Re: [Jbeta] J IDE
On 5/31/06, Eric Iverson [EMAIL PROTECTED] wrote: I think it is a good thing that we provide the J IDE in addtion to the old tried and true tools. Certainly:) However i too have been limiting my use of it. Primarily for one reason: When J crashes it takes the ijs windows with it:( If i use another editor then i can use its autobackup and, in any case, be independent of J's stability. i would like to figure out how to get TextPad's IDE capability to mesh with J's ... if anyone has explored that particular route. Or maybe someday i should beak down and learn eMacs. i have used TextPad for years so it is comfortable to me, but i am sure there are alternatives which are also easy to switch to and may be even more easily meshable with J. Perhaps if the J windows had a more robust autobackup mechanism of some kind that could be switched on... Along with an easier mechanism to access recently used directories in the file finders. i don't know what the best extension is, i just know it is pretty rickety to put ones thoughts on papers which are so hard to secure. Maybe when i figure out how to hack the system files, and stick my user directory in place of the TEMP defaults etc i can use the save key more freely/frequently: this mostly applies to the ijx window. But the main desired improvement, for me, would be to insulate J's IDE from the runtime environment. The ijs editor is quirky/different, but hey the source is open, so it can be hacked. i would like to use it, if it were secure. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
Yes, when J crashes you may lose unsaved code, this is why intead of Ctrl+W, I use Ctrl+S Ctrl+W all the time, esp. with things like DLL calls. To salvage ijx you can save it periodically too, it will then just increment file name/number in temp for the next ijx session. ctrl w reminds me to ctrl s any way so that is not a worry. TextPad used to be my favorite too. Though it has nice features like numbered replace, it's too slow on open. Intead, now I use EditPlus. Their IDE features is like invoking a compiler. But that's a one time thing, I don't think they can maintain a terminal session, can they? They do have an IDE partner they collaborate with, but i have not yet figured out if it is usefully in the J context. Also it's not the right direction: if you use J, you need to use J tools. If you see something lacking, it's better to improve the J tools. Would be nice. Problem is my need is for something which is neither ijs nor ijx. More like a programmer notebook. Remember TTY's? We used to cut out pages, paste in, highlight, and annotate snippets and put them in three ring binders. Maybe the Mathematica notebook would be a closer analogy. Probably if i were more disciplined an ijs could act like that, but it would kind of pollute the script files and i would have need for a way to strip it down on completion. The ijx is closer, but one still needs the caution of saving before every execution which could potentially trash the environment For example, there are issues with find dialog box and the script window. Use cases: Find new word. - Open find: Ctrl+F - Type word - Press enter to find Issue: default button is Find from Top, I believe it should be Find Next Found word. Now want to find next word. Issue: expect to press the same key (in many viewers it's F3 for initial find and all the Next, and Shift+F3 or F7 is find new) or an adjesent key, like Ctrl+G. Instead need to use mouse (horrible) or press Ctrl+F just to return to the dialog box. Now want to dismiss the dialog. Press Escape -- instead the script window closes -- this is a major setback. So need to press Ctrl+F Escape, or use mouse. Other issues: there is no tab ordering, i gotta haul out the mouse to restart the search etc. BTW, autoclosing scripts on Escape is an annoyance rather than convenience. Agreed, that is why for now i maintain a separate note booking environment. But it would also be possible to insulate the two environments so that there was no need to constantly be saving IJS's... at least i can guess it is. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
Eric On 5/31/06, Eric Iverson [EMAIL PROTECTED] wrote: If you have edited several scripts at the same time and a run of one of them is dangerous (might crash) then an S of each before the first W would be wise. The simple habit of always running after editing largely removes this concern. If i knew it would crash, i wouldn't do it;-) Also the ijs window lacks a line stack for trying variations so it is not really where one would like to have the inner loop of coding. i tend to treat it as where i put the final results, so i can resume the next day etc. Thus it lacks the latest fragments. As you point out it is easy to save the ijx file by a manual step. One could easily provide a function key shortcut that did this. We could, if pressed, provide an option that automatically did a save on every entry, but I would initially be against this. i would rather see an IDE with some isolation from the crash prone development environment. Some IDE users develop a strong habit that non-transient stuff is put in scripts and view the ijx window as a completely transient scratchpad. At least for some styles of work and development this works well and is what the IDE caters to. Options and extensions can be added fairly easily, but should be carefully considered. In particular where a change can be implemented as a private customization a user should live with it for a while themselves before proposing it for others. i do think the notebook concept is not addressed well by either ijs or ijx. And is something closer to the heart of actual development. And perhaps the IDE could be used for a while as intended. Who knows, maybe it works a bit better when used that way than one might think at first blush. Yeah, i'm still learning how it works, and still fluid, but i think the defaults don't do it, and it will require in the end either an external IDE component or some heavy investment in customizing the interface. The whole safety issue keeps me from the burden of the second option. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
Eric wrote: I think it is a good thing that we provide the J IDE in addtion to the old ried and true tools. greg wrote: Certainly:) However i too have been limiting my use of it. Primarily for one reason: When J crashes it takes the ijs windows with it:( If i use another editor then i can use its autobackup and, in any case, be independent of J's stability. On 5/31/06, Fraser Jackson [EMAIL PROTECTED] wrote: I do not understand this. The User manual clearly states that when running a script window it is saved as a file. If you are developing multiple script windows and running only one of them, a simple Ctrl+S of those windows will save them before trying others. However there is no need for the Ctrl+s Ctrl+W in Olegs contribution when running a script window. The session manager does it automatically. Undeniable. Except that it is not guaranteed that code is run from an ijs window: it can be run from the ijx window. Indeed there are big advantages to running it there: you have access to a statement stack allowing experimental variations much more quickly. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
On 5/31/06, Miller, Raul D [EMAIL PROTECTED] wrote: Note that when you save the file under a different name you do not have any good way of getting rid of the old copy. However, if you do not save under a different name, J will delete it if you are not careful. That moving between the temp and the user directory in the browse window involves several clicks and some visual searching makes this somewhat worse. The biggest pet peeve which has been prompting me into studying the startup code:-) i would like to see something like a recently used directory list for saves to allow one to file ijs ijx files under appropriate project directories. Possible remedies include: (1) Move the temp directory inside the user directory. (2) In Save As, if the existing file is a blank file in the temp directory, delete it. (Alternatively, remember the numbered file in the temp directory and ask if it should be deleted (either right away or later when the window is closed).) ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
On 5/31/06, Oleg Kobchenko [EMAIL PROTECTED] wrote: --- greg heil [EMAIL PROTECTED] wrote: i do think the notebook concept is not addressed well by either ijs or ijx. And is something closer to the heart of actual development. Is this what you are thinking, similar to the idea in session executed lines at http://www.jsoftware.com/jwiki/System/Interpreter/Requests ie http://www.jsoftware.com/jwiki/System/Interpreter/Requests#head-a120f2ebd5e9e9ef684d831edc0c45fe390dc6f6 Yes those are addressing a similar issue. Maybe a bit more of the Mathematica Notebook, hypertexty stuff too. But yes the execution environment needs more integration with a documentation arm. But most especially i need to document with facility and security. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
On 5/31/06, Chris Burke [EMAIL PROTECTED] wrote: I don't think it good practice to develop code in ijx, and use ijs only for storing the code. This might work for trivial code fragments but would quickly bog down for larger systems. Heh, maybe you should remove the statement stack. It is irresistibly enticing;) Generally i agree though. But programming is more than building systems, it is also an experimental science. One needs a notebook _and_ a report writer. i do mostly work from tacits, so scripts are not much of a focus. A lot of other machinery fails in this need as well. Eg i am not getting much mileage out of the debug environment. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
On 5/31/06, Chris Burke [EMAIL PROTECTED] wrote: How to quantify not getting much mileage? The debugger is not perfect, but most things work pretty well, and I use it all the time. What specific things does it lack? There seems to be no way of teasing a tacit apart. No indication of where the failure occurred - i know that is hard to provide. System code likes to run lean and lacks the back out traces. But it would be nice to get the values of [ and ]. For some reason the stack always reports x and y empty ... It would be nice if one could put a marker inside the tacit and get the value when it reaches that point. Ie internal stop/traces ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] J IDE
On 6/6/06, RS HUI [EMAIL PROTECTED] wrote: - What is ur? - What are x and y such that the following fail: x rute y x rutt y Ah, OK, i'll send you the script again Both verbs work, just that the tacit one does not get the debug GUI to report its argument values. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] x. spelling error default
On 5/31/06, Eric Iverson [EMAIL PROTECTED] wrote: If I understand the request you can already get it with edit|configure. Edit|Configure|Color press Add, select arguments, press OK, press OK... Voila. I guess the question is what if anything should be added to the default config. Those interested should experiment and report back their experience. Specifically wrt argument coloring i have a couple niggles 1) dotted args (x., y) aren't supported rather they seem to get color from the primitives which includes copulas and parenthesis 2) trailing white space is included (specifically noticeable if underlining is used as i prefer) 3) attributes do not add eg noun italic + argument underline gives only italicized, non-underlined noun arguments - whether noun is place before or after argument in the precedence ordering ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] Re: caching of files in the IDE
Then there is the oddity that if i do a File | Refresh Tabs (while in Project or Other tabs) in the Project Manager and then go to either the Project or Other tabs and click on a Files item ... the highlighted item reflects a file in the Source tab not what it should be for that tab... i have had to shut the Project Manager down and restart it when it somehow lost the life project file. Have not been able to replicate that condition as yet. Seems again to be related to a caching action where it gets in a state which is other than what is reflected in the gui. BTW my posts still do not show up in my mail box ... gMail does not have this difficulty with other list boxes so i presume there is some interaction between the JForum server and gMail that means i need to CC myself and trust that the post actually has gone through? i recall the subject has been brought up before so i am hoping someone using gMail can let me know what i have miss-set in gMail. i can not find any settings to discard such mail. i did check this time and saw my previous post made it to the archives, so i presume everyone else saw it. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Re: caching of files in the IDE
Chris wrote: The ijx and ijs windows are read and written on request, and do not continually check whether they have been updated by an external process. This practice is common in Windows text editors, but whether J should do so is debatable - it involves a lot of mechanism that is probably more trouble than it is worth. It is easiest just to avoid editing a script in J and in an external editor at the same time. Sort of makes it difficult to create an alternative IDE w/o that support. As i said earlier i have been i need of a mechanism for maintaining my state independently of J's stability - hence the alternate open window. Project files are normal script files loaded with 0!:0, and do not interpret NB. comment lines. They are intended to be read and written by the Project Manager. In normal use, you need not modify them directly, and if you are doing so, you are likely missing some feature in Project Manager that will do it for you. If you do modify them outside Project Manager, take care to preserve their format. It looks like you changed it somehow. Hmm, guess i need to figure out how to get PM to modify them then ... perils of bootstrapping to understanding;-) Thanks for the guidance. If a project is opened in Project Manager, it is not going to continually re-read the project, just in case an external program has changed it - in this case, exit Project Manager, make your changes, then reload it. OK greg heil wrote: Then there is the oddity that if i do a File | Refresh Tabs (while in Project or Other tabs) in the Project Manager and then go to either the Project or Other tabs and click on a Files item ... the highlighted item reflects a file in the Source tab not what it should be for that tab... Is this on Windows or Linux? I have not seen any problems on Windows, but have occasionally noticed Project Manager tabs not being refreshed properly under Linux. I think this happens when a tab is empty, and will try to track this down for the release. W2K AS The tabs are not (necessarily) empty clicking on an item line replaces its content with the corresponding item line in Source. i have had to shut the Project Manager down and restart it when it somehow lost the life project file. Have not been able to replicate that condition as yet. Seems again to be related to a caching action where it gets in a state which is other than what is reflected in the gui. I don't understand this. PM should not lose a project file - though in normal use you can be prompted whether to save changes, and perhaps this was not done. i don't think i have ever been prompted in PM. i have received messages about the non existence of files etc, ie error messages, but no prompt dialogs. One change I could make is to always save changes immediately, rather than prompting for them. As long as you are backing up properly, and in particular using the snapshot feature, then saving changes immediately should not be a problem. Any thoughts on this? Probably beyond my sophistication. Don't know about the snapshot procedure yet. i am just trying to understand how the basic Life project works... and stumbling. But i gather i need to exit the PM whenever i change any file. This does seem awkward and i had sort of thought its File | Refresh Tabs did the equivalent of that - like a Revert item in the File menu of a text editor: makes sure everything is in sync with what is on the disk. That might be a good middle way:-) ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] self postings WAS: caching of files in the IDE
On 6/15/06, Chris Burke [EMAIL PROTECTED] wrote: greg heil wrote: this. Otherwise i think there may be a Mailman setting that does prevent posts being echoed back to the author. It is a common feature for mailers. Usually though one has access, individually, to setting that option, but there seem to be no end user controls on this listbox... Thats it! You must have set the Mailman option to not receive your own posts. See http://www.jsoftware.com/jwiki/System/Forums, section Options/Unsubscribe for how to change your options. Ah i do have some control there! i looked in at my settings which would all have been in default as i had not touched them, they all looked safe, and correct: in particular Receive your own posts to the list? is set to Yes There is this potential setting as a source of the issue: Avoid duplicate copies of messages? which comes into effect When you are listed explicitly in the To: or Cc: headers of a list message, which i am not (though i have started putting myself on BCC due to not receiving them) In any case it is now set to Yes and i will change it to No to force it, hopefully, to send me copies. Seems harmless and might be the problem. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] self postings WAS: caching of files in the IDE
Bill and Björn: On 6/15/06, Björn Helgason [EMAIL PROTECTED] wrote If you use gmail you can cc or bcc to yourself as well as to the list and then get your own mal in the inbox That is what i am resorting to. Otherwise if you want to read your own mail to the list before someone replies on it you can read it in sent mail Yes, but i want it sent to the mailbox that my j gMail is received through not the gMail box that it is sent through. Not necessarily the same thing;-) In gmail you can put you mail in groups one or more and then find it in the groups and you do not need to keep a copy of it in the inbox Really? i looked for that capability and could not find it. i got the impression that Googles mantra was search, don't sort into folders. There are a limited number of Labels which can be set with filters and i do use those, though not in my writing gMailbox.. It is a good practice not to keep the mail in the inbox in my writing gMailbox i Only use the Inbox It is a purely transactional account. i delete everything once i have attended to the matter. and no file movements are automated, full manual. If i need something which turned out to not be fully attended to after deletion i pull it from the deletion by search or, in the rare instance it is past the Google deletion daemon's harvest time, i go back to the separate gMail account which archives the mailing - my Archive gMailboxes. So far it seems a good system. Would like to automate it. Anyone understand AJAX? There are bugs in the interaction of Googles AJAX, for sure. Eg periodically it crashes FireFox. That seems mostly related to the Go Back functionality possibly in relation to its https security code. So it would not surprise me if there were a bug in gMail which is the issue. As Bill Lam suggests... On 6/15/06, bill lam [EMAIL PROTECTED] wrote: I've reported this problem earlier and some members confimed the same problem. I subscribe to another mail list using mailman+gmail and have the same problem. My experience can be summarised as. mailman + gmail = can not receive own post mailman + non-gmail = no problem non-mailman + gmail = no tested We should keep our eyes on #3 there may be issues with Yahoogroups too. I guess this is a problem of gmail itself, May be Chris could test it using a gmail account. Some one should do it in an isolated account with full privileges on both sides of the problem for sure. It is becoming difficult to manage the debugging from here ... eg i need to do all the BCC'ing to ensure i archive my half of the story, and thus i cannot test all the variables. It might be a worthy investigation. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] receiving own posts
On 6/15/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: WRT receiving your own posts (or not): I assume you've turned on (or perhaps toggled off and back on) the option Receive your own posts to the list? in your list options? No i had not touched the controls, they were in default. Have you tried subscribing another mail address (not under gmail) to see if that works? No, just trying to survive on this box;-) There's also an option Receive acknowledgement mail when you send mail to the list? That is set to to No. Is an acknowledgement the same as the outgoing Posting ? If so then perhaps it is poor wording ... and is the issue. Let us know... i will switch to Yes ~ greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] receiving own posts
On 6/15/06, greg heil [EMAIL PROTECTED] wrote: On 6/15/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: There's also an option Receive acknowledgement mail when you send mail to the list? That is set to to No. Is an acknowledgement the same as the outgoing Posting ? If so then perhaps it is poor wording ... and is the issue. Let us know... i will switch to Yes Well so far it just fills my mailbox with meaningless (no content) ACK's: From: [EMAIL PROTECTED] [EMAIL PROTECTED] Mailed-By: jsoftware.com To: [EMAIL PROTECTED] Date: Jun 15, 2006 10:57 PM Subject: Beta post acknowledgement Not enamoured with the idea of leaving it on ... unless that is a cure ... and the only cure. So far i have no evidence it is even a cure. as the only copies i am getting are my BCC's ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] form edit undo
Any value to the behaviour of undo in an ijs window: which wipes the file out, if performed after a Form Edit modification? A second undo finishes the undo of the Form Edit operation. A subsequent redo goes back to the clean slate, with a second redo again implementing the form edit operation. The extra stage of a blank screen could be kinda distressing... -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] keystrokes in dialogs
Thanks Bill two separate issues: -- IDE GUI issues On 7/7/06, bill lam [EMAIL PROTECTED] wrote: greg heil wrote: Dunno if this is new in the betas, but there do not seem to be keystroke equivalents for dialog buttons - though there are underlines Did you request something like the keypreview in VB? i am using pure J, no Visual Basic, if that is what you mean. Though in that instance i am referring to the J GUI's dialog, for Find (there are similar issues with other dialogs). i consider it a bug/fault to have keystroke indications and not respond to keystrokes. Perhaps it is buried in Configure somewhere to turn them on? Can't find it Maybe related to -- my GUI issues If I remember correctly, there is a flag to enable or disable keypreview. If it is enabled every keystroke event will be reported as if for the parent form. Inside parent form's event, user can further allow/disallow the same key event be sent to child control by modifying an output parameter. Any citation? Since J event handling does not allow output parameter, there may be difficulty to implement what you requested. i was taking what hints i could from the end of the 2D Graphics section of J4C chapter 34 on graphics (p205). Henry suggests getting char Events from a call like formname_isigraphname_eventname_formlocale_ '' That event is supposed to populate sysdata. Is that what you need/mean by an output parameter? Also hoping to get/utilise Mouse Events, but i need to first figure out the 4 locatives of that callback to get the data patched into my code when an event occurs. ... And possibly set the flag you speak of? As always documentation pointers appreciated. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] keystrokes in dialogs
On 7/7/06, Henry Rich [EMAIL PROTECTED] wrote: greg wrote: formname_isigraphname_eventname_formlocale_ '' That event is supposed to populate sysdata. Is that what you need/mean by an output parameter? Also hoping to get/utilise Mouse Events, but i need to first figure out the 4 locatives of that callback to get the data patched into my code when an event occurs. ... And possibly set the flag you speak of? Only 1 locative. formname_isigraphname_eventname is the name of the verb, and formlocale is the name of the locale the form is running in. Multiple locatives apply only to objects, as in a__b__c . i have a file win.ijs with my GUI code, which is loaded direct to base. It successfully receives signals to events like schell_sctrl_fkey and embeds an isigraph spaceG. So... i want to be signaled when an arrow key is hit and then access which key was stroked... How would i reference such a keystroke event, form the name of that call back's verb? Thanks ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] pocketpc beta available
On 7/8/06, Eric Iverson [EMAIL PROTECTED] wrote: The 601 o beta is available for the pockepc. http://www.jsoftware.com/download/j601obeta_pocketpc.zip I'm afraid the answer is that J601 (and later) require Windows CE 4.20 and we won't be doing the considerable work to get it to work on earlier versions of wince. Arrgh. i completely sympathise with the lack of resources you have to make such things possible:( i expect i am in the same boat with my old iPaq 3870 though. i hope at some point J is successful enough that you can spawn off a legacy version into an open source project which can be more accommodating to compatibility related issues. Quite a serious issue all around and getting more so day by day. i do wonder if a portable Linux target would not be a more stable target though? i would prefer to buy into such a device than remain indefinitely on the Micro$oft escalator. Always having to buy new hardware is not good ecology. And it limits what we, in turn, can make available to our clients. Hopefully ISI will never require me to go to an Activated version of MS Windows in order to use/develop J under Windows. Or at least they would keep parity with Linux version betas in such an eventuality. Is Linux development advancing enough to become the primary platform soon? Sorry, i just had to wince at this one. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] accelerator keystrokes
In the keystrokes in dialogs thread in Programming i wrote: Though in that instance i am referring to the J GUI's dialog, for Find (there are similar issues with other dialogs). i consider it a bug/fault to have keystroke indications and not respond to keystrokes. Perhaps it is buried in Configure somewhere to turn them on? Can't find it OK ~system/extras/help/user/acc_keys.htm says these are accelerator keys and should respond to an Alt+the character. But they do not, at least not for me on j601binobeta_win (though they do in j504binb_win), eg in the Find dialog. And i cannot get them to work in my own dialogs. Am i the only one with this issue? And of course i am still assuming others see my posts, wish i could get to the bottom of this nondeliverance issue of mailman/gmail. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Pocket PC Beta
On 7/12/06, Oleg Kobchenko [EMAIL PROTECTED] wrote: One should be careful and read the startup guide. First time I was trying to do a reset, it didn't work. As it turned out I was poking the microphone, ... it was lucky to have survived. Unfortunately people don't come with such manuals. Early on i did the same thing to my ear. Fortunately i had two ears, and so i can still hear with the left ear. Never did resolve that reboot issue though. Wrt to the PPC. if i restrict myself to linguistic issues rather than gui ones is there a reasonable way to emulate J6 code on a PPC v3 machine? Some things are deprecated going from 5 to 6 but it would be nice to be able to continue to use my iPaq 3870 to practice J coding. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] accelerator keystrokes
Find dialog has been reworked. That's great! There is now some keystroke interaction and the ability to repeat finds accounts for maybe 40% of usage so one can significantly often avoid reopening the find dialog. However reopening the find dialog can be painful and the keystrokes defined in the find dialog lose their significance back in the document. Perhaps if the find keystrokes were the same, eg ctrl-F in both places, and all buttons had hot keys in the document context... following some user configurable/standard convention. Or a quick stop gap would be to make it easy to just get back to the find dialog :( ctrl-f loses its meaning to open the find dialog, and multiple Alt-Shift-Tab is arduous and slow ): Say you use ctrl-f to get you whichever of find/replace that was last open, (similar to the current logic with no find window open) and the new key defs become active in the document window. That, with a few more hot keys, would do it. But better might be to follow a convention document like whatever serves for such at Microsoft. First trick is to get everything hot keyed in both contexts, then second, to plug in your key remapping capability. ~greg -- The p beta for Windows is available at: http://www.jsoftware.com/download/j601pbeta_win.exe Most problems discussed in the forum have been addressed. isigraph control now supports events for middle button, mouse wheel, and focus (gained and lost). See Help|User|Window Driver Command Reference Overview|isigraph events. Printer support is now complete. See labs Print Basic and Print. Find dialog has been reworked. PCRE (Perl Compatible Regular Expression) now works with UTF-8 encoding. Use with data that is not UTF-8 can give unexpected results. Verb rxutf8_jregex_ with an argument of 0 selects the old, non UTF-8 behavior and an argument of 1 selects the new UTF-8 behaviour. The PCRE shared library is now built with the July 2006 6.7 source (previous version was with theSept 2004 5.0 source). -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] x m} y
On 7/21/06, Roger Hui [EMAIL PROTECTED] wrote: If m is not a gerund, x m} y is formed by replacing by x those parts of y selected by m{ (an error is signalled if such selection requires fill). i like that: it will expose any problems in users code, and open the way in the future for new functionality in the space where an error is declared! ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] accelerator keystrokes
The lack of keystroke equivalents seems pervasive, at least throughout J601beta... Eg these dialogs all have underlined letters that do not respond to ctrl or alt keystroke combinations: Studio | Lab Edit | Configure Run | Project Manager File | Recent Edit | Find in Files This one does seem to work... File | Open But it uses Alt-O for open which follows the windows convention, though i think the J convention is to use control key? ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] accelerator keystrokes
[EMAIL PROTECTED] wrote: I don't understand the complaint. Things work for me as I would expect in windows. Let's follow through on a specific case. start J mouse click Studio menu mouse click Labs... Alt+R (underlined shortcut letter on the Run button) selected lab is run Does this not work for you? Please give a detailed example (as simple as possible) of what you think is wrong. That helps to know it is supposed to be ALT:-) And from a clean slate it works. In particular the pure keystroke sequence AltS-L-AltR works. However there seems to be a state change that happens... yuck:( And with lots of experimentation i am not getting a repeatable way of causing it. But it is not long before i lose the ability to run As-l-Ar Once gone i can not get it back. This makes me think it is not a keyboard issue from my overly smart keyboard: no amount of multiple toggles on the alt key etc seems to bring it back, but rebooting J works like a charm. Ah! This is repeatable: As-l-Ar As-l-Ar As-l-Ar ... First two do a run of the first Lab ... thereafter i have lost Alt key capability: including the other hot keys on the form. And at least on some other forms, eg Studio | Author no longer responds to Ac to close the form. Closing the form reopening fails to help Utterly peculiar and i suppose there could be something idiosyncratic about my machine. So if anyone else can repeat it... ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] j601tbeta_win.exe
On 8/19/06, Eric Iverson [EMAIL PROTECTED] wrote: What do those apps do with keyboard input in the same case. That is, as the focus for the wheel moves does the focus for up/down arrow also move. If so, I think it simply means that mouse move events over the control do an automatic setfocus. Can you verify if this is or is not the case? Eg for the case of FireFox and my text editor (TextPad) the arrow keys affect whatever is in focus and the wheel affects whatever the cursor is over. Independently. i have used this feature a lot to scroll around w/o grabbing focus and bringing a window to the top. i can scroll any window while typing into another. Is the gl2 events Lab awaiting resolution of these issues? i wouldn't mind sneaking a look at a beta of that. i am feeling a bit lost as i came in trying to figure out GUI (J's and generally) ... at the same time it is being all turned upside down;-} Oh well, all for the best i am sure. i would appreciate any pointers to material covering what is known about how the J6 events are accessed etc ... i will need quite a range of services from those events. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] u M. (Memo)
On 9/8/06, Roger Hui [EMAIL PROTECTED] wrote: The point of fib is not that it is better than f7a, but that M. removes the exponential time penalty associated with a straightforward recursive algorithm. It can also reduce the recursion stack overhead. Using a memoized K verb i was able to push two levels deeper than others had with Ackermanns function a few years ago. ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] end of a long road
On 9/7/06, Dan Bron [EMAIL PROTECTED] wrote: The J601 beta is over and the J601 release is now available for all platforms. Aww, I was hoping to find out what came after Beta Z. Well someone had to say it! i think we should all jump up and down and scream to demand to know what would have happened! Or at least we should have been more diligent and found more bugs and delayed the release more;-) When, if ever, will this opportunity re-arise?? ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] cannot use J6.02
Looks like i am stuck at J6.01, as J6.02 says get Visual C++ 2005 SP1 Redistributable Package (x86) While trying to install the above package a closing dialogue box says: This installation package is not supported by this processor type. Contact your product vendor. Probably the problem is that i am running a non activation MS system called 2000 rather that the newer XP and Vista systems which require (buying ) call home capabilities. As i am unlikely to update to vista or XP i am likely stuck at J6.01. Unless the capabilities needed in that MS product are included in some future J release. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] cannot use J6.02
According to Chris (thanks!) this bounced Subject: Re: [Jbeta] cannot use J6.02 From: greg heil [EMAIL PROTECTED] Date: Wed, 3 Oct 2007 17:03:05 -0700 To: Beta forum beta@jsoftware.com On 10/3/07, greg heil [EMAIL PROTECTED] wrote: On 10/3/07, Eric Iverson [EMAIL PROTECTED] wrote: I hadn't realized that our move to the new MS development environment implied a decomittment of Windows 2000. That is unfortunate. Yes :-( Ah, following Chris' advice i downloaded it again and it worked. (seems i do not know German so well) Seems i had downloaded the 64b version. Now i have a J6.02 system on my MS 2000 OS! ~greg -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] Re: [Jprogramming] Hex conversion error
Crashes 602 beta d too. On Nov 28, 2007 4:32 AM, Arie Groeneveld [EMAIL PROTECTED] wrote: ... dfh17=: 16 #. 16x | (H,h) i. ] dfh17 14#'1' ... This works fine under 6.01 on W2K (gets a value error for H) but crashes 6.02c - after a few seconds, it closes the main window. ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Problem migrating a project to beta
On Jan 25, 2008 6:32 PM, Henry Rich [EMAIL PROTECTED] wrote: This problem is always with us when there are incompatible changes. I'm just saying that if 1!:45 could be quietly supported, the problem would disappear, and wouldn't that be nice? 1!:45 works for you in j602?? it works for me it J601 but fails in J602. Whether i get a domain error on boxing the arguments seems to depend on the number of arguments with 1 or fewer demanding an unboxed argument. Altogether this demands a memorization of the form of each foreign. ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] form editor
Ric There was indeed a typo as you suggested. It should have been a * as it was in the actual dialog. Windows does not let me cut and paste from a dialog;-( Everywhere else i cut and pasted. Eg the verb dice_roll_button is the same as yours... and the labs. my entire lab ijs is: DICE=: 0 : 0 pc dice; menupop File; menu new New ; menu open Open ; menusep; menu exit Exit ; menupopz; xywh 38 1 34 12;cc ok button;cn OK; xywh 38 16 34 12;cc cancel button;cn Cancel; xywh 2 1 34 12;cc roll button;cn Roll; xywh 2 16 34 12;cc values button; pas 6 6;pcenter; rem form end; ) dice_run=: 3 : 0 wd DICE NB. initialize form here wd 'pshow;' ) dice_close=: 3 : 0 wd'pclose' ) dice_cancel_button=: 3 : 0 dice_close'' ) dice_ok_button=: 3 : 0 wdinfo 'OK button pressed' ) dice_roll_button=: 3 : 0 wd 'set values *',::?6 6 ) dice_run'' On Feb 10, 2008 2:16 AM, Sherlock, Ric [EMAIL PROTECTED] wrote: I think you have a typo in your dice_roll_button verb. The lab specifies the verb as follows: dice_roll_button=: 3 : 0 wd 'set values *',::?6 6 ) I.e. you have a @ instead of a * -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of greg heil Sent: Sunday, 10 February 2008 19:33 To: Beta forum Subject: [Jbeta] form editor Running the Form Editor lab it evokes a domain error at item 25: domain error: dice_roll_button wd'set values @',::?6 6 This in 6.02h on Win2K ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] issues not yet resolved
pmove is still not working as of 6.02b(q)... On Sat, Feb 16, 2008 at 6:04 PM, greg heil [EMAIL PROTECTED] wrote: pmove information is lost (ie that line in the form noun is stripped back to pas info) when the form editor uses several options including nomin, nomax, right bottom. Putting the pmove on a separate line from pas loses the pmove line even if nothing is done in the FE, except OK to close. This is true of 6.01 and 6.02b(h). ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Gtk IDE
a nice principle is to have the UI duplicated: mousing, keyboarding, few mixes. Then all (most) in either. ~greg krsnadas.org -- fromChris Burke cbu...@jsoftware.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date18 May 2010 04:19 subject Re: [Jbeta] Gtk IDE On Tuesday, May 18, 2010 01:00 AM, Dan Bron wrote: Chris Burke wrote: Right double-click won't work, but I instead made it so that you need to hold down Ctrl (or Shift) for the double-click to load Find in Files, otherwise the word is selected as usual. I suggest that left-doubleclick with no metakey modifier means highlight the word, as you've describe here, and left-SINGLEclick modified with CTRL means find in files, but left-SINGLEclick modified with SHIFT means edit 'selectedword'. That is, the metakey modifiers apply to single-clicking, not double-clicking, and that ctrl- and shift- have different meanings. This is more consonant with current Windows conventions. Also, edit is a nice productivity tool, and SHIFT-leftclick to edit a definition will be familiar with Dyalog users. Good idea, thanks. The Ctrl key is defined as suggested. For Shift, it is defined as search word assigned - the problem is that we don't know exactly where a definition is sourced, since there could be several such. It could be improved so that if there is only one hit, go to it. -- frombill lam bbill@gmail.com reply-toBeta forum beta@jsoftware.com to beta@jsoftware.com date17 May 2010 21:04 subject Re: [Jbeta] Gtk IDE ???, 18 ??? 2010, Sherlock Ric ?(?): On Win7, 32-bit Choosing View|Toggle Line numbers in the edit Window without an open script causes J to crash. If a script file is open in the edit Window then it seems to work fine. Ric source code are available in addons svn, do 'svn update' will fetch related into directory gui/gtk and gui/gtkide And then you may debug and patch. -- fromChris Burke cbu...@jsoftware.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date17 May 2010 09:31 subject Re: [Jbeta] Gtk IDE On Monday, May 17, 2010 08:14 PM, Sherlock, Ric wrote: From: Dan Bron Right now, the only complaint I have is that in the IJX window, pressing home brings me to the 0th column (left-hand-margin), so I have to hit it twice to get to the true beginning of the line. This is in contrast to the old C++ IDE, and is a productivity friction, for me. I hadn't noticed that yet but now that you point it out I agree that the J6 behaviour (both in IJX IJS windows) was preferable. Fixed now, thanks. In a similar vein, I initially liked the idea that double-clicking on a word did a Find-in-Files on that word. However with more extended use I find that I am very used to double-clicking within a word to select the whole word (this is common behaviour across many editors and word processors including J6). I think I would prefer to bind the Find-in-Files behaviour to something else (is double-right click possible?) Right double-click won't work, but I instead made it so that you need to hold down Ctrl (or Shift) for the double-click to load Find in Files, otherwise the word is selected as usual. -- fromDon Guinn dongu...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date17 May 2010 05:40 subject Re: [Jbeta] Gtk IDE Looks good. No comments yet except it would help if you would add the date and possibly time the download was created. That way one doesn't have to retrieve the download to see if it is newer than what was already downloaded. Kind of like the Please listen to all the options as they have changed message that's six months old. Also, looks like the J Beta page is out of date as it lists betas for several systems and does not reflect this download. On Sun, May 16, 2010 at 9:30 PM, Chris Burke cbu...@jsoftware.com wrote: New beta zips are on the website. These fix a bug in pacman in Windows. Also, the library is not yet available in pacman. So until pacman is working fully in J7, please download the beta zips. -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] Syncing Jsoftware-hosted JHS session
i have gone into a JHS session hosted on the Jsoftware server with two separate machines. Can work ... but syncing the history between the two machine/sessions seems to require: 1 logout 2 password 3 ijx button in FF. In contrast DropBox is fine with just a refresh in FF. Are their alternative/easier/faster approaches available/planned? ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Syncing Jsoftware-hosted JHS session
Eric Yes! jlog _ on a second browser session reSynchs them to their commonality. This is better easier/faster than my three step method. But i have 3 machines in my work area and it would be nice to do a faster synch. Is there a way to map that to a keystroke? ~greg krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date11 August 2010 16:32 subject Re: [Jbeta] Syncing Jsoftware-hosted JHS session Not quite sure what you mean by syncing the history. There is only 1 J task and it is always in sync. I assume you mean the views in the different browser sessions. Normally a browser interaction only displays its input and output. This means it doesn't see actions in another browser session. This is good if you are focused on what you are doing in the current session. You can manipulate the log to your hearts content. Right now there are just 2 options: jlog 0 NB. clear the local log and just see new transactions jlog _ NB. show the entire log of all transactions since J started -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date11 August 2010 16:24 subject Syncing Jsoftware-hosted JHS session i have gone into a JHS session hosted on the Jsoftware server with two separate machines. Can work ... but syncing the history between the two machine/sessions seems to require: 1 logout 2 password 3 ijx button in FF. In contrast DropBox is fine with just a refresh in FF. Are their alternative/easier/faster approaches available/planned? ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Syncing Jsoftware-hosted JHS session
Eric Sounds very good. i think though on the JUM that file is a system file and needs admin privleges to edit? ~greg krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date11 August 2010 17:06 subject Re: [Jbeta] Syncing Jsoftware-hosted JHS session Not a single keystroke, but a simple series in the current shortcut scheme. For exmaple ctrl+/ r could be defined to be show entire session log. Not sure this would be good practice in general, but could be very nice in certain work scenarios. Currently no easy way to edit IDE menus or shortcuts, but you could do the following (untested) to make jijx shortcut r run jlog _ *** ~addons/ide/jhs/jijx.ijs edit function doshortcut to be as follows: function doshortcut(c) { switch(c) { case 'r': jdo(jlog _,true,[]); default: dostdshortcut(c); break; } } case 'r': -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date11 August 2010 16:44 subject Re: [Jbeta] Syncing Jsoftware-hosted JHS session Eric Yes! jlog _ on a second browser session reSynchs them to their commonality. This is better easier/faster than my three step method. But i have 3 machines in my work area and it would be nice to do a faster synch. Is there a way to map that to a keystroke? ~greg krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date11 August 2010 16:32 subject Re: [Jbeta] Syncing Jsoftware-hosted JHS session Not quite sure what you mean by syncing the history. There is only 1 J task and it is always in sync. I assume you mean the views in the different browser sessions. Normally a browser interaction only displays its input and output. This means it doesn't see actions in another browser session. This is good if you are focused on what you are doing in the current session. You can manipulate the log to your hearts content. Right now there are just 2 options: jlog 0 NB. clear the local log and just see new transactions jlog _ NB. show the entire log of all transactions since J started -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date11 August 2010 16:24 subject Syncing Jsoftware-hosted JHS session i have gone into a JHS session hosted on the Jsoftware server with two separate machines. Can work ... but syncing the history between the two machine/sessions seems to require: 1 logout 2 password 3 ijx button in FF. In contrast DropBox is fine with just a refresh in FF. Are their alternative/easier/faster approaches available/planned? ~greg krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] JUM JHS updated
Eric How does one kill' a task on the JUM? i got a new login with 1.0.25 - and it is stuck there - and the older login is stuck at an even older release:-( Loging out, killing cookies, does not seem to be enough. greg ~krsnadas.org -- fromDon Guinn dongu...@gmail.com to Beta forum beta@jsoftware.com date16 October 2010 14:55 subject Re: [Jbeta] JUM JHS updated Message on jum form: SECURITY! Log out when you are done! Click jijx logout button or close browser (not just tabs!). Don't quite understand what you want done. There is no jijx window available to logout. I'm running Chrome. Why do all the sessions in Chrome need to be stopped? Isn't just closing the tab sufficient? Closing the browser does not really close the browser. It is still running, though no forms open. -- On Sat, Oct 16, 2010 at 4:38 PM, Eric Iverson eric.b.iver...@gmail.comwrote: JUM JHS has been updated to latest version. You will have to use JUM to kill your current task and start a new one with the new version. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] JHS JAL update 1.0.29 available
JUM does not seem to obey USER/config/jhs.cfg ...at least the pswd is not synced greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date18 October 2010 08:57 subject Re: [Jbeta] JHS JAL update 1.0.29 available Tanks for the upgradeable button -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date18 October 2010 08:49 subject [Jbeta] JHS JAL update 1.0.29 available JHS JAL update 1.0.29 available. The latest release fixes problems with: ijs bs/del standard browser shortctus on ijs page file page support for delete/new file etc. This release has been installed at Jsoftware JUM JHS. You need to kill and restart your task to run this latest version. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] default switching jhs
i set up a htm page consisting of an ijx, ijs and help frame from the JUM: htmlheadtitlexsh/title/head frameset cols=44%,* frame src=http://www.jsoftware.com:53692/jijx; name=ijx, scrolling='auto' frameset rows=77%,* frame src=http://www.jsoftware.com:53692/jijs; name=ijs, scrolling='auto' frame src=http://www.jsoftware.com:53692/jhelp; name=jhelp, scrolling='auto' /frameset /frameset /html The next step is to set the IJS file to be a particular file (o/w the system keeps generating a new one). But i cannot find a URL for such a file. I can modify any file, but i cannot get the URL for that modification... greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date24 October 2010 08:14 subject Re: [Jbeta] default switching jhs Finally got some time to test your suggestion here. Pretty nice! javascript:window.open(prompt('url'), '_blank','status=0,scrollbars=0,location=0,menubar=0') javascript:window.open(http://www.jsoftware.com;, '_blank','status=0,scrollbars=0,location=0,menubar=0') javascript:window.open(http://www.jsoftware.com,mywindow;) javascript:window.open(http://127.0.0.1:65001/jhelp;, '_blank','status=0,scrollbars=0,location=0,menubar=0') javascript:window.open(http://127.0.0.1:65001/jijx;, '_blank','status=0,scrollbars=0,location=0,menubar=0') javascript:window.open(http://127.0.0.1:65001/jijx,mywindow;) -- 2010/10/22 Raul Miller rauldmil...@gmail.com On Fri, Oct 22, 2010 at 8:59 AM, Brian Schott schott.br...@gmail.com wrote: My understanding of browser interfaces is that such tessellation of frames/windows is not possible by the end user, but if it is I would like to know how, please. You can have multiple browser windows though. And you can size and arrange them as you please. I think the biggest limitation here is the extra real-estate imposed by the browser itself. However, chrome minimizes that, and also it is possible to instruct the browser to open a window without decorations (using javascript's window.open() method -- though if you are willing to use that you also have some other options available to you). Though, actually, you could make a bookmark and edit. For example, change it to: javascript:window.open(prompt('url'), '_blank', 'status=0,scrollbars=0,location=0,menubar=0') This lets you type in a url and get a window visiting that url with a relative minimum of decoration real-estate. You could also make multiple bookmarks, replacing the prompt('url') with whatever url you want (in quotes). Or instead of having the bookmark always spawn a new window you could change the second parameter from '_blank' to some unique name for that bookmark and the bookmark would then bring the named window to the front if it were already open... FYI, --- Raul -- from Alex Rufon alex_ru...@ist-systems.com to Beta forum beta@jsoftware.com date 22 October 2010 07:01 subjectRe: [Jbeta] default switching jhs Come to think of it ... I also do this a lot. I would normally have the IJX, PM and 1 to 3 IJS files open at the same time. You see, I sort of put all my global variables and global initialization scripts in two different files: defs.ijs - global variables, sql commands, structures, constants init.ijs - initialization scripts I sort of keep this two files open (particularly defs.ijs) ... In JHS, can I open scripts in another TAB or another browser window? Would that help Brian? Although I don't have time to look at JHS, I'm looking at java web application development and later on integrating J through/as web service. I'm currently drowning in reading materials for VAADIN http://vaadin.com/home DROOLS http://www.jboss.org/drools SPRING Framework http://www.springsource.org/ My ears are bleeding ... :P -- From: beta-boun...@jsoftware.com [mailto:beta-boun...@jsoftware.com] On Behalf Of Brian Schott Sent: Friday, October 22, 2010 9:00 PM To: Beta forum Subject: Re: [Jbeta] default switching jhs When Bjorn mentioned labs in the context of having frames it finally reminded me of how desirable it is to have multiple windows visible at once in J. On the Mac I have always been able to reposition and resize multiple windows to tessellate them, and that has been most desirable with labs and with simulation ogl graphics that are refreshed when code lines are entered in the ijx window (Eric has not quite gotten that to work for me in the last two revisions, btw). My understanding of browser interfaces is that such tessellation of frames/windows is not possible by the end user, but if it is I would like to know how, please. (B=) -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date22 October 2010 01:12 subject Re: [Jbeta] default switching jhs It is one thing what you can do as a knowledgeable user and send the pages into frames, windows or
Re: [Jbeta] default switching jhs
The 'URL' stays the same but the JUM computer keeps a referent to a value which changes. The ijs referent is always http://www.jsoftware.com:53692/jijs# Addons get an error page. i can also affect an ijs file by the jfile link ... but again that 'URL' is fixed and not related to its position in the file structure. So to get a particular ijs file in the ijs frame so far i need to do something outside of a url, perhaps in a frame - i cannot, so far, affect it by giving a URL in a preset htm file. Perhaps some additional, defining, information is/can be passed to the j machine after a ?? greg ~krsnadas.org -- fromRaul Miller rauldmil...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 06:38 subject Re: [Jbeta] default switching jhs On Sun, Oct 24, 2010 at 5:38 PM, greg heil ghei...@gmail.com wrote: i set up a htm page consisting of an ijx, ijs and help frame from the JUM: htmlheadtitlexsh/title/head frameset cols=44%,* frame src=http://www.jsoftware.com:53692/jijx; name=ijx, scrolling='auto' frameset rows=77%,* frame src=http://www.jsoftware.com:53692/jijs; name=ijs, scrolling='auto' frame src=http://www.jsoftware.com:53692/jhelp; name=jhelp, scrolling='auto' /frameset /frameset /html The next step is to set the IJS file to be a particular file (o/w the system keeps generating a new one). But i cannot find a URL for such a file. I can modify any file, but i cannot get the URL for that modification... Hmm... I have not been using jhs recently, but one additional possibility is that you can name a frame -- frame name=examplename ... -- and then use that name as the second argument to window.open -- window.open('url', 'examplename'). Or you can use that frame as the target of an anchor -- a target=examplename href=url. This does not solve the problem you are trying to solve (getting the url of an ijs ahead of time) but maybe you can open an ijs window and copy the url out of it? In other words, perhaps something like this: frame name=ijstarget ... And then open the ijs window, copy its url and visit: javascript:window.open('URLGOESHERE', 'ijstarget') (replacing URLGOESHERE with the url you copied). -- Raul -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] default switching jhs
Ah, i put a file frame in like this htmlheadtitlexfsh/title/head frameset cols=44%,* frameset rows=44%,* frame src=http://www.jsoftware.com:53692/jfile; name=file, scrolling='auto' frame src=http://www.jsoftware.com:53692/jijx; name=ijx, scrolling='auto' /frameset frameset rows=33%,* frame src=http://www.jsoftware.com:53692/jijs; name=ijs, scrolling='auto' frame src=http://www.jsoftware.com:53692/jhelp; name=jhelp, scrolling='auto' /frameset /frameset /html then i navigated in the file frame to the correct ijs window and was able to affect the ijx from there. To see that effect i navigated to the ijx frame and did a esc-j. To minimize keyboarding for navigation the ijs needs to be next to the ijs frame... unless there is another mechanism, than tabbing, for moving inter frames..? The help frame seems the only one affected by the scrolling option, and the ijx frame the only one not refreshable by a right click. greg ~krsnadas.org -- fromBrian Schott schott.br...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 10:00 subject Re: [Jbeta] default switching jhs 2010/10/25 Björn Helgason gos...@gmail.com: It looks good to have ijx and ijs side by side like this. Yes, this juxtaposition is good. It works to send lines from the ijs to the ijx for execution but the ijx is not refreshed. Yes, that is very undesirable. I created an ijs with an error on purpose and the error message only appears in the ijs frame, not the ijx frame even after a refresh. After you do a refresh you see that the lines are there. I refresh by typing Esc,j in the jijx window. Is there another way to refresh? It is interesting that the ijx frame is different from the ijs and help when you press the right mouse button. Yes, I wonder if that is a function of the ordering of the frames, where the first one is treated differently? -- fromBrian Schott schott.br...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 09:08 subject Re: [Jbeta] default switching jhs Greg, I also tried your frame approach below successfully in the online jum/jhs with Mac Safari and FireFox. I found one workaround for your jijs problem was to change the link in the ijs frame from jijs to jfile and then to select the file I wanted. I am having some difficulty navigating among the frames using only keystrokes. My only choice so far is to use the Tab and Shift+Tab keys to move forward and back but that is very limiting because the sequence of moves includes more than the three frames/panes. For example in each frame that has the menus, the keystrokes stop at each menu. I wish only the Esc,1 sequence would take me to the menus. Btw, I also wish there was an easy way to use the Esc key to escape from the menus, but so far my only alternative to escaping from a selected menu selection such as link|ijx is to arrow right or left and then press enter. Thanks for the html template. -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 07:47 subject Re: [Jbeta] default switching jhs It looks good to have ijx and ijs side by side like this. It works to send lines from the ijs to the ijx for execution but the ijx is not refreshed. After you do a refresh you see that the lines are there. It is interesting that the ijx frame is different from the ijs and help when you press the right mouse button. -- 2010/10/25 Björn Helgason gos...@gmail.com Thx! This is pretty good. I created a local version a file called test71.html - test71.html htmlheadtitlexsh/title/head frameset cols=44%,* frame src=http://127.0.0.1:65001/jijx; name=ijx, scrolling='auto' frameset rows=77%,* frame src=http://127.0.0.1:65001/jijs; name=ijs, scrolling='auto' frame src=http://127.0.0.1:65001/jhelp; name=jhelp, scrolling='auto' /frameset /frameset /html --- Nice thing is I can change the contents of each frame as needed and the other info stays unchanged. It should be easy to put a control of this into a config file. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] default switching jhs
Ah one can navigate to a particular ijs file using jfile, when one opens it, it opens into an ijs frame. Using the 'link' mechanism one can mutate a frame to any frame type. One cannot create or destroy frames from the browser (so far as i know) so the arrangement is set. The main limitation to having a large number of frames to play with is navigational overhead. Three frames is perhaps the best compromise - as all frames are adjacent, though tabbing the menus becomes tiresome. Jumping between windows/tabs is easier, and more context preserving. From what i can tell the JUM only provides a single ijx context. greg ~krsnadas.org -- fromBrian Schott schott.br...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 11:14 subject Re: [Jbeta] default switching jhs Greg, Your newer version looks good in the browser but is not really a good alternative for me because when I refresh the jijs frame with Esc,J (capital J) I just get a new tempn.ijs, not the file I selected in the file menu. I think my solution worked better, with the jijs and jfile using the same frame. [From the messages you attached below, I am not sure you are replying to my message where I just replaced jijs with jfile in your original.] -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] default switching jhs
i do have two logins - but cookies prevent one from using both ijx's. So far i have not been able to get two working, even by switching cookies (using the cookieswap plugin for ff (chrome?)). greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 12:14 subject Re: [Jbeta] default switching jhs You could have more J sessions and more ports if you want independent ijx windows. This J browser combination is opening up a lot of interesting options. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] default switching jhs
In the JUM cookies are use for login and context. It may well be possible to get around this in ones local sandbox environment and have parallel ijx's - while still having difficulties in talking to another machine. greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 12:49 subject Re: [Jbeta] default switching jhs You set the port in c:/j701/addons/ide/jhs/config/jhs_default.ijs I do not know how cookies come into it. Have not tried to use two ports yet. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] JHS JAL update 1.0.34 available
The action menu shortcuts seem to run out of steam after 3 entries which garner q, w e. The shortcuts 1 2 seem to be tied the menu and body of a frame (perhaps a case could be made for using 3 to go to the 'scratch form). Maybe this would be a good place to try a second level? Eg q1, q2 q3... or another shortcut key, eg my editor uses ctrl-Q - though in some desktop apps this is tied to quit. The multiple demo uses the IP address of the callers machine... This may not have a J engine on it (and in many cases it is multiple machines) Perhaps if it were pointed at the currently used J engine (IP and port) it would be more useful? What is the shortcut r tied to in the ijs actions menu? It says it is tied to 'run' in the action menu but seems to have no effect on the ijx box... 'run display' which has no shortcut is the useful run for the file... i would like r tied to 'run display' (for symmetry with scratch) and 'run' be run current line. A third run item might be selection. The scratch area can be made to a larger size by dragging, but cannot be made smaller. Perhaps if it were initially a single line in size it would be more flexible. greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date25 October 2010 14:59 subject [Jbeta] JHS JAL update 1.0.34 available JHS JAL update 1.0.34 available. Try debug (see help). Try ijx action menu (see help). The idea is that this facilitates a very simple project manager. Try jdemo9 (thanks to the users who did the groundwork here). Try ijx studio|scratch.. and scratch. This puts a small text area at the top of the ijx window where you can enter multiple lines and then run them with esc r shortcut. The ijs readonly problems with some browsers has not been fixed. If this is a big problem for you, I can provide a temporary workaround. The varous cut/paste problems have not been sorted out. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] how unsplit window after run jdemo9
Do you mean go backwards in history? i use the keyboard shortcuts of alt left and right arrows alot for navigating the history. Or one can go to an entirely new URL by dragging the URL into the window. greg ~krsnadas.org -- fromMurray Eisenberg mur...@math.umass.edu to beta@jsoftware.com date26 October 2010 10:16 subject [Jbeta] how unsplit window after run jdemo9 How does one unsplit the split window created by running jdemo9? -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] paste of marked up copy
If i put 1 in the paste buffer (from an html source, eg ijx or ijs windows) and i paste into ijx and return, ijx evaluates 1. Good. If i do 5 pastes so i get 1 on the command line and i still get 1 as the evaluation result! Interspersing the pastes of 1 with numbers so i have 012131415 i get 15. Doing the same setup but moving the cursor(^): 0121^31415 i get 13 Doing the same setup but moving the cursor(^): 012^131415 i get 13 Doing the same setup but moving the cursor(^): 01^2131415 i get 12 Doing the same setup but moving the cursor(^): 0^12131415 i get 12 Doing the same setup but moving the cursor(^): 012131^415 i get 14 Doing the same setup but doubling the typed numbers and moving the cursor(^): 0012213^3144155 i get 133 Doing the same setup and moving the cursor(^): 0012213314^4155 i get 144 my guess is from the cursor position the IJX is taking the text of last paste and the tokens until the next paste as the input line. i seem to get the same results using , or ' for delimeters and either ff or chrome (on W7 and XP) in the JUM If i do a paste of a number cut from some window where that digit has a style, that style is preserved in the paste line - so my guess is the current (where the cursor is) 'text' of a markup plus the succeeding typed entries only, are used. This was verified by taking copies from a prepared page, and found to go up only one markup level. Previous and succeeding markups are cut out. They are not in the scroll buffer nor executed - though they were pasted. So a solution is to eliminate markup, before inserting from the paste buffer. After that its elimination becomes problematic as the user can 'jump' to any text spot at any time and hierarchy (which correct XML and its tools requires) is not necessarily preserved. greg ~krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] paste of marked up copy
If x has a string, doing y=:x,'Z' works, but the same line in the scratch area gets: |domain error: jloadnoun__ | y=:x,'Z' |[-0] greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date28 October 2010 06:46 subject Re: [Jbeta] paste of marked up copy There is a bug in dealing with pasted html (previously reported by Ric). I think I have a fix for this now and it will be in the next update. A workaround for now is to do paste/composition in the scratch window and then run or cut/paste to the ijx last line. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] paste of marked up copy
Eric Hmm. i ran into a related(?) problem when i worked on a string called s adding to a s to create t. ... trying to rename variables and investigate further i can no longer successfully log into either of my accounts, on two machines. Both show the entire IJX page before giving another login (with an invalid login err msg) as i attempt to do anything. Although the history is lost if i do a kill and a start. So i cannot refine this issue on the JUM. From the dropdowns of that IJX page it appears that 51724 is stuck at an earlier version, while 53692 has the 'scratch' option. The dropdowns and their headers are shown, even though the login is unsuccessful. So there are peeks at my accounts even though the login mechanism gives failure. greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date7 November 2010 16:28 subject Re: [Jbeta] paste of marked up copy The problem is using names x and y. The scratch window lines are run in a verb and that verb has local copies of x and y. -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date6 November 2010 08:42 subject Re: [Jbeta] paste of marked up copy If x has a string, doing y=:x,'Z' works, but the same line in the scratch area gets: |domain error: jloadnoun__ | y=:x,'Z' |[-0] greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date28 October 2010 06:46 subject Re: [Jbeta] paste of marked up copy There is a bug in dealing with pasted html (previously reported by Ric). I think I have a fix for this now and it will be in the next update. A workaround for now is to do paste/composition in the scratch window and then run or cut/paste to the ijx last line. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] paste of marked up copy
Eric As i noted, i did go to the JUM (http://www.jsoftware.com:50001/jum) and killed and restarted both accounts/tasks. The result was only in losing the history. i am still unable to get anything but an invalid login to both accounts. With the account 51724 it comes immediately back to a login page saying invalid login. With the 53692 account it does the same, but in addition it gives the drop downs menus of the IJX window. Runing an entry in the scratch window just gets yet another invalid login entry: they are piled on the same page with successive numbers Invalid login(i)... Creating yet a third account, 54922, yields the same pattern as 53692 on Chrome/W7. Trying again on a different machine (XP3) with Chrome and FF i get the same result on the new third account. Perhaps there is some interaction with IP numbers? greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date8 November 2010 05:14 subject Re: [Jbeta] paste of marked up copy I can only suggest using JUM to kill and restart both tasks. -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date7 November 2010 17:42 subject Re: [Jbeta] paste of marked up copy Eric Hmm. i ran into a related(?) problem when i worked on a string called s adding to a s to create t. ... trying to rename variables and investigate further i can no longer successfully log into either of my accounts, on two machines. Both show the entire IJX page before giving another login (with an invalid login err msg) as i attempt to do anything. Although the history is lost if i do a kill and a start. So i cannot refine this issue on the JUM. From the dropdowns of that IJX page it appears that 51724 is stuck at an earlier version, while 53692 has the 'scratch' option. The dropdowns and their headers are shown, even though the login is unsuccessful. So there are peeks at my accounts even though the login mechanism gives failure. greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date7 November 2010 16:28 subject Re: [Jbeta] paste of marked up copy The problem is using names x and y. The scratch window lines are run in a verb and that verb has local copies of x and y. -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date6 November 2010 08:42 subject Re: [Jbeta] paste of marked up copy If x has a string, doing y=:x,'Z' works, but the same line in the scratch area gets: |domain error: jloadnoun__ | y=:x,'Z' |[-0] greg ~krsnadas.org -- fromEric Iverson eric.b.iver...@gmail.com to Beta forum beta@jsoftware.com date28 October 2010 06:46 subject Re: [Jbeta] paste of marked up copy There is a bug in dealing with pasted html (previously reported by Ric). I think I have a fix for this now and it will be in the next update. A workaround for now is to do paste/composition in the scratch window and then run or cut/paste to the ijx last line. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] pacman permissions
options require internet access Yes, there should/needs to be access for a 'clean-room' machine, one which can/does not have continuous internet access. A zip file or a like snapshot would satisfy this need. greg ~krsnadas.org -- fromSherlock, Ric r.g.sherl...@massey.ac.nz reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date9 November 2010 14:20 subject [Jbeta] pacman permissions The GUI interfaces for pacman in J602, J701 jhs and J701 gtk all currently check to see if you have Administrator writes to the ~addons folder before letting you run pacman. I think this is a mistake. The user should be able to open pacman and browse the available addons and see what is installed/not installed without requiring write access. IMO rather than fail if 'update' fails due to checkaccess, the attempt to start pacman should succeed, but perhaps there should be a message explaining that the local catalog was last updated x days ago. From the following page http://www.jsoftware.com/jwiki/JAL/Package%20Manager/jpkg The following options require write access to the ~addons directory: 'remove' The following options require internet access as well as write access to the ~addons directory: 'install' , 'reinstall' , 'update' , 'upgrade' All other options only require read access to the ~addons directory. (history, manifest, show, search, showinstalled, shownotinstalled, showupgrade, status ) -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] pacman permissions
Yes some (multiple?) packages which can sneakernetted through to a Faraday caged machine. That machine should be fully useful and fluent J-wise. greg ~krsnadas.org -- fromSherlock, Ric r.g.sherl...@massey.ac.nz reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date9 November 2010 14:58 subject Re: [Jbeta] pacman permissions Just wanted to confirm that you mean more than just the ability to interact with pacman while offline? (which pacman already does) You mean you want to be able to copy a zip file (downloaded from JAL) to an offline machine that it can use to update its local catalog? -- From: greg heil Sent: Wednesday, 10 November 2010 11:48 options require internet access Yes, there should/needs to be access for a 'clean-room' machine, one which can/does not have continuous internet access. A zip file or a like snapshot would satisfy this need. greg ~krsnadas.org -- from Sherlock, Ric r.g.sherl...@massey.ac.nz date 9 November 2010 14:20 The GUI interfaces for pacman in J602, J701 jhs and J701 gtk all currently check to see if you have Administrator writes to the ~addons folder before letting you run pacman. I think this is a mistake. The user should be able to open pacman and browse the available addons and see what is installed/not installed without requiring write access. IMO rather than fail if 'update' fails due to checkaccess, the attempt to start pacman should succeed, but perhaps there should be a message explaining that the local catalog was last updated x days ago. From the following page http://www.jsoftware.com/jwiki/JAL/Package%20Manager/jpkg The following options require write access to the ~addons directory: 'remove' The following options require internet access as well as write access to the ~addons directory: 'install' , 'reinstall' , 'update' , 'upgrade' All other options only require read access to the ~addons directory. (history, manifest, show, search, showinstalled, shownotinstalled, showupgrade, status ) -- For information about J forums see http://www.jsoftware.com/forums.htm
[Jbeta] mailing browser
watch out for line wraps... almost a mantra on these fora;-) Is there a way, perhaps using mark down, to get around that? Generally it seems most email client apps seem to mess up the writers intentions:-(( Perhaps there is a way if only in our fora (not necessarily the wider word of emailed communication) to move beyond that into at least a modicum of formatting, without the heaviness of proprietary solutions. Just putting the pre tag around an outgoing post can help tremendously. Yahoos! online rich text mailer is very good. Is there a way to get something like that working for incoming and outgoing mailings? On the outgoing end i would like control over a raw html'd post. On the incoming i would at least like to avoid the line wraps so frequently mentioned, and avoid the alternative parts! The deficit may be at the posting client agent end:-| Perhaps a custom posting client agent could me modeled in j7? If it can sit in the JUM, anyone with an account, anywhere can post markups (or downs). A reader client agent likely needs to sit on gigabytes of secure data and thus likely off the JUM. The poster only part, only needs the inbox, or about 5 lines of header info for each post - for correctly linked replies. Or a way to include appropriate headers. Perhaps bridge apps between receiver and poster agents. Is there an existing a solution that suffices? i (and others?) have been frustrated for decades by this state of affairs. The whole may be a snarled mess, but at least, with conventions or more, it could be ameliorated within a semi isolated ecosystem. greg ~krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] mailing browser
Raul Unfortunately the line length can grow (beyond the limit of perhaps 64-70 allowed) by the addition(s) of , sigh. Most mail agents wrap like you said at some predetermined value. One can wrap the whole in html's pre tag which is relatively simple - but tricking or creating clients to use that convention is much harder. Another convention is in Markdown: where a long line is only ended with two spaces. A reader could be made using that convention: It would run together all lines not demarked by two LF's or two spaces. Unwrapping the others. It would have to rid unused 's - which might be sad if they are needed, though escaping, in that rare case, is an option. Markup symbols are also treated badly by variable width fonts, which are the norm for most readers. A wrap in a pre tag could make it all constant width, much much better for the symbology of J like languages. i do like the idea of piggybacking on internet wide proto-conventions... It makes for the possibility of reinforcing feedback loops amongst communities. Markdown is good, though its href capabilities are a bit convoluted compred to my preferred ~ method: ~ denotes a path, usually a file at a (default) location. A default TLD may be set with a ~, and dates, times, ftps, emails etc... If the reader client were made facile enough it could be generally used, even if not the vendors own. If it spoke SMTP possibly it could speak to multiple vendor servers... if it had r/w access to their inbox at a raw level. greg ~krsnadas.org -- fromRaul Miller rauldmil...@gmail.com to Beta forum beta@jsoftware.com cc Chat forum c...@jsoftware.com date11 January 2011 14:09 subject Re: [Jbeta] mailing browser On Tue, Jan 11, 2011 at 4:24 PM, greg heil ghei...@gmail.com wrote: watch out for line wraps... almost a mantra on these fora;-) Is there a way, perhaps using mark down, to get around that? In the general case, no -- each mail client implements its own rules. That said, if you keep your lines short enough (64 characters wide should be short enough, 70 might be short enough), no mail client should wrap the lines. The problem, here, is that discovering line width often means using some editor other than the one provided by the mail client. Another possibility involves wrapping the logical line in something that would recover from extraneous line breaks. For example: .0 :0-.LF LINE GOES HERE ) However, this is bulky and distracting. Another possibility involves creating a J line-end indicator which would allow a script approach that ignores the email introduced line ends. To my knowledge no one has ever bothered with such a thing. There may be other possibilities. -- Raul -- fromgreg heil ghei...@gmail.com to Chat forum c...@jsoftware.com, Beta forum beta@jsoftware.com date11 January 2011 13:24 subject mailing browser watch out for line wraps... almost a mantra on these fora;-) Is there a way, perhaps using mark down, to get around that? Generally it seems most email client apps seem to mess up the writers intentions:-(( Perhaps there is a way if only in our fora (not necessarily the wider word of emailed communication) to move beyond that into at least a modicum of formatting, without the heaviness of proprietary solutions. Just putting the pre tag around an outgoing post can help tremendously. Yahoos! online rich text mailer is very good. Is there a way to get something like that working for incoming and outgoing mailings? On the outgoing end i would like control over a raw html'd post. On the incoming i would at least like to avoid the line wraps so frequently mentioned, and avoid the alternative parts! The deficit may be at the posting client agent end:-| Perhaps a custom posting client agent could me modeled in j7? If it can sit in the JUM, anyone with an account, anywhere can post markups (or downs). A reader client agent likely needs to sit on gigabytes of secure data and thus likely off the JUM. The poster only part, only needs the inbox, or about 5 lines of header info for each post - for correctly linked replies. Or a way to include appropriate headers. Perhaps bridge apps between receiver and poster agents. Is there an existing a solution that suffices? i (and others?) have been frustrated for decades by this state of affairs. The whole may be a snarled mess, but at least, with conventions or more, it could be ameliorated within a semi isolated ecosystem. greg ~krsnadas.org -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] mailing browser
Bjorn The problem is that solution is verbose (for the user) and only works when the user has a J environment handy, it requires the user to shuttle environments ... i am hoping for a solution for more general readers and problems. greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date11 January 2011 14:24 subject Re: [Jbeta] mailing browser t=.'if you really want to write a long line' t=.t,' and split it on many' t=.t,' lines and send it off' t=.t,' you can send it off in small chunks' t if you really want to write a long line and split it on many lines and send it off you can send it off in small chunks -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] mailing browser
Bjorn What if the line wrap occurs in the middle, say between the N the B, or the B and the L, or ... Mail agents are capable of making hash of anything ... by introducing randomness. greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com reply-toBeta forum beta@jsoftware.com to Beta forum beta@jsoftware.com date11 January 2011 16:14 subject Re: [Jbeta] mailing browser If it really matters that the result line is not wrapped you might add NB. LC (Line Continues) if you really want to write a long line NB. LC and split it on many NB. LC lines and send it offNB. LC you can send it off in small chunks -- fromgreg heil ghei...@gmail.com to Beta forum beta@jsoftware.com date11 January 2011 15:53 subject Re: [Jbeta] mailing browser Bjorn The problem is that solution is verbose (for the user) and only works when the user has a J environment handy, it requires the user to shuttle environments ... i am hoping for a solution for more general readers and problems. greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date11 January 2011 14:24 subject Re: [Jbeta] mailing browser t=.'if you really want to write a long line' t=.t,' and split it on many' t=.t,' lines and send it off' t=.t,' you can send it off in small chunks' t if you really want to write a long line and split it on many lines and send it off you can send it off in small chunks -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] [Jchat] mailing browser
Alan Yes mine too:-( it is the way of modern mail agents, at least mine, gMail. Is there another which lets you choose the font for viewing? i have an audience that i send monospaced text too, poetry if you will, none of them ever receive it in a monospaced form. One would have to be in a HTML forum using Monospaced fonts for that to happen. Many fora, such as these, are text only as the 'attachments' can take up a lot of space and potentially be dangerous. greg ~krsnadas.org -- fromRaul Miller rauldmil...@gmail.com to Chat forum c...@jsoftware.com date12 January 2011 02:58 subject Re: [Jbeta] [Jchat] mailing browser On Wed, Jan 12, 2011 at 1:59 AM, Alan K. Stebbens a...@stebbens.org wrote: For example, this ASCII table presents nicely only because I'm using a monospace font on it. ;:'now is the time for all good men to come to the aid of their country' +---+--+---++---+---++---+--++--+---+---+--+-+---+ |now|is|the|time|for|all|good|men|to|come|to|the|aid|of|their|country| +---+--+---++---+---++---+--++--+---+---+--+-+---+ Interestingly enough, it was displayed using a proportional font, in my browser. -- Raul -- fromAlan K. Stebbens a...@stebbens.org to Beta forum beta@jsoftware.com cc Chat forum c...@jsoftware.com date11 January 2011 22:59 subject Re: [Jbeta] mailing browser On Jan 11, 2011, at 1:24 PM, greg heil wrote: watch out for line wraps... almost a mantra on these fora;-) Is there a way, perhaps using mark down, to get around that? Generally it seems most email client apps seem to mess up the writers intentions:-(( Perhaps there is a way if only in our fora (not necessarily the wider word of emailed communication) to move beyond that into at least a modicum of formatting, without the heaviness of proprietary solutions. Greg, Using HTML or RTF in your mail client should prevent the mailer from messing up the presentation. I use HTML all the time for tabular content so that I can use a monospace font on that content, even if the prose is my normal, variable width Verdana font. For example, this ASCII table presents nicely only because I'm using a monospace font on it. ;:'now is the time for all good men to come to the aid of their country' +---+--+---++---+---++---+--++--+---+---+--+-+---+ |now|is|the|time|for|all|good|men|to|come|to|the|aid|of|their|country| +---+--+---++---+---++---+--++--+---+---+--+-+---+ -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Installation without access to internet
Yes! Just a second as i do not have an answer. But i do have two clean room machines (Ubuntu and W7) which would very much like to have J on them. greg ~krsnadas.org -- fromdavid alis david.a...@gmail.com to Beta forum beta@jsoftware.com date13 January 2011 04:59 subject [Jbeta] Installation without access to internet Apologies if this already been dealt with. I have copied the file j701a_linux64.sh to a linux server that has absolutely no access to the internet. How do (1) I install the gtk based ide (2) how would i install and keep current the various packages, addons? Thanks David -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Installation without access to internet
Bill It is not yet done ... what is it recommended? To begin, and to continue... greg ~krsnadas.org -- frombill lam bbill@gmail.com reply-toBeta forum beta@jsoftware.com to beta@jsoftware.com date13 January 2011 17:18 subject Re: [Jbeta] Installation without access to internet That depends on how these 'clean room' machines got installed initially. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] Installation without access to internet
That sounds easy, and in line with my projected state. i wonder how i let J know of differing placements on the two machines/OS's... Or are there a parts to keep? greg ~krsnadas.org -- fromTikkanz tikk...@gmail.com to Beta forum beta@jsoftware.com date13 January 2011 18:18 subject Re: [Jbeta] Installation without access to internet If I were doing this I think the easiest option would be to also have an installation on a connected machine. I'd install and update the connected machine as usual and then copy the j701 folder from the connected machine to the clean-room machine using a USB pen drive as required. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] e-mail etc
Björn Yes if there were a programmatic email client app made one could add a mode for squirting through formatted etc arbitrary text- as it could handle both coding and decoding. greg ~krsnadas.org -- fromBjörn Helgason gos...@gmail.com to Beta forum beta@jsoftware.com date18 January 2011 05:39 subject [Jbeta] e-mail etc Now that there is such an excellent interface to J in a browser it is interesting to know if there is anyone who has made an e-mail gui in J that can be used to communicate with the Forum? Or at least to read the Forum. It would be interesting to be able to connect through J with the wiki information too. Possibly to get some off line database from these sources as well. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] transfer with hopper bookmark
Eric my interaction with the site says it is a Flash one. For me, w/o Flash, it is write only. More famously Apple mobile eschews Flash. Has any one succeeded with an Apple mobile? i was able to reverse engineer a page but have not been able to do anything with the PublicURL ... Its hash is about 6 characters longer than the (successful) one you got. So my question is how can i get to a workable hash from the reverse engineered one? Or maybe their Flash software goes through a submital process with .tl? In which case i would be locked out until they did an SVG implementation. greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:19 June 2012 14:37 subject: [Jbeta] transfer with hopper bookmark I took the following steps: used jt to package up ~addons/math misc folder dragged jattach.html and dropped it on my hopper web page got the hopper URL to the file in clipboard pasted the link here http://hop.tl/H70cbqe2g9nKV7_UQs In theory, you you can read this message on your iPad, tap the link, and continue the process to get ~addons/math misc folder installed. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] transfer with hopper bookmark
Eric my complaint about Gethopper is that for Flash disabled devices (such as iOS ones) it is a WRITE ONLY service. One CANNOT even get the URL a PASTE is written to w/o the Flash interface. Going a datum one way (From a Flash device) to a non flash device is fine. The other way runs into the lack of interface difficulty. Eg knowing what the link is that is created on an iOS device is impossible w/o allowing Flash. So it cannot be read even on a Flash enabled desktop, because no URL can be created. Presumably the URL shown in your email was created on a non-iOS device. What i am hoping for is that URL may be reconstructable even w/o the Flash interface... greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:20 June 2012 06:33 subject: Re: [Jbeta] transfer with hopper bookmark I am not following your comments. My use of hopper has been simple and it works for me. I'm not sure it is a useful thing at all and present it more as a curiousity. If you find it useful, great. If not, don't use it. In my example, hopper offers nothing over just including the file as an attachment in the email. I'm not sure if there are use cases where hopper provides interesting value. From my iPad, reading that email, and clicking the link worked fine, and there was nothing about flash. -- from:greg heil ghei...@gmail.com to: Beta forum beta@jsoftware.com date:19 June 2012 15:28 subject: Re: [Jbeta] transfer with hopper bookmark Eric my interaction with the site says it is a Flash one. For me, w/o Flash, it is write only. More famously Apple mobile eschews Flash. Has any one succeeded with an Apple mobile? i was able to reverse engineer a page but have not been able to do anything with the PublicURL ... Its hash is about 6 characters longer than the (successful) one you got. So my question is how can i get to a workable hash from the reverse engineered one? Or maybe their Flash software goes through a submital process with .tl? In which case i would be locked out until they did an SVG implementation. greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:19 June 2012 14:37 subject: [Jbeta] transfer with hopper bookmark I took the following steps: used jt to package up ~addons/math misc folder dragged jattach.html and dropped it on my hopper web page got the hopper URL to the file in clipboard pasted the link here http://hop.tl/H70cbqe2g9nKV7_UQs In theory, you you can read this message on your iPad, tap the link, and continue the process to get ~addons/math misc folder installed. -- For information about J forums see http://www.jsoftware.com/forums.htm
Re: [Jbeta] transfer with hopper bookmark
Eric i did get a Gethopper account. But lacking Flash their interface it does not show the widget for getting a URL for a pasted snippet. It is possible through their Bookmarklet interface for iOS devices, and other non-Flash users, to get full use of the service. i think Bookmarklets are only in JS. Something for me to try. Or perhaps someone else has? greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:20 June 2012 07:10 subject: Re: [Jbeta] transfer with hopper bookmark So hopper isn't that interesting for the iPAD (one way only). Good to know. I'm surprised as they certainly talk about iOS. Did you try www.gethopper.com in safari with a hopper account? Judging solely on my 1 message might not do it service. -- from:greg heil ghei...@gmail.com to: Beta forum beta@jsoftware.com date:20 June 2012 06:58 subject: Re: [Jbeta] transfer with hopper bookmark Eric my complaint about Gethopper is that for Flash disabled devices (such as iOS ones) it is a WRITE ONLY service. One CANNOT even get the URL a PASTE is written to w/o the Flash interface. Going a datum one way (From a Flash device) to a non flash device is fine. The other way runs into the lack of interface difficulty. Eg knowing what the link is that is created on an iOS device is impossible w/o allowing Flash. So it cannot be read even on a Flash enabled desktop, because no URL can be created. Presumably the URL shown in your email was created on a non-iOS device. What i am hoping for is that URL may be reconstructable even w/o the Flash interface... greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:20 June 2012 06:33 subject: Re: [Jbeta] transfer with hopper bookmark I am not following your comments. My use of hopper has been simple and it works for me. I'm not sure it is a useful thing at all and present it more as a curiousity. If you find it useful, great. If not, don't use it. In my example, hopper offers nothing over just including the file as an attachment in the email. I'm not sure if there are use cases where hopper provides interesting value. From my iPad, reading that email, and clicking the link worked fine, and there was nothing about flash. -- from:greg heil ghei...@gmail.com to: Beta forum beta@jsoftware.com date:19 June 2012 15:28 subject: Re: [Jbeta] transfer with hopper bookmark Eric my interaction with the site says it is a Flash one. For me, w/o Flash, it is write only. More famously Apple mobile eschews Flash. Has any one succeeded with an Apple mobile? i was able to reverse engineer a page but have not been able to do anything with the PublicURL ... Its hash is about 6 characters longer than the (successful) one you got. So my question is how can i get to a workable hash from the reverse engineered one? Or maybe their Flash software goes through a submital process with .tl? In which case i would be locked out until they did an SVG implementation. greg ~krsnadas.org -- from:Eric Iverson eric.b.iver...@gmail.com to: Beta forum beta@jsoftware.com date:19 June 2012 14:37 subject: [Jbeta] transfer with hopper bookmark I took the following steps: used jt to package up ~addons/math misc folder dragged jattach.html and dropped it on my hopper web page got the hopper URL to the file in clipboard pasted the link here http://hop.tl/H70cbqe2g9nKV7_UQs In theory, you you can read this message on your iPad, tap the link, and continue the process to get ~addons/math misc folder installed. -- For information about J forums see http://www.jsoftware.com/forums.htm