Numbered Bugs - 2/25/87 Category Page Count Card Bugs 2 4 Cross-file Link Bugs 4 2 NoteFile Bugs 6 1 Structure Bugs 7 1 Interface Bugs 8 5 Parameters Bugs 12 2 Converter Bugs 14 1 NCP (ProgInt) Bugs 15 6 I&R (Inspect&Repair) Bugs 19 1 Text Card Bugs 20 1 Browser Card Bugs 21 5 Document Card Bugs 24 1 Sketch Card Bugs 25 1 Graph Card Bugs 26 1 Miscellaneous Bugs 27 1 Lyric Bugs 29 2 Total 35 Card Bugs ! 320. Tang: Card and NoteFile protected operations need rethinking and fixing Date: 20 Nov 86 10:34 PST From: Tang.pa Subject: NoteCards 1.3k: Uncertainty about NC blocking To: NoteCardsSupport.pa cc: Tang.pa NoteCards 1.3k System Date: 14-Nov-86 22:33:44 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (204#110#) Microcode version: 26,32 Memory size: 13400 Frequency: Intermittent Impact: Minor I don't have a good formulation of this problem, but: With the new capability of having several NF open, there is a great temptation to do several different notefile operations at once, and I have occasionally noticed that breaks can result. For instance, I might Compact one NF while I'm closing another, or once I was Copying one NF while doing some work in another's browser, and I got a break "end of file something not copied". Another example, is that if you choose Close Cards off of the session icon menu and you select a list of cards to close, and while those cards are in the process of closing, you mouse a card icon to open a card, it lets you do that and opens the card, but halts the close card operation, and you have to go kill OLDMOUSE through the PSW and use Close Cards again to select the remaining unclosed cards. This used to work ok in 1.2, but doesn't in 1.3. These experiences have taught me simply not to try to do too many things at once. This is probably some tedious programming for this stage of the game, so you might just warn users NOT to simultaneously execute notecards operations. ~jt End of message 156.ml Trigg: Card NC.CloseNoteCards delete unused args Date: 2 Sep 86 16:26 PDT From: Trigg.pa Subject: NoteCards 1.3k: NC.CloseNoteCards takes two unused args To: NoteCardsSupport.pa cc: Trigg.pa Lisp System Date: 11-May-86 15:19:08 Machine: Dorado (MONK) Microcode version: 26,32 Memory size: 40000 NC.CloseNoteCards takes two args NoCheckFlg and Don'tClearFlg which are unused. I know it's a pain to find and fix the callers not to pass these args, but I think it's of some importance to clean up this sort of crud. - Randy End of message 400. PIrish: Card closing should do a FORCEOUTPUT 414. Gobbel: Card can get wedged by doing operation while Assign Title is in progress Date: 28 Jan 87 12:33 PST From: Gobbel.pa Subject: NoteCards 1.3k: Starting another op w/ Assign Title in progress loses To: NoteCardsSupport.pa cc: Gobbel.pa NoteCards 1.3k System Date: 6-Jan-87 16:42:38 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (DandeLisp) Microcode version: 26,32 Memory size: 13777 Frequency: Always Impact: Serious Scenario is: create a card, no title, select Close And Save. Assign Title window comes up. With this window up, select Delete Card. Assign Title window goes away, replaced w/ window that says "Can't Delete, Assign Title in progress". Card is now wedged, since there's no way to complete (or anything) the Assign Title. -Randy End of message Cross-file Link Bugs ! 345. Trigg: Cross-file Links weirdness Date: 19 Dec 86 15:27 PST From: Trigg.pa Subject: NoteCards 1.3k: Cross-file links weirdness To: NoteCardsSupport.pa, Tang cc: Trigg.pa NoteCards 1.3k System Date: 15-Dec-86 19:09:04 Lisp System Date: 11-May-86 15:19:08 Machine: Dorado (MONK) Microcode version: 26,32 Memory size: 40000 Frequency: Always Impact: Moderate This came up out of discussions with John Tang: suppose card A lives in one notefile, while filebox B lives in another notefile. Then, say you file A in box B. Because this is a cross-file link, a cross-file link card A1 is created in A's notefile with a FiledCard link pointing to A. (There's also a cross-filelink card in B's notefile, but that's irrelevant to this story.) Now, it turns out that if you close A without filing it anywhere else, NC won't object, because A has a filing link to it from within the notefile (from the crossfilelink card A1). Worse yet, if you delete the filing link from B to A, NC STILL won't object. This is because A1 still exists in A's notefile. We could check that filing links not be from crossfilelink cards before pronouncing the card properly filed. Other ideas? - Randy End of message 374. Kelley: Cross-file links halt Browser card Date: 8 Jan 87 23:52 PST From: Kelley.pa Subject: NoteCards 1.3k: Cross-file-links halt Browser card To: NoteCardsSupport.pa cc: Kelley.pa NoteCards 1.3k System Date: 12-Dec-86 18:19:40 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (222#72#) Microcode version: 26,32 Memory size: 15777 Frequency: >> Always, Intermittent, Once << Impact: >> Fatal, Serious, Moderate, Annoying, Minor << The Browser card graph terminates at cross-file-link cards even though it opens the linked NoteFile. It should continue graphing the structure. -- kirk End of message NoteFile Bugs ! 147.ml Kelley: NoteFile CopyCards should check index size before doing copy Print a message: Reset NewNoteFileInitialSize to larger than x. Do Compact. Reset NewNoteFileInitialSize. Re-do Copy Cards. Structure Bugs ! 185.hl Halasz: Structure Copy (NC.CopyCards) should assign a writelock when put down Date: 10 Sep 86 15:41 PDT From: Halasz.pa Subject: NoteCards 1.3k: Copy structure does not work in NEXT> To: NoteCardsSupport.pa cc: Halasz.pa Due to fact that copied cards do not own write locks when they are being put down. -- Frank End of message Interface Bugs ! 259. Gobbel: Interface shift-selecting something and then deleting it before finishing the shift-selection causes breaks. Date: 20 Oct 86 21:54 PDT From: Gobbel.pa Subject: Notecards 1.3k: dangling thingies To: NotecardsSupport.pa cc: Gobbel.pa It looks like there may be a rather pervasive problem with dangling references in the shift-select scheme for designating fileboxes, link destinations, etc. I tried selecting a filebox which I deleted before finishing the filing operation, and sure enough, it dies. Haven't tried other cases, but I'd be surprised if they weren't broken too. -Randy End of message ! 282. Tang: Interface problems with windows being pushed incorrectly in the horizontal direction Date: 5 Nov 86 12:30 PST From: Tang.pa Subject: NoteCards 1.3k: More detail comments on user interface To: NoteCardsSupport.pa cc: Tang.pa Lisp System Date: 11-May-86 15:19:08 Machine: Dove (204#110#) Microcode version: 26,32 Memory size: 13400 Frequency: Always Impact: Annoying More discoveries about the session icon. I tend to want to put it at the top of the screen roughly in the middle. I've noticed that when I do operations like Copy, with long file names, it not only pushes the icon down from the top (as I mentioned earlier), but also pushes it horizontally to the right (because Copy has 2 file names involved in the prompt window, it is a rather long prompt window message--longer than 1/2 the screen). Again, this tends to push the session icon underneath an active window, making it hard to find. Again, I could get used to it, and I learned that Randy avoids it by keeping the icon near the top and hugging the right side of the screen, and maybe I'm just too used to the 1.2 way when THE Prompt Window was used, but you asked for naive reactions to 1.3k. ~jt Date: 30 Jun 86 23:56 PDT From: Trigg.pa Subject: Re: NoteCards 1.3k: Session icon getting shoved down In-reply-to: Halasz.pa's message of 30 Jun 86 23:22 PDT To: Halasz.pa cc: Trigg.pa, NoteCardsSupport.pa My vote would be to change the code uniformly for all NC. It would push the session icon (or card or notefile icon) down after remembering the old position. Then, when the prompt window interaction is over, the window is moved back to its old position. - Randy End of message Date: 4 Sep 86 09:25 PDT From: Newman.pasa Subject: NoteCards 1.3k: message window moves main menu window To: NoteCardsSupport.pa NoteCards 1.3k System Date: 15-Jul-86 12:56:24 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (207#276#) Microcode version: 26,32 Memory size: 15777 Frequency: Once Impact: Minor While copying a notefile using the main menu copy notefile item, a prompt window was created that was so long, it could not fit on top of the menu window justified left or right without running off the screen. What happened was, the prompt window opened with its left edge at the left edge of the screen, and it moved the main menu window to attach to the prompt window's lower right corner. When the prompt window closed, the menu window stayed where it was. A minor problem, but somewhat disconcerting. >>Dave End of message Date: 31 Oct 86 17:44 PST From: Tang.pa Subject: NoteCards 1.3k: Naive reactions to 1.3k To: NoteCardsSupport.pa Lisp System Date: 29-Oct-86 22:35:30 Machine: Dorado (LeParc) Microcode version: 26,32 Memory size: 40000 Frequency: Always Impact: Annoying Randy asked for my initial reactions to 1.3k. One thing I noticed is that it's rather easy for me to lose the NoteCards main menu. I typically put the NC main menu at the very top of my screen, so I don't cover it up with any other windows or notecards. But, because the auxiliary message window gets occasionally attached to the NC main menu, it pushes the main menu down an arbitrary distance from the top of the screen. The way I use NC, it inevitably pushes it underneath a window, then I can't get at it later. It's harder for me to avoid overlaying the main menu when it's a couple inches from the top of the screen rather than at the very top. Perhaps the auxiliary message window could attach to the bottom or side of the menu instead of the top. Nit picky detail comments: I kinda miss having Close Cards (being a frequently used operation) that you could just click on the main menu (instead of now having to make a selection off the main menu, then drag over to make an additional selection) I rather dislike having to hold down a key to make the multiple card selections (like in Close Cards), but I guess it provides functionality that I haven't learned to appreciate yet. ~jt End of message ! 234.hl Halasz: Interface select cards should just look for mouse up instead of both up. Date: 12 Sep 86 10:47 PDT From: Newman.pasa Subject: NoteCards 1.3k: When asked to select a file box (filing a card), pushing "SHIFT" too early causes problems To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 15-Jul-86 12:56:24 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#53#) Microcode version: 26,32 Memory size: 16400 Frequency: Always Impact: Annoying If you are filing a card, and are to the point where you are asked to select the file box, push the shift key down, left-click in the body of a filebox to bring its title to the top, then left-click in the title bar just brought to the top in order to push that file box into the prompt window. The selection will not go into the box, and one must release the shift key, put it down again, and then select the title bar. The only real problem that I have with this is that it is natural to expect that it will work given the way that similar actions work when shift-selecting from one TEdit window to another. I guess this is just another consistency-of-interfaces issue. >>Dave End of message Date: 19 Sep 86 23:03 PDT From: Kelley.pa Subject: Re: NoteCards 1.3k: When asked to select a file box (filing a card), pushing "SHIFT" too early causes problems In-reply-to: Newman.pasa's message of 12 Sep 86 10:47 PDT To: Newman.pasa cc: NoteCardsSupport.pa Yes this is a problem. In fact, attempting to continue to select in the title or down in the body with the SHIFT key down will cause NC.CopyButtonEventFn to break "TRYING to \SETUPGETCH BEYOND END OF TEXT". Simple cross-TEdit shift selects do seem to work correctly (i.e. without having to let up on the SHIFT key after bringing the card totop). -- kirk End of message 378. Newman: Interface - would like ability to specify font for all menus 439. Kelley: Interface - Null String titles don't display with small box in Link icons It is possible programatically to actually create a title with an empty string. This is the source of some bugs. The link image boxer does not work. The space for the link is there, but it does not have an empty box indicating it. This is a new bug maybe introduced by Peggy's new stuff. Parameters Bugs 367. PIrish: Parameters - Inconsistent initializing of parameters from init file Date: 7 Jan 87 12:29 PST From: PIrish.pa Subject: NoteCards 1.3k: Inconsistent initializing of parameters from init file To: NoteCardsSupport.pa cc: PIrish.pa NoteCards 1.3k System Date: 6-Jan-87 16:42:38 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (Leprechaun) Microcode version: 0,0 Memory size: 13400 Frequency: Once Impact: Minor I modified my initialization of NoteCards parameters in my init file, and then loaded a new sysout and loadup. Most of the parameters were set correctly, but two were not: MarkersInFileBoxes and DefaultLinkIconAttachBitmap. I had both initialized to No in my init file, and they ended up being Yes. I noticed when modifying these parameters that some were being set with T and NIL, and I changed them to Yes and No. Could this be related to the problem? -Peggy End of message 371. Kelley: Parameter MenuLingersAfterNoteFileClose has bug when set to No Date: 7 Jan 87 16:32 PST From: Kelley.pa Subject: NoteCards 1.3k: MenuLingersAfterNoteFileClose No bug To: NoteCardsSupport.pa cc: Kelley.pa NoteCards 1.3k System Date: 12-Dec-86 18:19:40 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (222#72#) Microcode version: 26,32 Memory size: 15777 Frequency: >> Always, Intermittent, Once << Impact: Annoying When MenuLingersAfterNoteFileClose is set to No (Reason its set to No: I got in big trouble editing an old version of a compacted file.) the "Menu" gets closed before the final status information is sent to it causeing funny writing on the region where it used to be. Especially a problem if the status information does not fit and a scroll signal from the user is needed. -- kirk End of message Converter Bugs 443. Marshall: Converter breaks on complicated sketch card Date: 23 Feb 87 16:40 PST From: Marshall.pasa Subject: NoteCards 1.3k: Another conversion snag To: NoteCardsSupport.pa cc: Marshall.pasa NoteCards 1.3k System Date: 14-Oct-86 12:32:25 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#52#) Microcode version: 26,32 Memory size: 16400 Frequency: >> Always, Intermittent, Once << Impact: >> Fatal, Serious, Moderate, Annoying, Minor << Same notefile as before: This time, a hash array overflowed in the process of NC.PutSketchSubstance-ing. Stack looks something like this: . . HASHARRAY HASHOVERFLOW PUTHASH HPRINT1 .(bunch more HPRINT1s here) . HPRINT1 NC.PutSketchSubstance NC.PutMainCardData . . and so forth I remember what this sketch looks like; it's big, contains no links but quite a number of other sketchobjs. Not sure about the cause, but it's probably not the sketch's fault (It looks so innocent...). I'm bailing out of the card substance and going on, but I think this might be symptomatic of a problem with converting complicated sketches. Cathy End of message NCP (ProgInt) Bugs 128.hm Hogg: NCP for SelectCards should not overwrite non-notecard window. Date: 12 Aug 86 16:18 PDT From: hogg.pa Subject: NoteCards 1.3k: NCP.SelectCards overwrites non-notecard window To: NoteCardsSupport.pa cc: hogg.pa Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (204#54#) Microcode version: 26,32 Memory size: 13777 Calling (NCP.SelectCards W) where W is a non-notecards window (e.g. the main exec, a normal TEdit or some random menu other than the session & file menus) causes a prompt window & the buttons to appear as usual but all messages (e.g. the names of the selected cards) go into W, instead of the prompt window! Moreover, the window W is also cleared first. Since the prompt window correctly appears (& the "target" cursor is in the prompt window as expected), I also expect the usual messages to go into the prompt window as well. Tad End of message 183.hu PIrish: NCP display of newly created card should not affect initial text. Date: 10 Sep 86 14:24 PDT From: PIrish.pa Subject: NoteCards 1.3k: Problem with NCP.LocalGlobalLink? To: NoteCardsSupport.pa cc: PIrish.pa Lisp System Date: 11-May-86 15:19:08 Machine: Dove (Leprechaun) Microcode version: 26,32 Memory size: 13400 Frequency: Always Impact: Serious Scenario, using Prog. Int. functions: - Create a notefile and open it. - Create some cards for this notefile. These cards have types that I have define. They are "form" types: text cards with some initial text. - Go through the cards and add links between them. - Checkpoint the file. Results: If I have the cards display when created, then all works correctly. If I don't have the cards display when created, then when I open the cards later I discover that the initial text has disappeared, but the links are still there. I have traced a few things, and my makefns appear to be getting called correctly, and the cards are initially created with text. But then, my debugging consisted of breaking my makefns and then evaling them with the nodisplayflag set to NIL so that I could see the cards being created. Why should the display of a card upon creation have anything to do with the initial text being put in the card, or staying put in the card? Or could it have something to so with adding the links? I am lost! Peggy End of message 184.ml PIrish: NCP.LocalGlobalLink using wrong display mode default. 342. Hogg: NCP.SelectCards error when InstigatingCard not open Date: 18 Dec 86 08:57 PST From: Hogg.pa Subject: NoteCards 1.3k: NCP.SelectCards error when InstigatingCard not open To: NoteCardsSupport.pa cc: Hogg.pa NoteCards 1.3k System Date: 11-Dec-86 19:23:11 Lisp System Date: 11-Dec-86 19:37:41 Machine: Dorado (Ventana) Microcode version: 26,32 Memory size: 40000 Frequency: Always (NCP.SelectCards card ...) breaks with ILLEGAL ARG NIL in ATTACHWINDOW if card is not open on the screen. I think either the card should be activated so that SelectCards has a window to use, or it should just use the default prompting (as if the card arg had been NIL). End of message 344. WJohnson: NCP.NoteFileFromFileName has problem Date: 19 Dec 86 15:03 PST From: Trigg.pa Subject: [wjohnson.pa: NCP.NoteFileFromFileName] To: Notecardssupport cc: Trigg.pa FYI: ----- Begin Forwarded Messages ----- Date: 19 Dec 86 12:25 PST From: wjohnson.pa Subject: NCP.NoteFileFromFileName To: Trigg Reply-to: wjohnson.pa I could never get it to work, Randy. Interestingly, here are the 2 lines of code that worked: (NCP.OpenNoteFile (QUOTE {DSK}VOYAGER>VOYAGER-RULES.NOTEFILE)) (SETQ INTERFACE.NOTEFILE (NCP.NoteFileFromFileName (QUOTE {DSK}VOYAGER>VOYAGER-RULES.NOTEFILE))) Here are the two lines of code that don't work (the next two lines in the function, in fact): (NCP.OpenNoteFile (QUOTE {DSK}VOYAGER>VOYAGER-KNOWLEDGE.NOTEFILE)) (SETQ KNOWLEDGE.NOTEFILE (NCP.NoteFileFromFileName (QUOTE {DSK}VOYAGER>VOYAGER-KNOWLEDGE.NOTEFILE))) I solved the problem for now by naming KNOWLEDGE.NOTEFILE to the value returned from opening the notefile; of course that won't work if the notefile is already open. Asides: 1] the FULLNAME of my file EQ'd the (fetch FullFileName of my notefile), So that isn't the problem. 2] Once I set KNOWLEDGE.NOTEFILE to (NCP.OpenNoteFile (QUOTE {DSK}VOYAGER>VOYAGER-KNOWLEDGE.NOTEFILE)), the above code that didn't work began working. Walt ----- Next Message ----- Date: 19 Dec 86 15:02 PST From: Trigg.pa Subject: Re: NCP.NoteFileFromFileName In-reply-to: wjohnson.pa's message of 19 Dec 86 12:25 PST To: wjohnson.pa cc: Trigg.pa Walt, it just occurred to me looking at your code that you should replace the lines: (NCP.OpenNoteFile (QUOTE {DSK}VOYAGER>VOYAGER-RULES.NOTEFILE)) (SETQ INTERFACE.NOTEFILE (NCP.NoteFileFromFileName (QUOTE {DSK}VOYAGER>VOYAGER-RULES.NOTEFILE))) with the expression: (SETQ INTERFACE.NOTEFILE (NCP.OpenNoteFile (QUOTE {DSK}VOYAGER>VOYAGER-RULES.NOTEFILE))) since NCP.OpenNoteFile returns the notefile object if successful. That still doesn't explain the problem with NCP.NoteFileFromFileName, of course. - Randy ----- End Forwarded Messages ----- End of message 389. Dixon: NCP - orphans not always recognized Date: 6 Jan 87 23:48 PST From: MikeDixon.pa Subject: orphans not always recognized To: NoteCardsSupport i have a program which creates a number of fileboxes and cards in them, all with nodisplayflg=T. i've noticed that if i then delete one of these fileboxes, it doesn't notice that the cards in it are now orphans. they seem to be still in the notefile (NCP.AllCards finds them) but as far as i can tell have no parents. this is a bug, no? .mike. End of message I&R (Inspect&Repair) Bugs 375. Kelley: I&R abort does not close notefile Date: 9 Jan 87 13:17 PST From: kelley.pa Subject: NoteCards 1.3L: Abort in I&R does not close notefile To: NoteCardsSupport.pa cc: kelley.pa NoteCards 1.3L System Date: 23-Dec-86 10:09:31 Lisp System Date: 7-Jan-87 03:56:58 Machine: Dandelion (222#146#) Microcode version: 0,231 Memory size: 15777 Frequency: >> Always, Intermittent, Once << Impact: >> Fatal, Serious, Moderate, Annoying, Minor << I think this is a Koto problem too. If the I&R process aborts due to some error, it does not close the notefile. -- kirk End of message Text Card Bugs 206.lh Newman: Text cards with untitled links should not TOTOP when changing title. TEDIT redisplays when object has not changed. Must check display mode first. Maintain window stacking order? Date: 16 Sep 86 09:45 PDT From: Newman.pasa Subject: NoteCards 1.3k: Untitled link icons still come to top when title of the card is changed To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 15-Jul-86 12:56:24 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#53#) Microcode version: 26,32 Memory size: 16400 Frequency: Always Impact: Minor When changing the title of a card, and a link icon representing a link that points to that card is on the screen, the card containing that icon must come to the top to update the title in the link icon. The problem is that when the link icon does not contain a title, the card containing it comes to the top anyway. >>Dave End of message Browser Card Bugs 384. Newman: Browser - Want browsers to avoid accumulation of extraneous material ��Apparently when a browser is recomputed many times, it appears to cause a garbage collect and t�hen crashes the system. 306. Newman: Browser having problems removing nodes which are the last links to virtual nodes - Disconnected Graph error Date: 14 Nov 86 15:34 PST From: Newman.pasa Subject: NoteCards 1.3k: Disconnected Graph error To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 14-Oct-86 12:32:25 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#53#) Microcode version: 26,32 Memory size: 16400 Frequency: Intermittent Impact: I built a browser. I removed a node from the browser because the browser was too complex, and I asked the system to relayout the graph. The system broke with a "disconnected graph" error. In this case (I know I have reported this before, but I don't think it was quite the same). There were a number of virtual nodes connected to the node I removed. They are not linked to any other part of the graph now that I have removed the unwanted node. They are, I think, the disconnected part of the graph. Also, I cannot remove the virtual nodes without removing all instances of that link from the graph. I used "Reconnect Nodes" from the menu after up-arrowing out of the break, and it seemed to fix the problem. >>Dave End of message 112. Cole: Browser limit should be detected before Put time From: cole.pa Subject: 1.2k: 455 Cards Breaks Browser To: NoteCardsSupport.pa cc: cole.pa Format: TEdit 1.2k System Date: 19-Feb-86 12:37:45 Lisp System Date: 30-May-86 11:49:05 Machine: Dandelion (204#103#) Microcode version: 26,32 Memory size: 13777 Frequency: Once out of once tried Impact: Fatal I suspected there might be troubles with large enough browsers.... After running Inspect & Repair (which DID cure previous complaints, so obviously a link had to be re-written), I opened the now "ok" notefile. I selected no link dashing, no arrowheads, in global parameters; gave the root to the browser, asked for every linktype -> direction except browser contents, set the depth to infinity, orientation to horizontal, and let fly. Ten minutes later (with .87 of Data, .81 of Dsk space from VSTATS) I got a browser. The original overview window was about 1/5 black. The expanded overview window (70-80 percent of the screen, filling most of the middle) was still not large enough to have more than faint grey lines marking the cards over heavy black lines marking the links. The shape of the bulk of the notefile was revealed. That done, I opted to close the notefile. This is where the break developed -- closing the browser broke the system. Here is a copy of the break: 37.lm Kurt: Browser hardcopy should not need Terminal 8 REgular Interpress font to print 432. Newman: Browser Expand Node destroys existing links Date: 10 Feb 87 17:45 PST From: Newman.pasa Subject: NoteCards 1.3k: Expand Browser Node destroys existing links To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 14-Oct-86 12:32:25 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#53#) Microcode version: 26,32 Memory size: 16400 Frequency: Once Impact: Serious Open the Geographic Information Browser in the DEMO notefile. Expand the Caribbean Area Map node. Observe that all the existing links drawn in the browser dissappear, and note that only the links from the Caribbean Area Map link icon to the other icons appear in the browser. This is bad. >>Dave P.S. Note that I am in the new sysout now, the SYSOUTDATE is 28-Jan-87 though the NoteCards system date still says 14-October-86. End of message Document Card Bugs 365. PIrish: Document Cards cannot follow cross-file links Date: 6 Jan 87 16:06 PST From: Gobbel.pa Subject: NoteCards 1.3k: Document Cards cannot follow cross-file links To: NoteCardsSupport.pa cc: Gobbel.pa NoteCards 1.3k System Date: 15-Dec-86 19:09:04 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (DandeLisp) Microcode version: 26,32 Memory size: 13777 Frequency: Always Impact: Annoying Document Card cannot follow cross-file links because it does not know about that card type. It should probably be updated to know how to handle them. -Peggy End of message Sketch Card Bugs 417. Newman: Sketch menus are still reversed Date: 29 Jan 87 15:43 PST From: Newman.pasa Subject: NoteCards 1.3k: A minor interface bug (already reported once) To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 14-Oct-86 12:32:25 Lisp System Date: 11-May-86 15:19:08 Machine: Dandelion (207#65#) Microcode version: 26,32 Memory size: 15777 Frequency: Always Impact: Minor The two menus used on sketch cards are still reversed. (I reported this once before) When I open a sketch card, the menu with the "Fix Menu" item is already fixed at the side of the card, and the menu without the "Fix Menu" item is brought up by middle-buttoning on the title bar. I thought that this was fixed already. >>Dave P.S. Did we ever figure out the confusion with my NoteCards systemdate? I note that it is wrong above, as I just made this sysout, and the sysoutdate is 27-Jan-87 End of message Graph Card Bugs 202.mh Newman: Grapher should not gratuitously place (GACHA 8 MRR) in substance Find some way to finesse this. Date: 15 Sep 86 14:17 PDT From: Newman.pasa Subject: NoteCards 1.3k: Browser break with FONT NOT FOUND on standalone machine To: NoteCardsSupport.pa cc: Newman.pasa NoteCards 1.3k System Date: 15-Jul-86 12:56:24 Lisp System Date: 11-May-86 15:19:08 Machine: Dove (207#53#) Microcode version: 26,32 Memory size: 16400 Frequency: >> Always, Intermittent, Once << Impact: >> Fatal, Serious, Moderate, Annoying, Minor << A Customer reported this problem. A browser created in a notefile on a machine on a network breaks when being brought up on a different machine that is standalone. The font not found break is for GACHA 8 MRR in a PRESS device. I hypothesize that the browser was hardcopied on the networked machine causing a FONTCREATE expression for PRESS to be placed on the BROWSER substance. Hence, when the standalone machine tries to bring the card up, it gets the break. >>Dave End of message Miscellaneous Bugs 377. Kelley: Miscellaneous - bad substance tale Date: 12 Jan 87 11:41 PST From: kelley.pa Subject: NoteCards 1.3L: bad substance tale. To: NoteCardsSupport.pa cc: kelley.pa NoteCards 1.3L System Date: 23-Dec-86 10:09:31 Lisp System Date: 7-Jan-87 03:56:58 Machine: Dandelion (222#146#) Microcode version: 0,231 Memory size: 15777 Frequency: >> Always, Intermittent, Once << Impact: >> Fatal, Serious, Moderate, Annoying, Minor << I'm doing a copy structure from one notefile to another. The Text card being copied is open on the screen, has a bunch of text and ended in four links to SEdit cards. The first of the four links was deleted right before the copy stucture by means of Delete Card with UseDeletedLinkIconIndicators nil. I suspect that is the culprit but this has worked in the past. NC.GetMainCardData is now reading a pointer for the length of the substance from the Text card stream that is inconsistent with the stream because when it calls NC.GetTextSubstance with that length, it COPYBYTES a slightly truncated substance into its temporary local corefile stream. This causes the subsequent opentextstream to break because tedit.build.pctb reads an end of file trying to get charlooks.list. After closeing the NoteFile Inspect and Repair w/o substances reported no errors, but in rebuilding links I&R broke with the same problem. I uparrowed out, closed the NoteFile, and did an I&R with substances. It found the bad substance, I backed it and the associated props up to the previous good version and proceeded. During the links rebuilding phase, 5 cards needed to be filed in to-be-filed. While making Child link I was asked if I wanted to follow a cross-file link. I said yes. It opened the NoteFile, could not find the card (valid failure), closed the NoteFile without killing NC.CacheTypesAndTitles so NC.CacheTypesAndTitles process broke. At the same time I was asked if I wanted to try another file. I said no. NC.InsertLinkInText then as a result broke trying to NC.MakeLink with a NIL CardObject. I did a (RETFROM 'NC.InsertLinkInText) and successfully finished the I&R. Upon Opening the NoteFile, I can now not bring up the original card as it dies on the same error. Could the substance for this card somehow have a length pointer for the new substance but substance start pointer for the old substance? Another I&R with substances revealed that my switch to the old version did not take. Furthermore, the links in inspecting the substance of the old version came up as delete-me icons with an error message for each I could not read because of scrolling off the tty window. After backing up the version again and successfully completing I&R, upon re-opening the file, the substance still had the same problem. Thats enough for now... Summary of bugs: 1. Delete Card with UseDeletedLinkIconIndicators nil does not always update the stream correctly. 2. I&R / cross-file link close of file should first kill NC.CacheTypesAndTitles process. 3. NIL return to query for name of new NoteFile to try in bad cross-file link should behave as if No new notefile were requested to be tried. 4. Backing up a substance in I&R should work and/or fix up the substance length pointers to a consistent state. 5. Inspecting substance position region prompt message should not scroll the bad links message off the TTY window before the user can read them. -- kirk End of message Lyric Bugs 398. Lyric - FINDFILE in Lyric does not use FILEDATES property Date: 18 Jan 87 20:37 PST From: kelley.pa Subject: NoteCards 1.3L: FINDFILE in Lyric does not use FILEDATES property To: NoteCardsSupport.pa cc: kelley.pa NoteCards 1.3L System Date: 23-Dec-86 10:09:31 Lisp System Date: 7-Jan-87 03:56:58 Machine: Dandelion (222#146#) Microcode version: 0,231 Memory size: 15777 Frequency: Always Impact: ED cannot load definitions It loks like SPELLFILE in Lyric will no longer use the FILEDATES property. Good news: we can get rid of the hack in NOTECARDSCOMS that changes filedates. Bad news: need to append the NoteCards release source directory to DIRECTORIES at NoteCards load time if (NOT (EQ NC.LOADFLG 'SYSLOAD)). -- kirk End of message 395. Lyric - I&R should deal with character class changes Date: 16 Jan 87 18:15 PST From: Kelley.pa Subject: I&R should deal with character class changes To: NoteCardsSupport cc: Kelley.pa Whenever TEDIT.PUT.PCTB sees a byte containing 255 [presumably followed by a non-zero byte], it prepends a 255 0 on the front of the destination stream. This has happened to me several times, though it seems to work ok most of the time even with extended characters in the substance. I have talked to John S about this and TEdit will do this for any kind of stream including binary streams. Since NoteFiles are extremely fragile to this kind of breakage, I&R, NC.OpenNoteFile actually, should make a special case to look for the 255 0 bytes at the beginning of the NoteFile if a version mismatch error occurs, and get rid of them rather than first asking the user if they want to convert verisons. -- kirk End of message(LIST ((PAGE NIL (PAPERSIZE Letter FOLIOINFO (ARABIC) STARTINGPAGE# 1) (0 0 612 792) ((FOLIO NIL (PARALOOKS (QUAD CENTERED) CHARLOOKS (SUPERSCRIPT 0 INVISIBLE OFF SELECTPOINT OFF PROTECTED OFF SIZE 10 FAMILY TIMESROMAN OVERLINE OFF STRIKEOUT OFF UNDERLINE OFF EXPANSION REGULAR SLOPE REGULAR WEIGHT BOLD INVERTED OFF USERINFO NIL STYLE NIL) FORMATINFO (ARABIC)) (162 36 288 36) NIL) (TEXT NIL NIL (72 72 468 648) NIL))) (PAGE NIL (PAPERSIZE Letter FOLIOINFO (ARABIC)) (0 0 612 792) ((FOLIO NIL (PARALOOKS (QUAD CENTERED) CHARLOOKS (SUPERSCRIPT 0 INVISIBLE OFF SELECTPOINT OFF PROTECTED OFF SIZE 10 FAMILY TIMESROMAN OVERLINE OFF STRIKEOUT OFF UNDERLINE OFF EXPANSION REGULAR SLOPE REGULAR WEIGHT BOLD INVERTED OFF USERINFO NIL STYLE NIL) FORMATINFO (ARABIC)) (162 36 288 36) NIL) (TEXT NIL NIL (72 72 468 648) NIL))) (PAGE NIL (PAPERSIZE Letter FOLIOINFO (ARABIC)) (0 0 612 792) ((FOLIO NIL (PARALOOKS (QUAD CENTERED) CHARLOOKS (SUPERSCRIPT 0 INVISIBLE OFF SELECTPOINT OFF PROTECTED OFF SIZE 10 FAMILY TIMESROMAN OVERLINE OFF STRIKEOUT OFF UNDERLINE OFF EXPANSION REGULAR SLOPE REGULAR WEIGHT BOLD INVERTED OFF USERINFO NIL STYLE NIL) FORMATINFO (ARABIC)) (162 36 288 36) NIL) (TEXT NIL NIL (72 72 468 648) NIL))))) 100 \�100 h�100 P�.HH�(()T(( TIMESROMAN� TIMESROMAN TIMESROMAN TIMESROMAN TIMESROMAN TIMESROMAN TIMESROMAN N 7 / % 6 | � � 7 @ % � 1U O / % C ( 4 / % | 5 { / < / % , 8 � M V > S { + X a @ % = 8 " � > / % � - 1 % � � � Y s / % � � w : � � IV &P P / % . � K > / % , v : 1 / % , 8 � Y J G % b � P : % & ! � 3 F � � D> M / % 3 1 * " _ N p ^ R t � e y 1 & : U N p h B Q ( / ' T 9 . > / % , 8 r T o ` / % = Q �x 2 / % � � m ; ' % % " C � � � � Y8 E / % > $ � � : G / % � , G / % ' � P P / % , 8 " � / - / % , 8 q � E > � � c Z � q � > K / % # K P � 9 6 � 3 ��lz�