]
CDot@ should deliver a single rev, IMHO [13:06]
SvenDowideity, WebHome@(2,3,4) is needless
WebHome@2, WebHome@3, WebHome@4 means the same thing
given what i know of the impl of ref and versions
i'm wondering if it makes more sense to make something like
[13:06]
CDotdoes @ implement @-1, @-2 etc?
as in @0 refers to the most recent, @-1 to the next most recent etc?
[13:08]
SvenDowideitSMOOLE[name='OtherTopic' AND author='Crawford' AND rev!='head'] [13:09]
pharveyCDot: Not at the moment. I thought that was too query-ish.
er, I mean - in Foswiki::Address, I mean.
[13:09]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Address [13:09]
CDotwhat's the other op to SMOOLE? [13:09]
SvenDowideit? [13:10]
CDot^op^operand
it needs another operand, dunnit?
[13:10]
SvenDowideitwhy?
this 'SMOOGLE' is essentially a reference to the store
and returns the resultset
[13:10]
CDotoic - you mean SMOOLE refers to the entire history of every topic in the wiki? [13:11]
SvenDowideitya
and name='WebHome' ... tells you the topic
[13:11]
CDotok, so write my example query, please [13:11]
SvenDowideitthe point being to make the full versions thing sepatate
that is it i think
SMOOLE[author='Crawford' AND rev!='head']
[13:11]
CDotCDot wanted the date [13:12]
SvenDowideitoh, that thing returns an array of either addresses or refs, dunno which [13:12]
pharveyDamn, http://trunk.foswiki.org/Development/QueryAcrossTopicRevisions#Examples aren't working [13:12]
SvenDowideitSMOOLE[author='Crawford' AND rev!='head']/date ? [13:12]
pharvey/info.date [13:12]
SvenDowideitor
shortcuts :)
SMOOLE[author='Crawford' AND rev!='head'].date ?
depending on impl detail
likely the later, but stuffed if i know
SMOOLE[author='Crawford' AND rev!='head'].info.date ?
SMOOLE[web='Tasks' AND author='Crawford' AND rev!='head'].info.date ?
[13:13]
CDotok, so you are making a special case of a query over all topics in th store [13:13]
SvenDowideitSMOOLE[web='Tasks' AND info.author='Crawford' AND rev!='head'].info.date ?
no
i'm making the query the head rev only a 'special case'
sorry, yes :)
[13:14]
pharveySMOOLE[author='Spammer']/address would be handy [13:14]
CDotCDot favours shortening SMOOLE to SMOO, as being more intention-revealing [13:14]
pharveylol [13:15]
SvenDowideitSMEGOL? [13:15]
pharveypharvey was using topics[...] initially, but okay :) [13:15]
SvenDowideitthat wya, i can trivially SEARCH for all OMGYOURKILLINGMYWIKI queryies [13:15]
CDoty. I had considered extending up to full-wiki queries, but avoided it due to the syntax complexities [13:16]
SvenDowideit?
i didn't realise you'd built a neutered versions[] thing
basically, i tried to fix versions last year when doing mongodb hoising for it
and its deref to topic then deref to all versions concept meant i had to invert everything the language did to make it perform
[13:16]
CDotCDot isn't wedded to versions[] - it just felt like the easiest way to do it. But I'm open to alternatives (always am) [13:18]
SvenDowideitso its not like versions is good for the backend either [13:18]
pharveySvenDowideit: before I jump off, I don't like the topic(form.name) form, prefer form.name.topic instead, but I wonder if this all means typing our schemas and attracting lepers [13:18]
SvenDowideitpharvey the reason
is that you might want to get the topic portion of _anything_
so a function that operates on strings will work
whereas an attribute on a topic object is just plain painful
[13:18]
pharveysounds annoying to hoist, but I guess I've never written a hoister, just debugged one :P
ok
[13:19]
SvenDowideitas you're forcing people to dereference all strings (ie create a meta obj in the backend) and then call topic on it
let alone that accessors are special
[13:19]
pharveyI guess I can find the energy to move my fingers up to the bracket () keys [13:20]
SvenDowideitSvenDowideit send you some sugar [13:20]
pharveyhehe [13:20]
CDotCDot was sucking the nectar from primrose flowers yesterday; delicious!
I can't help feeling that 'topics' is the wrong term
[13:21]
SvenDowideitSvenDowideit ponders SEARCH{"SMOO[author~'Sven*']" topic="Item1*" web="Tasks" [13:22]
FoswikiBothttp://foswiki.org/Tasks/Item1 [ Item1: set up mailercontrib crons ] [13:22]
pharveyon that note, g'night, and thanks again CDot for humouring my brief outbursts, I've yet to work on any other project that can work through arguments as successfully as we do in Foswiki :) [13:22]
SvenDowideitCDot topics? [13:22]
CDotlooking for a more friendly smoo [13:22]
SvenDowideityou mean topic('WebHome'/name)
for what?
[13:22]
CDotfor SMOO [13:22]
SvenDowideitoh, no, its got to be in your face obvious
######[] is good
[13:23]
CDotgood god, no it isn't! [13:23]
SvenDowideitits better than a string that looks similar to a normal thing in the query language atm
ie ##### is better than topics[]
[13:23]
CDotsure, but it's relative [13:24]
SvenDowideitits much better, given that we might have topic() [13:24]
CDot2 versus 1, whereas I want 10 [13:24]
SvenDowideitthis is why i originally suggested implementing it as an extra param [13:24]
CDotCDot wishes keyboards had an infinity symbol on them [13:24]
SvenDowideitie
SEARCH{"author~'Sven*'" topic="Item1*" web="Tasks" allversions="on"
though i recon that'd be scary ++ too
er, i meant to write
SEARCH{"author~'Sven*'" topic="Item1*" web="Tasks" noversion="off"
[13:24]
CDotCDot throws a rock at SvenDowideit's head [13:26]
SvenDowideitthat would be nice
SvenDowideit probly won't sleep now, and i didn't sleep much last night
been trying to get too much done :(
[13:26]
CDotrecent research suggests microsleeping is much healthier
or rather, mini-sleeping
CDot looks for the article
[13:27]
SvenDowideitslashdot mentioned it too
seen it, and whatever
[13:27]
FoswikiBotSorry, I haven't seen it, and whatever. [13:27]
CDotok, so the upshot of this discussion is I'm removing the versions stuff from 1.2 [13:27]
SvenDowideitbecause we can't find a syntax in on hour? [13:28]
CDoty [13:28]
SvenDowideitor because you need to totally re-write it all [13:28]
CDotwell, maybe that too [13:28]
SvenDowideitpersonally i'm a little shocked at how fast this '1.2' proposal is being kicked [13:29]
CDotkicked? [13:29]
SvenDowideiti hope it finds us an RM for it :) [13:29]
CDotgac410 agreed to RM it [13:29]
SvenDowideitum, eh? [13:29]
CDotI asked him a few days ago
if he would be willing
[13:29]
SvenDowideitwhen i talk to him, he says he's hoping that someone else will [13:30]
CDotCDot would not be as active if there was no RM there [13:30]
SvenDowideitbut thats what i'm having trouble with
there's lots of quick talk, and not much discussion for the non-irc-ers
though i admit its much like irc has everyone, so the dynamics not rubbish
there's just an aspect of chinese whispers going on between tzs
[13:30]
CDottrue. which is why it's important to focus down the set of "sensible" topics ofr the release, and then making sure they are documented [13:31]
SvenDowideitand george is left as the piggy in the middle
no
its important not to alienate people by driving forward one view of 'sensible' before they even know whats going on
atm, i'm hoping that stuff isn't just yoinked away from us before we even know that there is something happeninf
[13:31]
CDothave there been any new proposals recently? AFAUK all the proposals have been there for moons [13:32]
SvenDowideitbasically, i asked george to write up something on f.o/Blog
1.2.0 __is__ a proposal
[13:32]
CDotthere are quite a few new *commitments* but the prooposals are old
y, it wouldn't hurt to blog it, agreed
[13:33]
SvenDowideitbut you seem to (imo in a good way) be driving something more than we've told people about
f.o/Blogs is the microblog for the project
[13:33]
CDotand fos-discus it to [13:33]
SvenDowideitadd a para there and get people talking
thats why i started to do the weekly un-newsletters
we want to bring more people into the discussion, not do things so fast that they cna't orient themselves
which is the same thing as declaring that versions is not in 1.2 because you and I, 2 overly influenced by the crap language legacy can't think up a better syntax
so http://foswiki.org/Development/GrandUnifiedSearchAPITwoZero#foo_7
uses logs[] to query the logs
in the Move Scope into the query
mmm, so i suggested revisions[web='Sandbox' AND author='JoeBloggs']
clearly influenced by your touchstone query
where the scope can then be used to query a previously named result set
nifty
then we get
revisions[web='Sandbox' AND author='JoeBloggs']
is the same query as
revisions[web='Sandbox'][author='JoeBloggs']
as revisions[web='Sandbox'] returns a set that the next [] queries over
[13:34]
CDotCDot is happier with the second form, but still feels a vague disquiet
need some lunch, and then a sit-down-and-think
not saying I don't like it, just saying I need to think through the implications
[13:45]
SvenDowideitSvenDowideit might try to goto bed again
hell y
that syntax just happened to be there when i shifted to the other borwser
its a tad hairy to have revisions[] and 'revisions'/[]
[13:46]
CDotthey mean very different things [13:49]
SvenDowideityes
but look similar
[13:49]
CDotCDot is thinking of a syntax like a * prefix [13:50]
SvenDowideitand so in the hands of most, will be confused [13:50]
CDote.g. *revisions[qqqq]
but that's nerdy
[13:50]
SvenDowideitooo, so * would prefix a resultset [13:50]
CDotright - but it's nerdy [13:50]
SvenDowideitie *logs[]
it should be a tad nerdy
you have to have a slightly wider pov
but ya pointer dereference for resutlset
vs directory dereference for turning a string into a topic ref
[13:50]
CDoty. * means "the world" - I was thinking of *.revisions but thought you would rebel
and also the conflict with the multiply op
CDot isn' abig fan of mysterious nonary ops
[13:51]
SvenDowideit*.logs[country(host) = 'AU']
*logs[country(host) = 'AU']
can't seem _much_ improvement
[13:52]
CDotwhat other op would you *ever* follow * with? [13:53]
SvenDowideiti was thinking more that * denotes an alias
ie
[13:53]
CDot*[name=''] is a query over... what? [13:53]
SvenDowideit@logs
* by itself either doesnt exist
as or..
we're going in different directions
[13:53]
CDotno, just thinking around the problem
CDot isn't going in any specific direction yet
[13:54]
SvenDowideity, in opposite directions :)
@(logs,revisions)[text ~ '*Spam*']
darnit :(
worse
@(resultset1, resutlsset2)[text ~ '*Spam*']
[13:54]
CDotSELECT * FROM logs,revisions WHERE text ~ '*Spam*'; [13:55]
SvenDowideitnono
logs&revs are different DBs
[13:56]
CDothmm [13:56]
SvenDowideitwith frustratingly different query indexes and engines [13:56]
CDotok [13:56]
SvenDowideitlogs is seriously our events.log
it works
but i didn't mush it into the query's scope
[13:56]
CDotCDot will worry about that after lunch. Back shortly. [13:57]
SvenDowideitSvenDowideit will hopefully be gone
back in a long time
[13:57]
mmm, so, if i ask for *revsions[author=PaulHarvey].name i get a non-unique list of topics
*revsions[author=PaulHarvey].name ~ 'Something*'
is the same as
*revsions[author=PaulHarvey AND name ~ 'Something*']
is the same as
*revsions[author=PaulHarvey][name ~ 'Something*']
gosh how that hurts
having all 3 forms in the query parse tree makes hoisting really complicated
and iirc there were some cases that i could not xform, and thus remained inefficient
course, in an IF to have something like
count(*revsions[author=PaulHarvey][name ~ 'Something*']) > 3
count(unique(*revsions[author=PaulHarvey][name ~ 'Something*'])) > 3
course for such simple eg's its easy
nesting ORs and ANDs is where it all came unstuck
SvenDowideit goes
[14:04]
....................... (idle for 1h54mn)
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/dRu-dA
[foswiki/master] Item10657:Item11505: - MichaelDaum
[16:03]
***GithubBot has left [16:03]
FoswikiBothttp://foswiki.org/Tasks/Item10657 [ Item10657: provide a solr multicore setup by default ] http://foswiki.org/Tasks/Item11505 [ Item11505: Javascript Error on JQuery 1.7.1 ] [16:03]
.... (idle for 17mn)
MichaelDaumCDot, is DISPLAY_IF new on trunk? or do we have it on 1.1.4 as well? [16:20]
CDottrunk only [16:20]
MichaelDaumwould adding it break configure on old foswikis? [16:20]
CDotno [16:20]
MichaelDaumk thanks [16:21]
jeskevery Topic has the 'Web Utilities' listed
no its only on the first Topic of the Web
i want to have the WebPreferences link and the others still available somehow
is there a nice way to hide those links maybe by having a hide/show button or so
just deleting them would mean for me not knowing how to get to those pages :)
some good ideas? :-)
or xamples?
[16:25]
GithubBot[foswiki] foswiki pushed 2 new commits to master: http://git.io/GNdv4w
[foswiki/master] Item11402: - MichaelDaum
[foswiki/master] Item11139: syntax error breaking installation - MichaelDaum
[16:30]
***GithubBot has left [16:30]
FoswikiBothttp://foswiki.org/Tasks/Item11402 [ Item11402: Problem saving topic when SolrPlugin is enabled ]
http://foswiki.org/Tasks/Item11139 [ Item11139: initial release ]
[16:30]
.... (idle for 15mn)
GithubBot[foswiki] foswiki pushed 3 new commits to master: http://git.io/3wwUaQ
[foswiki/master] Item11561: fixed installer breakage - MichaelDaum
[foswiki/master] Item11562: - MichaelDaum
[foswiki/master] Item11562: preparing release - MichaelDaum
[16:45]
***GithubBot has left [16:45]
FoswikiBothttp://foswiki.org/Tasks/Item11561 [ Item11561: syntax error in pm stub breaks configure and pseudo-install ]
http://foswiki.org/Tasks/Item11562 [ Item11562: protect exclamation marks in formfield editors ]
[16:45]
FlorianPHello; Is it possible to use $1 $2 in %INCLUDE when i use a pattern?
like pattern=".*?\-\-statTopViews\-\-\>[^\|]+\|[^\|]+\|[^\|]+\|[^\|]+\|[^\|]+\|[^\|]+\|[^\|]+\|([0-9 ])(.*))\|.*" output : $1 | $2
[16:59]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/WsOoqg
[foswiki/master] Item11433: check access rights before extracting the TopicTitle - MichaelDaum
[17:00]
***GithubBot has left [17:00]
FoswikiBothttp://foswiki.org/Tasks/Item11433 [ Item11433: use TopicTitle property as a link text only when permitted by view rights ] [17:00]
.... (idle for 15mn)
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/dwV-lw
[foswiki/master] Item11433: check access rights before extracting the TopicTitle - MichaelDaum
[17:15]
***GithubBot has left [17:15]
.... (idle for 15mn)
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/IWN1tA
[foswiki/master] Item11433: wochenende - MichaelDaum
[17:30]
***GithubBot has left [17:30]
FoswikiBothttp://foswiki.org/Tasks/Item11433 [ Item11433: use TopicTitle property as a link text only when permitted by view rights ] [17:30]
........... (idle for 54mn)
AlexanderStHi there, does anyone have experience with the PublishPlugin?
i would like to get a html file from a specific topic but with out all the logging stuff of the Publish Plugin
[18:24]
.................................. (idle for 2h46mn)
BabarSvenDowideit: http://foswiki.org/Blog/NewsLetter2012-02-24 ... your searches suck. It pulls tasks so old it's embarassing
eg: Item513
[21:11]
FoswikiBothttp://foswiki.org/Tasks/Item513 [ Item513 ] [21:11]
gac410And ToPDFPlugin was withdrawn by the author -
hm... shouldn't the WikiUsers topic be updated by RegistrationAgent?
[21:14]
ArthurClemenswell, it also means the tasks need pruning [21:17]
gac410gac410 notices off and on that it's updated by AdminUser [21:17]
....... (idle for 30mn)
SvenDowideitBabar, that was the idea [21:47]
Babarthe idea was to dig up mud from t-days? [21:47]
SvenDowideitthe idea is to remind peolpe of things they've forgotten
if its open, and old, then we'll get it
next up i'll randomise more
[21:48]
BabarBabar spent most of his day pruning the tasks at work [21:49]
SvenDowideit:) [21:49]
Babarwas around 120 when I started
96 now
[21:49]
SvenDowideitSvenDowideit runs for the hills again :) [21:51]
Babarspeaking of which... I need to figure out a way to enable git push to git.foswiki.org [21:54]
............ (idle for 55mn)
Item11506 [22:49]
FoswikiBothttp://foswiki.org/Tasks/Item11506 [ Item11506: Value of GTIME{"$week"} is not ISO8601 conform ] [22:49]
Babarok, done.
RHA!
Babar wants //
to implement pharvey's request
Foswiki::Time::formatTime( $_[1]->{time} // time(), $_[1]->{_DEFAULT} || '',
[22:57]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Time [23:01]
GithubBot[foswiki] foswiki pushed 2 new commits to Release01x01: http://git.io/w30nnQ
[foswiki/Release01x01] Item11506: TDD: Write tests for $we - OlivierRaginel
[foswiki/Release01x01] Item11506: add $we to have zero-padded week number - OlivierRaginel
[23:05]
***GithubBot has left [23:05]
GithubBot[foswiki] foswiki pushed 2 new commits to master: http://git.io/3PT7NQ
[foswiki/master] Item11506: TDD: Write tests for $we - OlivierRaginel
[foswiki/master] Item11506: add $we to have zero-padded week number - OlivierRaginel
[23:06]
***GithubBot has left [23:06]
pharveyBabar, I guess we can use that syntax when RHEL & Solaris are shipping perl 5.10 in... 2015? :)
Actually, RH already does 5.10, doesn't it
yep
Solaris 11 also ships 5.10
[23:11]
Babaroh crap. If I had this time feature to GMTIME and the like... I have to document it!
Babar was worrying about unit tests, but there are no unit tests about time functions :)
[23:16]
pharveyBabar, I can document it, if you like [23:30]
Babarthat's very nice of you
but first I'd like to be sure my implementation works
and for that, I'd need unit tests :)
so... I'll write a bunch
[23:31]
pharveyI'm a bit mad - I often write POD or doc first, then implement :P
well, just to clarify in my mind how I'd like some public thing to work
[23:33]
BabarI usually write the tests for that :)
but sure, write the doc, or at least the doc for GMTIME, and I'll implement
(and merge into the release branch)
[23:34]
crap...
I just ran the full unit test suite.. with perlbrew
so missing half the required modules :(
wondering where I should put tests for GMTIME, SERVERTIME, etc...
[23:42]

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)