Re: Compiler-chain issues(?) with DIA code from cvs
On Tue, 04 May 2004, Ben Combee wrote: > You really should be using pilrc 3.1 or 3.2 beta -- the 3.0 version has > problems with auto control sizing and also doesn't set some control > attributes correctly. Well, I finally got back to a machine I could sync to and tried a new version compiled (in conjunction with) pilrc 3.2b1 and I see exactly the same behavior... Can someone comment on what version of prc-tools they are using? I have 2.2.90 -- the most current version for Debian's 'sid' distribution. However, I see that it is fairly dated (Mar 2003). Should I be upgrading to 2.3 off the sourceforge.org site? -- Brad -- Brad Sawatzky <[EMAIL PROTECTED]> University of Virginia Physics Department Ph: (434) 924-6580Fax: (434) 924-7909 ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Compiler-chain issues(?) with DIA code from cvs
From: "Ben Combee" <[EMAIL PROTECTED]> > What is the viewer.rcp.s file? It's produced by a combination of cpp and my doaddition.pl script that does addition within wordlist resources. Alex ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Compiler-chain issues(?) with DIA code from cvs
At 09:36 AM 5/6/2004, you wrote: Thanks for the note. I've upgraded pilrc to 3.2b1 -- it generates a viewer.rcp.s file identical to pilrc 3.0 though... I didn't think to do a cmp with the .ro files (if that would even be meaningful). What is the viewer.rcp.s file? I don't see how PilRC can actually generate assembly code, which is the traditional meaning of .s. I just looked at the Makefile for Plucker, and I'm still not seeing how a .s file comes into play. The changes to PilRC would alter the .ro files, since those are the files that actually contain all of your resources. -- Ben Combee, DTS technical lead, PalmSource, Inc. Read "Combee on Palm OS" at http://palmos.combee.net/ ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Compiler-chain issues(?) with DIA code from cvs
On Tue, 04 May 2004, Ben Combee wrote: > At 09:16 PM 5/4/2004, Brad Sawatzky wrote: > >I get a consistent crash (requires pin-reset) under the following > >conditions: > > - completely fresh install on the PDA (all prefs, etc deleted using > >FileZ) > > - plucker viewer 1.7.1 (current from today's cvs) > > - compiled under linux (debian sid) using > > prc-tools 2.2.90 > > pilrc 3.0 > > You really should be using pilrc 3.1 or 3.2 beta -- the 3.0 version has > problems with auto control sizing and also doesn't set some control > attributes correctly. Thanks for the note. I've upgraded pilrc to 3.2b1 -- it generates a viewer.rcp.s file identical to pilrc 3.0 though... I didn't think to do a cmp with the .ro files (if that would even be meaningful). > >Unrelated (AFAIK) compile-time quirk: > >- A 'make' from either the plucker-cvs/ or viewer/ directories fails when > > it enters plucker-cvs/viewer/fonts. (cd plucker-cvs/viewer/fonts; make) > > compiles cleanly. The problem is that the root makefile attempts to > > compile fontconf.c using m68k-palmos-gcc rather than just gcc... A little more poking shows that the variable CC isn't being reset in viewer/font/Makefile, so it is inheriting the CC=m68k-palmos-gcc set in viewer/Makefile. The strikes me as the logical way for make to behave -- do other linux developers see this problem, or is it just me? -- Brad -- Brad Sawatzky <[EMAIL PROTECTED]> University of Virginia Physics Department Ph: (434) 924-6580Fax: (434) 924-7909 ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Compiler-chain issues(?) with DIA code from cvs
At 09:16 PM 5/4/2004, you wrote: The PDA is a Garmin iQue 3600, a PalmOS 5.2.1 device, 320x480 screen. I get a consistent crash (requires pin-reset) under the following conditions: - completely fresh install on the PDA (all prefs, etc deleted using FileZ) - plucker viewer 1.7.1 (current from today's cvs) - compiled under linux (debian sid) using prc-tools 2.2.90 pilrc 3.0 You really should be using pilrc 3.1 or 3.2 beta -- the 3.0 version has problems with auto control sizing and also doesn't set some control attributes correctly. The only compile-time warnings are of the form: viewer.rcp:17020: warning : Form has editable field(s) without Graffiti State Indicator viewer.rcp:17228: warning : Form has editable field(s) without Graffiti State Indicator viewer.rcp:17445: warning : Form has editable field(s) without Graffiti State Indicator Those are just warnings -- in PilRC 3.1, you can add NOGRAFFITISTATEINDICATOR to those forms to tell PilRC that you actually intended to have an editable field without giving the user the ability to see what writing mode they are in. -- Ben Combee, DTS technical lead, PalmSource, Inc. Read "Combee on Palm OS" at http://palmos.combee.net/ ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Compiler-chain issues(?) with DIA code from cvs
On Tue, 4 May 2004, Brad Sawatzky wrote: > I get a consistent crash (requires pin-reset) under the following > conditions: > - completely fresh install on the PDA (all prefs, etc deleted using > FileZ) > - plucker viewer 1.7.1 (current from today's cvs) > - compiled under linux (debian sid) using > prc-tools 2.2.90 > pilrc 3.0 > ./configure --enable-armlets > The error occurs when switching to the mainform (the form that displays the > actual document text, etc), but only if the toolbar is being drawn. The > crash occurs immediately after drawing the toolbar but before any page text > is drawn. If the toolbar is disabled in the prefs (ie. while on the > library form) then the mainform will render successfully. There still > seems to be a leak though(?), as toggling the DIA a few times will soon > lead to crash. Send me a copy of your viewer.rcp.s file. Alex -- Dr. Alexander R. Pruss || e-mail: [EMAIL PROTECTED] Philosophy Department || online papers and home page: Georgetown University || www.georgetown.edu/faculty/ap85 Washington, DC 20057|| U.S.A. || - "Philosophiam discimus non ut tantum sciamus, sed ut boni efficiamur." - Paul of Worczyn (1424) ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Compiler-chain issues(?) with DIA code from cvs
The PDA is a Garmin iQue 3600, a PalmOS 5.2.1 device, 320x480 screen. I get a consistent crash (requires pin-reset) under the following conditions: - completely fresh install on the PDA (all prefs, etc deleted using FileZ) - plucker viewer 1.7.1 (current from today's cvs) - compiled under linux (debian sid) using prc-tools 2.2.90 pilrc 3.0 ./configure --enable-armlets The error occurs when switching to the mainform (the form that displays the actual document text, etc), but only if the toolbar is being drawn. The crash occurs immediately after drawing the toolbar but before any page text is drawn. If the toolbar is disabled in the prefs (ie. while on the library form) then the mainform will render successfully. There still seems to be a leak though(?), as toggling the DIA a few times will soon lead to crash. FWIW, the error dialog reads: MemoryMgr.c (line 3635) Null handle The only compile-time warnings are of the form: viewer.rcp:17020: warning : Form has editable field(s) without Graffiti State Indicator viewer.rcp:17228: warning : Form has editable field(s) without Graffiti State Indicator viewer.rcp:17445: warning : Form has editable field(s) without Graffiti State Indicator The daily snapshot works fine. Earlier versions of of the plucker code (prior to the introduction of the DIA code) compile and run cleanly. Can someone tell me what prc-tools and pilrc versions are being used to compile the snapshots? I have a hunch that pilrc 3.0 is doing something unfortunate. Unrelated (AFAIK) compile-time quirk: - A 'make' from either the plucker-cvs/ or viewer/ directories fails when it enters plucker-cvs/viewer/fonts. (cd plucker-cvs/viewer/fonts; make) compiles cleanly. The problem is that the root makefile attempts to compile fontconf.c using m68k-palmos-gcc rather than just gcc... Any thoughts? -- Brad -- Brad Sawatzky <[EMAIL PROTECTED]> University of Virginia Physics Department Ph: (434) 924-6580Fax: (434) 924-7909 ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev