Shop OBEX P1 Docs P2 Docs Learn Events
Open Propeller Project #3: Propeller IDE V0.1 Package Available - Page 20 — Parallax Forums

Open Propeller Project #3: Propeller IDE V0.1 Package Available

1171820222329

Comments

  • mindrobotsmindrobots Posts: 6,506
    edited 2014-03-28 11:09
    jazzed wrote: »
    See, I told you I could make an IDE just for your favorite language ;-) Made it last year actually.

    Ok, this settles it (mostly). I have to go back to using an external loader. The code is exactly the same as in the propeller load program, but the Qt environment is getting in the way of the strict propeller loader timing requirements. There is an alternative ... but it involves another program also. Fortunately the beanie still works. Unfortunately, we will need a delay before printing anything to the console.

    I'm a simple person and I have simple needs!! :lol:

    I kind of like the separate loader idea. I don't think there are really any performance penalties now days with a wrapper program on a PC calling worker/helper programs (gee, you could make an operating system where a bunch of small tools could be strung together to get work done). Having the separate components also makes it easier to implement new feature/fixes in each area (and some day maybe only have to touch one area). IMHO
  • RoadsterRoadster Posts: 209
    edited 2014-03-28 17:23
    jazzed wrote: »
    Ok. Can you try the attached .zip and the program Loader.exe to load your .binary program? Loader.exe is a small GUI program - just double-click it. There is also a test.binary in the package. It is a key data-point to find out if the loader works for you and everyone else for that matter. The Loader.exe simply runs the propeller-load.exe program which is known to be very reliable.
    I tried the loader program and it worked every time, however this is what I noticed:
    when I start the program for the first time each load is very slow, but if I click load again it loads much quicker
    If I close the loader program and start it again it is very slow on the first load again but each load after that are much quicker.

    I tried other binary files of different sizes and got the same results as above.
  • dgatelydgately Posts: 1,623
    edited 2014-03-28 17:40
    jazzed wrote: »
    Does anyone feel strongly about keeping the grey line-number bar?

    How about a green one?

    Hmm, in a time when operating systems are going towards less color... I made sure the Find & Replace option was sans-color :smile:



    Green looks OK, to me. As long as it's different than the color of any of the syntax colors, should work well!


    Dennis
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-03-28 19:54
    I need to test v0.4 on my Win7 under Parallels on the Mac. That one got all buggered up with v0.3 too! I'll try that tomorrow.
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-03-29 09:12
    Update! It's "tomorrow" and v0.3 worked fine running on a Win7 VM under Parallels. It even recognized my P2 and wanted to program it!! :lol:

    Jazz Loader also worked - it was able to pick up binaries and send them to a Quickstart. YAY!!

    EDIT: I meant v0.4 WORKED FINE!!!
  • jazzedjazzed Posts: 11,803
    edited 2014-03-29 10:39
    mindrobots wrote: »
    Update! It's "tomorrow" and v0.3 worked fine running on a Win7 VM under Parallels. It even recognized my P2 and wanted to program it!! :lol:

    Jazz Loader also worked - it was able to pick up binaries and send them to a Quickstart. YAY!!
    Ugh. So if it's tomorrow's tomorrow will it fail again?

    I'm replacing the built-in loader with the reliable external "Loader" based program. I think I've figured out a way around needing a startup delay.

    I've decided not to use the green line number bar, even though everyone insists on the editor being colorful. There are some changes in the highlighter though.

    Roy, Dennis and other contributors, by the end of the day today I should have a big push.
  • dgatelydgately Posts: 1,623
    edited 2014-03-29 10:46
    jazzed wrote: »
    Roy, Dennis and other contributors, by the end of the day today I should have a big push.

    Thanks Jazzed, will check that push out!

    External loader will be for all systems? (Mac, Win, Linux)?


    dgately
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-03-29 10:54
    jazzed wrote: »
    Ugh. So if it's tomorrow's tomorrow will it fail again?

    I'm replacing the built-in loader with the reliable external "Loader" based program. I think I've figured out a way around needing a startup delay.

    I've decided not to use the green line number bar, even though everyone insists on the editor being colorful. There are some changes in the highlighter though.

    Roy, Dennis and other contributors, by the end of the day today I should have a big push.

    Sorry for any confusion -

    V0.4 WORKS GREAT on Win7 under Parallels too!!


    NO LOADING PROBLEMS!!
  • KMyersKMyers Posts: 433
    edited 2014-03-29 11:07
    Working on my win 7 system and the loader works also. How do you select the com port that loader uses?

    The prop id does infact work as it should. It just notifies me that some of my USB com ports dont have a Prop connected my bad. Also it compilies my SHT-11 program but complains about the port my lcd is on but bot loads ram and burns it on the activity board..

    The Debug is disabled for the time being, right?
  • jazzedjazzed Posts: 11,803
    edited 2014-03-29 14:06
    KMyers wrote: »
    Working on my win 7 system and the loader works also. How do you select the com port that loader uses?
    The little GUI loader just uses the first Propeller port. There is no control for it.

    KMyers wrote: »
    The Debug is disabled for the time being, right?
    It isn't a production feature.


    @Rick, I'm glad V0.4 worked for you. Unfortunately we have one failure report (from Roadster), so to me that means V0.4 is DOA.

    Working on that external load thing now. Unfortunately the Port sharing thing I wanted to do won't be possible on Windows.
  • KMyersKMyers Posts: 433
    edited 2014-03-29 14:18
    Thanks Jazzed, explains why it went to my C3...
  • jazzedjazzed Posts: 11,803
    edited 2014-03-30 13:13
    Hi.

    @Roadster (and everyone else) please try V0.5 posted here. Test the Hat and program loading.

    This is similar to V0.4 except I turned off an unnecessary monitor thread that may have caused V0.4 load failures.

    This should also fix the Parallax Serial Terminal.spin enter problem (uncheck Terminal -> Options "Enter is NL" to use old behavior).

    Thanks.
  • JonnyMacJonnyMac Posts: 8,955
    edited 2014-03-30 14:26
    I don't know how it works (and don't need to), but I'm very happy about how fast programs download and open the debug window. Thanks, guys. Today I was playing with my Dynamixel tester on a pan/tilt rig which uses a lot of PST-compatible formatting commands.

    I also appreciate that I can change the font in the editor and debug window -- finally given up on Parallax font.
    1024 x 640 - 98K
  • JonnyMacJonnyMac Posts: 8,955
    edited 2014-03-30 15:02
    Long-term, will we be able to turn off method name suggestions/replacements? Now that I'm actually trying to write a program, I'm finding them more annoying than helpful. That I have the list of methods available to me in the lower-left panel is enough (for me; I realize others may like this feature).
  • jazzedjazzed Posts: 11,803
    edited 2014-03-30 17:15
    JonnyMac wrote: »
    Long-term, will we be able to turn off method name suggestions/replacements? Now that I'm actually trying to write a program, I'm finding them more annoying than helpful. That I have the list of methods available to me in the lower-left panel is enough (for me; I realize others may like this feature).
    Yes, these can all be made optional. Which do you find least annoying? Or conversely, which is most helpful?

    If you hit escape, the suggestion or autocomplete list should go away. Enter will accept an item from either. Tab also accepts suggestions. I can't get tab to use an entry from the autocomplete list.
  • RoadsterRoadster Posts: 209
    edited 2014-03-31 06:37
    jazzed wrote: »
    Hi.

    @Roadster (and everyone else) please try V0.5 posted here. Test the Hat and program loading.

    This is similar to V0.4 except I turned off an unnecessary monitor thread that may have caused V0.4 load failures.

    This should also fix the Parallax Serial Terminal.spin enter problem (uncheck Terminal -> Options "Enter is NL" to use old behavior).

    Thanks.

    This version 0.5 works most of the time now, but when it fails you get no message saying it failed.

    The terminal program is now working correctly
  • jazzedjazzed Posts: 11,803
    edited 2014-03-31 08:36
    Roadster wrote: »
    This version 0.5 works most of the time now, but when it fails you get no message saying it failed.
    I suppose that's progress. Wish I could reproduce the problem.

    There is one more thing to do that may improve it; will post V0.6 later.

    Does the beanie button work all the time? What happens when the loader fails?
  • RoadsterRoadster Posts: 209
    edited 2014-03-31 11:30
    jazzed wrote: »
    Does the beanie button work all the time?
    Yes
    jazzed wrote: »
    What happens when the loader fails?
    When I hit the debug button, the loader pops up and closes and the serial terminal never opens
  • RsadeikaRsadeika Posts: 3,824
    edited 2014-04-01 03:56
    V0.5, I am working on a larger Spin program to see if I run into any substantial problems. A problem that started to appear, in a random fashion, is the use of terminal debug screen, sometimes it works, and sometimes it does not. Sometimes when I hit the debug icon, the terminal screen does not appear. Yesterday, when I was working on the program, the session worked as expected. Today, the problem is occurring quite frequently, even exiting the program does not fix the problem all the time. Sometimes I have to do a numerous amount of exit/start sequences. Maybe Windows does not like QT5?

    I noticed that the auto insert feature can become quite annoying sometimes, especially when you have a lot of objects that you are using. The problem is that the list gets to be very long and you are forced to make a choice as opposed too just typing in what you really want. Maybe a way to turn on/off the feature would be nice or somehow shorten the list to just the particular method that is required.

    When I use the hammer icon, maybe it should state that the program is OK. Is the zip icon a functioning choice? After I am done with the program I will be zipping it as a project.

    <edit> Concerning the debug terminal window, the correct COM port has to be selected before the debug window will work. When I start a Propeller IDE session the connected COM port is not always selected. It seems like in SimpleIDE that was not a problem, and the functioning of the debug(terminal) screen was also more built in. I know this is supposed to be a clone of the Propeller Tool, and not the utility of SimpleIDE. <edit>

    Ray
  • jazzedjazzed Posts: 11,803
    edited 2014-04-01 10:22
    Hi Folks.

    I've made a branch of the repository for testing loader ideas. We are very close to having a good solution, but need more testing. Hopefully the branch that folks can build will increase test coverage.

    What is a branch? A repository is a tree of code. Usually the main or default branch is the trunk, and branches are variations.

    How do I use the loader branch? Close Qt Creator or other tools that may have files open. Do these commands: "hg pull" then "hg update loader" Then build as normal.

    BTW, I've had some recent exposure to git that I don't like at all. I'm not inclined to use it.
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-04-01 12:56
    With new buildable options out there would you like testing across all platforms or still testing concentrating on Windows. Is the loading issue an Windows only issue or is that an unknown at this time?

    Ready to play, regardless.
  • jazzedjazzed Posts: 11,803
    edited 2014-04-01 18:10
    mindrobots wrote: »
    With new buildable options out there would you like testing across all platforms or still testing concentrating on Windows. Is the loading issue an Windows only issue or is that an unknown at this time?

    Ready to play, regardless.

    Hi Rick,

    I've tested the "loader" branch on Windows, Mac, and Linux. So, it "should" work for anyone who can build from source. That is: Please test it if you have a chance.

    There is now a Windows RT Qt toolchain with Qt5.3 beta if anyone is interested in trying that. I've been building with Qt5.3 libraries also.
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-04-01 19:07
    jazzed wrote: »
    Hi Folks.

    I've made a branch of the repository for testing loader ideas. We are very close to having a good solution, but need more testing. Hopefully the branch that folks can build will increase test coverage.

    What is a branch? A repository is a tree of code. Usually the main or default branch is the trunk, and branches are variations.

    How do I use the loader branch? Close Qt Creator or other tools that may have files open. Do these commands: "hg pull" then "hg update loader" Then build as normal.

    BTW, I've had some recent exposure to git that I don't like at all. I'm not inclined to use it.

    Hi Steve,

    I just did the "hg pull" and "hg update loader" on my Win7 laptop and then did "Rebuild All" for both the debug and release. When I executed either of the .exe files that were created via bouble click from Windows Explorer, I got this error dialog:

    pide-error-2014_04_01.png

    It DOES run fine if I use the "Run" green arrow from inside QT. In the past, I have always been able to build working .exe files. Like a said, not a big deal now, just a note for the difference.

    So far, Beanie and loading has worked on this version. The debug terminal also handles the enter key properly.

    Off to see if I can break anything else in the Windows world........
    530 x 226 - 12K
  • jazzedjazzed Posts: 11,803
    edited 2014-04-01 20:34
    Rick,

    I'm not too worried about that error because the environment is not setup correctly if you're just double-clicking from the release/debug folders. The error is a little different from the one I get though ....


    attachment.php?attachmentid=107910&d=1396409622


    ;-)
    501 x 179 - 13K
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-04-01 20:40
    jazzed wrote: »
    Rick,

    I'm not too worried about that error because the environment is not setup correctly if you're just double-clicking from the release/debug folders. The error is a little different from the one I get though ....


    attachment.php?attachmentid=107910&d=1396409622


    ;-)

    Got it! Friends don't let friends drink and code.....because they might have to do the code review in the morning! :smile:

    I'm good to go!
  • jazzedjazzed Posts: 11,803
    edited 2014-04-01 20:59
    LOL.

    It took me two weeks to recover from my first and only "4 authorized beers at work" coding attempt.
  • mindrobotsmindrobots Posts: 6,506
    edited 2014-04-01 21:03
    jazzed wrote: »
    LOL.

    It took me two weeks to recover from my first and only "4 authorized beers at work" coding attempt.

    That's funny!! I code like that sober!!:frown:
  • jazzedjazzed Posts: 11,803
    edited 2014-04-01 21:54
    Posted a V0.6 windows package here: https://drive.google.com/file/d/0BzcfH7bdVTbtLTlzb1k1S2JoYjQ/edit?usp=sharing

    V0.6 represents what is currently in the "loader" branch. If this is solid (Roadster?), I'll merge changes to the default and retire the loader branch.

    @Roadster, can you please test that and let me know if the loader behavior is better?
  • dgatelydgately Posts: 1,623
    edited 2014-04-01 22:37
    Jazzed,

    Built and ran on Mac OS X (10.9.2). Had no problems ID'ing my Activity Board, loading, debugging! The console worked, running my test code from last month's postings.

    dgately
  • lmclarenlmclaren Posts: 104
    edited 2014-04-02 02:47
    Testing with V0.6

    With only one Propeller chip connected it seems to work fine as far as the communications are concerned.

    If I connect a second serial interface (Parallax USB int) that I use for debugging, the beanie finds both but I don't get any option on which is used for the programming or debugging on the terminal.

    It still seems to program fine, just can't use the second port.


    When using the auto complete, if I have the name of a object with both functions and constants, I can hit the "." and see both functions and constants but I can not use a constant, eg, I must have a "." before it instead of "#", maybe if you can not use the "#" to bring up the list, how about it you choose a constant, it removes the "." and puts in a "#"?

    thanks again for your hard work.

    Lee
Sign In or Register to comment.