From ef2e2f5b3a8bfc1301fb057af49eabd4337f65c0 Mon Sep 17 00:00:00 2001 From: benfry Date: Sun, 6 Feb 2005 20:55:26 +0000 Subject: [PATCH] beginning rev 76 --- build/macosx/dist.sh | 23 ++---- core/done.txt | 66 ++++++++++++++++ core/todo.txt | 65 +--------------- done.txt | 181 +++++++++++++++++++++++++++++++++++++++++++ todo.txt | 181 +------------------------------------------ 5 files changed, 257 insertions(+), 259 deletions(-) diff --git a/build/macosx/dist.sh b/build/macosx/dist.sh index f5e467662..629fb4764 100755 --- a/build/macosx/dist.sh +++ b/build/macosx/dist.sh @@ -100,22 +100,15 @@ find processing -name "*~" -exec rm -f {} ';' find processing -name "._*" -exec rm -f {} ';' find processing -name "CVS" -exec rm -rf {} ';' +stuff -f sitx processing-$REVISION + # zip it all up for release -#mv processing "Processing $REVISION" -#mv processing "processing-$REVISION" -NICE_FOLDER="Processing $SHORT_REVISION" -DMG_NAME="processing-$REVISION" -mv processing "$NICE_FOLDER" - -#stuff -f sitx processing-$REVISION -#WHERE=`pwd` -chmod +x mkdmg -#./mkdmg $WHERE/processing-$REVISION -#./mkdmg processing-$REVISION -./mkdmg "$NICE_FOLDER" "Processing" -mv "$NICE_FOLDER.dmg" "$DMG_NAME.dmg" - -# if there is a command line tool to make a dmg from this dir.. hmm +#NICE_FOLDER="Processing $SHORT_REVISION" +#DMG_NAME="processing-$REVISION" +#mv processing "$NICE_FOLDER" +#chmod +x mkdmg +#./mkdmg "$NICE_FOLDER" "Processing" +#mv "$NICE_FOLDER.dmg" "$DMG_NAME.dmg" # actually, could probably use: # open processing-uncomp.dmg diff --git a/core/done.txt b/core/done.txt index 04fe07de9..2863a43a8 100644 --- a/core/done.txt +++ b/core/done.txt @@ -1,3 +1,69 @@ +0075 +X textureMode(NORMAL_SPACE) screws up the image() command +X image() appears to require IMAGE_SPACE to function properly. +X added focusGained() and focusLost() +X lots of changes to internal naming of vars +X screenX(x, y) and screenY(x, y) added for noDepth() +X add TRIANGLE_FAN +X eyeX, eyeY etc have been renamed cameraX/Y/Z, and cameraNear/Far +X modify targa and tiff writing routines to break into header writing +X writeTIFF, writeHeaderTIFF, writeTGA, writeHeaderTGA +X MAJOR CHANGE: RGBA becomes ARGB for accuracy +o uv coords > 1 shouldn't clamp, they should just repeat ala opengl +o actually i think opengl has a setting for it +o remove need to use depth() at the beginning +X need only be set once +X pmouseX is broken again +X remove pmouseX/Y altogether +X maintain things a bit different +o email the beta@ list to see how people are using pmouseX +X changed PMovie.length to PMovie.duration +X move around/simplify loadImage() inside PApplet +X working to add more die() statements inside PApplet +o can ALPHA fonts work using the other replace modes? + +fixed in previous releases +X text stuff +X text() with alignment doesn't work for multiple lines +X don't change the size of a font when in screen space mode +X patch rotated text (from visualclusto) into bfont +X what sort of api? textSpace(ROTATED) ? +X rotated text has a bug for when it goes offscreen + +opengl +X why is the thing hanging until 'stop' is hit? +X what happens when stop is hit that sets it free? +X (at what point does it start working properly?) +X cache needs to also make things a power of 2 +X if images are already a power of 2, then needn't re-alloc +X cacheIndex needs to be set to -1 when the image is modified +X or maybe have a modified() function? +X debug why certain spots are having errors (see 'problem here' notes) +X INVALID_OPERATION after drawing lines for cube +X fix noLoop bug +X remove errors when drawing textures +X reverse y coordinates +X make play button un-highlight with opengl +X also make window move messages work properly +X very necessary, since opens window at 100x100 +X problem was the threading issues +X bring back renderer menu +X reading/saving pref for opengl renderer +X remove cameraMode(PERSPECTIVE) on each frame +X why is the first one failing? +X still threading issues with running opengl +X threading really horks up dual processor machine.. +X first run hangs until quit +X though doesn't seem to replicate when p5 is restarted +X make sure background() gets called at least once with opengl +X otherwise screen never actually updates +X beginFrame() around setup() +X draw mode stuff happens inside setup.. +o or maybe need to get better at size() inside of draw() ? +X make this consistent with the regular PApplet +X otherwise things are going to be weird/difficult for debugging + + 0074 core X removed zbuffer precision hack in PLine to improve z ordering X no longer set g.dimensions = 3 when using vertex(x, y, 0) diff --git a/core/todo.txt b/core/todo.txt index 7e22022af..d32d8f429 100644 --- a/core/todo.txt +++ b/core/todo.txt @@ -1,67 +1,4 @@ -0075 -X textureMode(NORMAL_SPACE) screws up the image() command -X image() appears to require IMAGE_SPACE to function properly. -X added focusGained() and focusLost() -X lots of changes to internal naming of vars -X screenX(x, y) and screenY(x, y) added for noDepth() -X add TRIANGLE_FAN -X eyeX, eyeY etc have been renamed cameraX/Y/Z, and cameraNear/Far -X modify targa and tiff writing routines to break into header writing -X writeTIFF, writeHeaderTIFF, writeTGA, writeHeaderTGA -X MAJOR CHANGE: RGBA becomes ARGB for accuracy -o uv coords > 1 shouldn't clamp, they should just repeat ala opengl -o actually i think opengl has a setting for it -o remove need to use depth() at the beginning -X need only be set once -X pmouseX is broken again -X remove pmouseX/Y altogether -X maintain things a bit different -o email the beta@ list to see how people are using pmouseX -X changed PMovie.length to PMovie.duration -X move around/simplify loadImage() inside PApplet -X working to add more die() statements inside PApplet -o can ALPHA fonts work using the other replace modes? - -fixed in previous releases -X text stuff -X text() with alignment doesn't work for multiple lines -X don't change the size of a font when in screen space mode -X patch rotated text (from visualclusto) into bfont -X what sort of api? textSpace(ROTATED) ? -X rotated text has a bug for when it goes offscreen - -opengl -X why is the thing hanging until 'stop' is hit? -X what happens when stop is hit that sets it free? -X (at what point does it start working properly?) -X cache needs to also make things a power of 2 -X if images are already a power of 2, then needn't re-alloc -X cacheIndex needs to be set to -1 when the image is modified -X or maybe have a modified() function? -X debug why certain spots are having errors (see 'problem here' notes) -X INVALID_OPERATION after drawing lines for cube -X fix noLoop bug -X remove errors when drawing textures -X reverse y coordinates -X make play button un-highlight with opengl -X also make window move messages work properly -X very necessary, since opens window at 100x100 -X problem was the threading issues -X bring back renderer menu -X reading/saving pref for opengl renderer -X remove cameraMode(PERSPECTIVE) on each frame -X why is the first one failing? -X still threading issues with running opengl -X threading really horks up dual processor machine.. -X first run hangs until quit -X though doesn't seem to replicate when p5 is restarted -X make sure background() gets called at least once with opengl -X otherwise screen never actually updates -X beginFrame() around setup() -X draw mode stuff happens inside setup.. -o or maybe need to get better at size() inside of draw() ? -X make this consistent with the regular PApplet -X otherwise things are going to be weird/difficult for debugging +0076 core _ when running externally, applets don't always get placed properly _ if size is never set, then doesn't always layout diff --git a/done.txt b/done.txt index 7240fc626..a9537240c 100644 --- a/done.txt +++ b/done.txt @@ -1,3 +1,184 @@ +0075 pde +X backported to 73 +X change horizontal offset to 5 instead of 4 +X yet another tweak to the message() function in PdeRuntime +X add -target 1.1 to all the library make entries +X find/replace +o add support for hitting 'ESC' +o option to ignore or match case +X when hitting 'find', select the text in the field +X add extra space to exterior of find (especially on windows) +X fixes to the horizontal scrollbar +X still not perfect, but some tweaks for improvement +X horizontal scroller is always present +X fixed, was a problem of sizing with the new inset size +X cosmetic fixes +X is PdeEditorHeader one pixel too tall +X move the tabs over just slightly +X resize box intrudes on the scroller for the console area +X need to set grow boxes intruding to false +X 1.3 version (deprecated): +X -Dcom.apple.mrj.application.growbox.intrudes=false +X 1.4 version (much nicer): -Dapple.awt.showGrowBox=false +X add mkdmg script to macosx build process +X could significantly speed things up +X make sure that the preproc stuff gets built once +X especially straight out of cvs +X menu option for Renderer > to "Core" or "OpenGL" +X on export, uses this setting to determine if jogl is added to project +o if additional tab is "public" class.. then make external? +X this seems too dangerous +X make font builder vertically resizable +X 200 pt font hoses things / hides buttons +X rename/delete work +X rename is now enabled on the first tab +X delete on first tab will also delete the entire sketch +X leaving an empty sketch will remove it +X avoids creating a zillion sketches when deleting/clearing up sketchbook +X disable 'hide' menu item for the main code tab + +sketchbook/prefs location +X read from registry key to get the proper name for these folders +X also find some way to do this on the mac +X on non-english windows (or at the broad) these will be named differently +X http://support.microsoft.com/?kbid=221837&sd=RMVP +X or is there a proper way to do this with java params? +X move to ~/Application Data/Processing for windows +X and ~/Library/Processing for macosx +X and ~/.processing for linux +X if it can't write prefs here, try creating a subfolder inside p5 +X and if that fails, then tell the user that they suck +X prompt for sketchbook location on linux +X or on mac/pc if it can't write anywhere else +X creating sketchbook / lab situations +X test to see if it's possible to write to "My Documents" +X if not, should bring up a prompt asking where to put sketchbook +X also put something in lib/preferences.txt for default location +X this way a course admin can change the default location +o need to check if volume is read-only, notify and quit if it is +o people are trying to run off the disk image +o actually that should be fine.. +o but make sure the prefs location can be written +o need to pay attention to when running from read-only drive +o reported by brandenberg +o "p5 will launch from the disk image, but will +o not draw the sketch name bar doesn't appear" +X include preferences.txt option to write to p5 folder first +X this can be altered by instructors and re-packaged +X fixed a couple bugs causing the jnilib stuff to crash +X put 'play' on a SwingWorker thread +X test this on a pc to see how it goes, especially with opengl +X doesn't seem to help anything +X PdeRuntime -> SystemOutSiphon, removed MIN_PRIORITY setting +o hack to not use console on the code folder +o no errors will propogate, but it should run fine +X lib/build should not be a subfolder of p5 +X p5 environment installed to /Applications on lab machines +X instead use a temp folder +X maybe when running on lab machines in that case, always java mode? +X or is it possible to add to java.class.path while app is running? +X have to use a special class loader +X new class loader for single files +X if more than one class defined, forces it to run externally +X (basically any time it sees "class" in the code.. +X may be subject to errors, but errs on side of just running ext) +X runs out of processing data folder +X fix for toolbar being one pixel off breaks the windows version +X need to only do the fix on the mac.. ugh +X macosx is using /Library instead of ~/Library.. get this fixed +X domain version wasn't in the stubs.. grr +X or maybe it's an int or something? +X just use reflection to call the function instead +X "use external editor" is broken +X throws exception on load() +X make preproc only build once (across osx, windows, linux) +X preproc: making all functions public that have no specifier +X this will make draw() etc all much easier +X as well as the library events +X focusGained/focusLost was added.. +o if a data file is in the sketch (not data) folder export breaks +o works fine in the editor, but on export gets a nullpointer ex +X no way around this, because needs to be able to read from local dir +X if saving a read-only sketch, default to sketchbook folder +X (before was prompting to save inside the same folder.. annoying) + +building libraries - not gonna build libs in p5 ide for 1.0 +o compiling - main file is a .java not a .pde +o also where no root .pde file with the same name.. (i.e. video) +o on "new library" fill in the default code for the lib +o record imports from java files as well +o ignore imports from java.* or maybe just things pde's classpath +o after export of library, rebuild "import library" menu +o write export-to-library +o need to be able to produce the serial, video, etc libs +o see if sonia project can compile + +libraries +X libraries could handle a series of events.. +X i.e. a 'completion' event, or 'new data' event +X these could be registered on obejcts in a general way +X implemented many events +X movieEvent, cameraEvent, soundEvent +X serialEvent, clientEvent, serverEvent +X if videoEvent exists, don't auto-read? +X though for serial this would only grab the last byte +X video bug on osx? an error message when quitting video sketches: +X java.io.IOException: Bad file descriptor +X at java.io.FileInputStream.readBytes(Native Method) etc... ). +o bring back some form of beginSerial/beginVideo +o openSerial(), serial().. +o should it prompt or use the first available if none specified? +X run library destroy after hitting 'stop' +X quicktime audio doesn't stop after hitting 'stop' +X weirdness external stop not calling stop(), just finished = true +X bring back serial event via reflection +X add serial.last() to api +X video - try to replicate bug that causes hang with camera +X happens when winvdig, camera or quicktime are not properly installed +X documented better in bugs.txt, and sometimes caught in Camera. + +fixed in previous releases +X dist.sh needs to auto-remove .DS_Store boogers +X write script to remove .DS_Store and CVS folders from dist +X # remove DS_Store files +X sudo find . -name ".DS_Store" -depth -exec rm {} \; +X add a "make archive" option +X doesn't rename sketch, just names zip file sketch + time stamp + +hanging bug - now fixed, need to post on the board about it +X keypressed hanging on applets with a code folder +X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1081450102 +X problems running external vm/vm is hanging +X seems to be happening because of virus scanning software (norton) +o may need to launch the applet (using 'start') +o and talk over a socket instead +X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1067867520;start=0 +X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1067643186 +X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1078714442;start=0 +X something in that one about mouse position halting or not +to try: 1) don't use a bufferedreader.. just read it and then use a +buffered reader around the byte array 2) don't do anything with the +input, and just read it off without signalling message() +* it may be something to do with the editor consoles. try shutting off +output to the console and see if it'll still crash. +* most likely related to io streams to the external java process. when +run via a disconnected process, using "cmd /c start java", the thing +will never hang. in that instance, the process terminates almost +immediately, and no i/o needs to happen (since it's a cmd prompt that +never shows up). +* it could also be a graphics sync bug that just gets more testy +because the environment, a second java process, is running at the same +time. mis.newPixels() may hork since it's over in the applet's thread, +and it might be calling repaint() or Toolkit.sync() to update the +image on-screen. +* shows up on key presses.. not sure if this is because of the actual +key press, or if it's because they're often accompanied by a println() +* blank spaces in filenames/parent folder often cause trouble.. not +sure if related. same for PATH and CLASSPATH. +* some virus scanning software, particularly older NAV versions cause +the trouble. + + 0074 pde X added florian's new icons diff --git a/todo.txt b/todo.txt index b5720b96d..7054d2d94 100644 --- a/todo.txt +++ b/todo.txt @@ -1,183 +1,4 @@ -0075 pde -X backported to 73 -X change horizontal offset to 5 instead of 4 -X yet another tweak to the message() function in PdeRuntime -X add -target 1.1 to all the library make entries -X find/replace -o add support for hitting 'ESC' -o option to ignore or match case -X when hitting 'find', select the text in the field -X add extra space to exterior of find (especially on windows) -X fixes to the horizontal scrollbar -X still not perfect, but some tweaks for improvement -X horizontal scroller is always present -X fixed, was a problem of sizing with the new inset size -X cosmetic fixes -X is PdeEditorHeader one pixel too tall -X move the tabs over just slightly -X resize box intrudes on the scroller for the console area -X need to set grow boxes intruding to false -X 1.3 version (deprecated): -X -Dcom.apple.mrj.application.growbox.intrudes=false -X 1.4 version (much nicer): -Dapple.awt.showGrowBox=false -X add mkdmg script to macosx build process -X could significantly speed things up -X make sure that the preproc stuff gets built once -X especially straight out of cvs -X menu option for Renderer > to "Core" or "OpenGL" -X on export, uses this setting to determine if jogl is added to project -o if additional tab is "public" class.. then make external? -X this seems too dangerous -X make font builder vertically resizable -X 200 pt font hoses things / hides buttons -X rename/delete work -X rename is now enabled on the first tab -X delete on first tab will also delete the entire sketch -X leaving an empty sketch will remove it -X avoids creating a zillion sketches when deleting/clearing up sketchbook -X disable 'hide' menu item for the main code tab - -sketchbook/prefs location -X read from registry key to get the proper name for these folders -X also find some way to do this on the mac -X on non-english windows (or at the broad) these will be named differently -X http://support.microsoft.com/?kbid=221837&sd=RMVP -X or is there a proper way to do this with java params? -X move to ~/Application Data/Processing for windows -X and ~/Library/Processing for macosx -X and ~/.processing for linux -X if it can't write prefs here, try creating a subfolder inside p5 -X and if that fails, then tell the user that they suck -X prompt for sketchbook location on linux -X or on mac/pc if it can't write anywhere else -X creating sketchbook / lab situations -X test to see if it's possible to write to "My Documents" -X if not, should bring up a prompt asking where to put sketchbook -X also put something in lib/preferences.txt for default location -X this way a course admin can change the default location -o need to check if volume is read-only, notify and quit if it is -o people are trying to run off the disk image -o actually that should be fine.. -o but make sure the prefs location can be written -o need to pay attention to when running from read-only drive -o reported by brandenberg -o "p5 will launch from the disk image, but will -o not draw the sketch name bar doesn't appear" -X include preferences.txt option to write to p5 folder first -X this can be altered by instructors and re-packaged -X fixed a couple bugs causing the jnilib stuff to crash -X put 'play' on a SwingWorker thread -X test this on a pc to see how it goes, especially with opengl -X doesn't seem to help anything -X PdeRuntime -> SystemOutSiphon, removed MIN_PRIORITY setting -o hack to not use console on the code folder -o no errors will propogate, but it should run fine -X lib/build should not be a subfolder of p5 -X p5 environment installed to /Applications on lab machines -X instead use a temp folder -X maybe when running on lab machines in that case, always java mode? -X or is it possible to add to java.class.path while app is running? -X have to use a special class loader -X new class loader for single files -X if more than one class defined, forces it to run externally -X (basically any time it sees "class" in the code.. -X may be subject to errors, but errs on side of just running ext) -X runs out of processing data folder -X fix for toolbar being one pixel off breaks the windows version -X need to only do the fix on the mac.. ugh -X macosx is using /Library instead of ~/Library.. get this fixed -X domain version wasn't in the stubs.. grr -X or maybe it's an int or something? -X just use reflection to call the function instead -X "use external editor" is broken -X throws exception on load() -X make preproc only build once (across osx, windows, linux) -X preproc: making all functions public that have no specifier -X this will make draw() etc all much easier -X as well as the library events -X focusGained/focusLost was added.. -o if a data file is in the sketch (not data) folder export breaks -o works fine in the editor, but on export gets a nullpointer ex -X no way around this, because needs to be able to read from local dir -X if saving a read-only sketch, default to sketchbook folder -X (before was prompting to save inside the same folder.. annoying) - -building libraries - not gonna build libs in p5 ide for 1.0 -o compiling - main file is a .java not a .pde -o also where no root .pde file with the same name.. (i.e. video) -o on "new library" fill in the default code for the lib -o record imports from java files as well -o ignore imports from java.* or maybe just things pde's classpath -o after export of library, rebuild "import library" menu -o write export-to-library -o need to be able to produce the serial, video, etc libs -o see if sonia project can compile - -libraries -X libraries could handle a series of events.. -X i.e. a 'completion' event, or 'new data' event -X these could be registered on obejcts in a general way -X implemented many events -X movieEvent, cameraEvent, soundEvent -X serialEvent, clientEvent, serverEvent -X if videoEvent exists, don't auto-read? -X though for serial this would only grab the last byte -X video bug on osx? an error message when quitting video sketches: -X java.io.IOException: Bad file descriptor -X at java.io.FileInputStream.readBytes(Native Method) etc... ). -o bring back some form of beginSerial/beginVideo -o openSerial(), serial().. -o should it prompt or use the first available if none specified? -X run library destroy after hitting 'stop' -X quicktime audio doesn't stop after hitting 'stop' -X weirdness external stop not calling stop(), just finished = true -X bring back serial event via reflection -X add serial.last() to api -X video - try to replicate bug that causes hang with camera -X happens when winvdig, camera or quicktime are not properly installed -X documented better in bugs.txt, and sometimes caught in Camera. - -fixed in previous releases -X dist.sh needs to auto-remove .DS_Store boogers -X write script to remove .DS_Store and CVS folders from dist -X # remove DS_Store files -X sudo find . -name ".DS_Store" -depth -exec rm {} \; -X add a "make archive" option -X doesn't rename sketch, just names zip file sketch + time stamp - -hanging bug - now fixed, need to post on the board about it -X keypressed hanging on applets with a code folder -X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1081450102 -X problems running external vm/vm is hanging -X seems to be happening because of virus scanning software (norton) -o may need to launch the applet (using 'start') -o and talk over a socket instead -X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1067867520;start=0 -X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1067643186 -X http://processing.org/discourse/yabb/YaBB.cgi?board=Proce55ing_software_bugs;action=display;num=1078714442;start=0 -X something in that one about mouse position halting or not -to try: 1) don't use a bufferedreader.. just read it and then use a -buffered reader around the byte array 2) don't do anything with the -input, and just read it off without signalling message() -* it may be something to do with the editor consoles. try shutting off -output to the console and see if it'll still crash. -* most likely related to io streams to the external java process. when -run via a disconnected process, using "cmd /c start java", the thing -will never hang. in that instance, the process terminates almost -immediately, and no i/o needs to happen (since it's a cmd prompt that -never shows up). -* it could also be a graphics sync bug that just gets more testy -because the environment, a second java process, is running at the same -time. mis.newPixels() may hork since it's over in the applet's thread, -and it might be calling repaint() or Toolkit.sync() to update the -image on-screen. -* shows up on key presses.. not sure if this is because of the actual -key press, or if it's because they're often accompanied by a println() -* blank spaces in filenames/parent folder often cause trouble.. not -sure if related. same for PATH and CLASSPATH. -* some virus scanning software, particularly older NAV versions cause -the trouble. - +0076 pde _ get an xml library in there _ nanoxml problems with manifest