From: Peter G. <pe...@ar...> - 2004-09-08 02:13:30
|
A new development snapshot (j 0.20.2.20, ABCL 0.0.3.18) is available: http://armedbear.org/j.zip (source) http://armedbear.org/j-jar.zip (just j.jar) For the adventurous few who build j from source, this snapshot features the bare beginnings of a port of slime to j. Slime is the "Superior Lisp Interaction Mode for Emacs", developed by Eric Marsden, Luke Gorrie, Helmut Eller and others. Those guys did all the hard work; I'm just cargo-culting. Note that slime-for-j won't work at all with this snapshot if you just download j-jar.zip and do "java -jar j.jar"; you must download j.zip (or use anonymous CVS) and build j from source. (This is not by design; I only discovered it myself after uploading the snapshot.) To start slime, do Alt-X, "slime", and wait for the "Slime connected!" message to appear in the status bar. Currently slime-for-j implements three things that work more or less correctly: 1. Arglists. Type, say, "(car " at the repl, and the signature of the function will appear in the status bar: "(car list)". This works by mapping the space key to (slime:slime-space) in Lisp source buffers and Lisp shells. For various reasons, not all functions come with arglist information (for example, unresolved autoloads don't), so don't be alarmed if nothing appears in the status bar. 2. Symbol completion. Type "*d-p-d" (or "f--e") at the repl and hit Tab. Less exotic completions work too: type "lam" and then keep hitting Tab to cycle through the possibilities. Unlike slime-for-emacs, the current incarnation of slime-for-j does not provide a choosable list when there are multiple possible completions; you need to keep hitting Tab to see them all. Symbol completion is mapped to Tab in the repl; it's available in Lisp source buffers too, but there it's mapped to Ctrl Alt I (since Tab is reserved for indenting). 3. EDIT-DEFINITION Slime maps Alt . to (slime:edit-definition) in the repl and in Lisp source buffers. EDIT-DEFINITION queries the superior Lisp for the source location of the function corresponding to the symbol at the location of the caret and jumps there if possible. (If the superior Lisp provides no useful information, EDIT-DEFINITION falls back to findTagAtDot.) Plus a couple of things that don't work quite right yet: 1. (slime:eval-region), mapped by default to Ctrl Alt R in Lisp source buffers. 2. (slime:eval-last-expression), mapped by default to Ctrl Alt E in Lisp source buffers. These commands evaluate a Lisp expression and print the result in the status bar. In their current state, these commands shouldn't cause any terrible trouble, but output sent to the repl buffer as a side effect may end up oddly placed. If you do simply Alt X, "slime", you'll get abcl (as a separate instance) in a slime buffer. If you have sbcl installed, and if you build j with the --enable-jpty configure option, you can do Alt X, "slime sbcl", to run sbcl in a j slime buffer (assuming sbcl is in your path). Currently sbcl is the only other Lisp that's supported; it seemed like the easiest one to start with since its threading model and socket support are very similar to abcl's. On another subject, I'd like to do a new official release of j by the end of the month. The last official release was on October 23, 2003, and it doesn't work well with Java 1.5 (which is also due to be released at the end of the month). Please remind me if there's anything that (realistically) should be done for this release. On my own hook, I'm not planning on anything much beyond further incremental improvements to slime and abcl, but if there are any outstanding editor bugs I'd like to fix them. Thanks for your support. -Peter |
From: Brad B. <br...@gl...> - 2004-09-08 13:28:11
|
Thanks for the note. Regarding your comment about editor bugs, I have a couple: 1) Enable multiline patterns in Replace in Files; 2) an incredibly hard-to-nail-down bug where j stops reponding entirely except for switching buffers. For #1, I have experimental features turned on in my prefs. Not sure, but I had thought multiline patterns were enabled at some point in the past. For #2, here's what happens ... sometimes: * Mode: xml / xsl * Java: both blackdown jdk 1.4.1 and sun jdk 1.4.2.05 * I use F11 a lot to comment out single lines. * Sometimes, when I hit shift-F11 to uncomment the line, j will just hang there. I can switch buffers using the buffer list, the entire j window redraws, I can select commands from the menus, but no commands (including the ones selected from the menus) execute, and I can't move the caret. * Force-quitting j and then restarting it is the only option, at which point it usually asks to recover a saved file. The recovery shows that the shift-F11 command completed successfully. I have been completely unable to determine the exact sequence of steps to get #2 to occur. I was just trying it again, and j worked perfectly. Sorry, but that's all I have. On Tue, 2004-09-07 at 19:13, Peter Graves wrote: > A new development snapshot (j 0.20.2.20, ABCL 0.0.3.18) is available: > > http://armedbear.org/j.zip (source) > http://armedbear.org/j-jar.zip (just j.jar) > > For the adventurous few who build j from source, this snapshot features > the bare beginnings of a port of slime to j. Slime is the "Superior > Lisp Interaction Mode for Emacs", developed by Eric Marsden, Luke > Gorrie, Helmut Eller and others. Those guys did all the hard work; I'm > just cargo-culting. > > Note that slime-for-j won't work at all with this snapshot if you just > download j-jar.zip and do "java -jar j.jar"; you must download j.zip > (or use anonymous CVS) and build j from source. (This is not by design; > I only discovered it myself after uploading the snapshot.) > > To start slime, do Alt-X, "slime", and wait for the "Slime connected!" > message to appear in the status bar. > > Currently slime-for-j implements three things that work more or less > correctly: > > 1. Arglists. > > Type, say, "(car " at the repl, and the signature of the function > will appear in the status bar: "(car list)". > > This works by mapping the space key to (slime:slime-space) in Lisp > source buffers and Lisp shells. > > For various reasons, not all functions come with arglist > information (for example, unresolved autoloads don't), so don't be > alarmed if nothing appears in the status bar. > > 2. Symbol completion. > > Type "*d-p-d" (or "f--e") at the repl and hit Tab. Less exotic > completions work too: type "lam" and then keep hitting Tab to cycle > through the possibilities. > > Unlike slime-for-emacs, the current incarnation of slime-for-j > does not provide a choosable list when there are multiple possible > completions; you need to keep hitting Tab to see them all. > > Symbol completion is mapped to Tab in the repl; it's available in > Lisp source buffers too, but there it's mapped to Ctrl Alt I (since > Tab is reserved for indenting). > > 3. EDIT-DEFINITION > > Slime maps Alt . to (slime:edit-definition) in the repl and in Lisp > source buffers. EDIT-DEFINITION queries the superior Lisp for the > source location of the function corresponding to the symbol at the > location of the caret and jumps there if possible. (If the superior > Lisp provides no useful information, EDIT-DEFINITION falls back to > findTagAtDot.) > > Plus a couple of things that don't work quite right yet: > > 1. (slime:eval-region), mapped by default to Ctrl Alt R in Lisp > source buffers. > > 2. (slime:eval-last-expression), mapped by default to Ctrl Alt E in > Lisp source buffers. > > These commands evaluate a Lisp expression and print the result in the > status bar. > > In their current state, these commands shouldn't cause any terrible > trouble, but output sent to the repl buffer as a side effect may end up > oddly placed. > > If you do simply Alt X, "slime", you'll get abcl (as a separate > instance) in a slime buffer. > > If you have sbcl installed, and if you build j with the --enable-jpty > configure option, you can do Alt X, "slime sbcl", to run sbcl in a j > slime buffer (assuming sbcl is in your path). Currently sbcl is the > only other Lisp that's supported; it seemed like the easiest one to > start with since its threading model and socket support are very > similar to abcl's. > > On another subject, I'd like to do a new official release of j by the > end of the month. The last official release was on October 23, 2003, > and it doesn't work well with Java 1.5 (which is also due to be > released at the end of the month). > > Please remind me if there's anything that (realistically) should be > done for this release. On my own hook, I'm not planning on anything > much beyond further incremental improvements to slime and abcl, but if > there are any outstanding editor bugs I'd like to fix them. > > Thanks for your support. > > -Peter > > > ------------------------------------------------------- > This SF.Net email is sponsored by BEA Weblogic Workshop > FREE Java Enterprise J2EE developer tools! > Get your free copy of BEA WebLogic Workshop 8.1 today. > http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click > _______________________________________________ > armedbear-j-users mailing list > arm...@li... > https://lists.sourceforge.net/lists/listinfo/armedbear-j-users _____________________________ Brad Brizendine, CTO, Glyphix |
From: Peter G. <pe...@ar...> - 2004-09-09 19:09:08
|
On Wed, 08 Sep 2004 at 06:29:21 -0700, Brad Brizendine wrote: > Thanks for the note. Regarding your comment about editor bugs, I have a > couple: 1) Enable multiline patterns in Replace in Files; 2) an > incredibly hard-to-nail-down bug where j stops reponding entirely except > for switching buffers. > > For #1, I have experimental features turned on in my prefs. Not sure, > but I had thought multiline patterns were enabled at some point in the > past. There doesn't seem to be any code to support multiline patterns in replaceInFiles, so unless I'm missing something, there's nothing to enable. > For #2, here's what happens ... sometimes: > > * Mode: xml / xsl > * Java: both blackdown jdk 1.4.1 and sun jdk 1.4.2.05 > * I use F11 a lot to comment out single lines. > * Sometimes, when I hit shift-F11 to uncomment the line, j will > just hang there. I can switch buffers using the buffer list, the > entire j window redraws, I can select commands from the menus, > but no commands (including the ones selected from the menus) > execute, and I can't move the caret. > * Force-quitting j and then restarting it is the only option, at > which point it usually asks to recover a saved file. The > recovery shows that the shift-F11 command completed > successfully. > > > I have been completely unable to determine the exact sequence of steps > to get #2 to occur. I was just trying it again, and j worked perfectly. > Sorry, but that's all I have. The next time this problem occurs, if you hit Ctrl \ in the xterm from which you've started j (assuming such a thing exists), you should get a full thread dump, including the JVM's best guess about any possible deadlock that might be happening. This might help narrow things down enough to find the bug. Thanks! -Peter |
From: Brad B. <br...@gl...> - 2004-09-09 19:24:35
|
Thanks for the response. Regarding multiline patterns, I should have been more clear: it's not a bug, it's just a feature I'd really like to have. And thanks for the note on how to start diagnosing my hang issue. I'll watch for it and let you know what I find. I appreciate all the work you're doing on j. Thanks! On Thu, 2004-09-09 at 12:08, Peter Graves wrote: > On Wed, 08 Sep 2004 at 06:29:21 -0700, Brad Brizendine wrote: > > Thanks for the note. Regarding your comment about editor bugs, I have a > > couple: 1) Enable multiline patterns in Replace in Files; 2) an > > incredibly hard-to-nail-down bug where j stops reponding entirely except > > for switching buffers. > > > > For #1, I have experimental features turned on in my prefs. Not sure, > > but I had thought multiline patterns were enabled at some point in the > > past. > > There doesn't seem to be any code to support multiline patterns in > replaceInFiles, so unless I'm missing something, there's nothing to > enable. > > > For #2, here's what happens ... sometimes: > > > > * Mode: xml / xsl > > * Java: both blackdown jdk 1.4.1 and sun jdk 1.4.2.05 > > * I use F11 a lot to comment out single lines. > > * Sometimes, when I hit shift-F11 to uncomment the line, j will > > just hang there. I can switch buffers using the buffer list, the > > entire j window redraws, I can select commands from the menus, > > but no commands (including the ones selected from the menus) > > execute, and I can't move the caret. > > * Force-quitting j and then restarting it is the only option, at > > which point it usually asks to recover a saved file. The > > recovery shows that the shift-F11 command completed > > successfully. > > > > > > I have been completely unable to determine the exact sequence of steps > > to get #2 to occur. I was just trying it again, and j worked perfectly. > > Sorry, but that's all I have. > > The next time this problem occurs, if you hit Ctrl \ in the xterm from > which you've started j (assuming such a thing exists), you should get a > full thread dump, including the JVM's best guess about any possible > deadlock that might be happening. This might help narrow things down > enough to find the bug. > > Thanks! > > -Peter _____________________________ Brad Brizendine, CTO, Glyphix |