Yeah the obj dir is missing, creates a file on error in the makefile; should fix that :P

Downloaded the latest toolchain on my iMac (macOS Sierra 10.12.1), installed the Wine Devel 1.9.23 .pkg and set up the environment variables.
Works fine so far. I was able to successfully build the demo0 and the demo1, transfer them to my TI-84 Plus CE-T and run them there via Asm(prgmDEMO0 and Asm(prgmDEMO1. 😃

The demo2, however, can't be compiled. Here's the output of make.exe:

Z:\Users\stefan\CEdev\examples\demo_2\src\main.c
Z:\USERS\STEFAN\CEDEV\EXAMPLES\DEMO_2\SRC\MAIN.C	(22,42) :	WARNING (197) No function prototype "lcd_GetBacklightLevel" in scope
Z:\USERS\STEFAN\CEDEV\EXAMPLES\DEMO_2\SRC\MAIN.C	(26,25) :	WARNING (197) No function prototype "lcd_SetBacklightLevel" in scope
Z:\USERS\STEFAN\CEDEV\EXAMPLES\DEMO_2\SRC\MAIN.C	(31,25) :	WARNING (197) No function prototype "lcd_SetBacklightLevel" in scope
Z:\USERS\STEFAN\CEDEV\EXAMPLES\DEMO_2\SRC\MAIN.C	(34,24) :	WARNING (197) No function prototype "lcd_SetBacklightLevel" in scope
ConvPNG 5.5 by M.Waltz

Converted icon 'iconc.png'
Linking...
WARNING (724) --> Symbol _lcd_SetBacklightLevel (obj\main.obj) is not defined.
WARNING (724) --> Symbol _lcd_GetBacklightLevel (obj\main.obj) is not defined.


SPACE ALLOCATION:
=================

Space                     Base         Top      Size      Used    Unused
------------------ ----------- ----------- --------- --------- ---------
RAM                   D:D1A87F    D:D1AA71   300000H      1F3H   2FFE0DH
                                          (  3145728       499   3145229)

ERROR: Object file(s) deleted because of option unresolved=fatal.
Consult bin/DEMO2.map for more information.
make: *** [bin/DEMO2.hex] Error -1

Both functions have no prototype in the main.c file. Is that meant to be so?

"A problem caused the program to stop working correctly. Windows will close the program and notify you if a solution is available."

[tdickerson@dickie] template $ wine ../../bin/make
CC: Unknown switch "..\INCLUDE\CE"
Linking...
WARNING (731) --> File "main.obj" is not found.
make: *** [template.hex] Error -1

I get a compile error under memset_fast due to size_t not being recognised.

If would be nice if you can have a mask for an image, that not every pixel of the image will be converted/in the output.

Get on it people. cracks whip

Both stdint.h and stdbool.h should be included in ti84pce.h.

Tilemap isn't rendered if global scope. wtf.

The resulting 8xp is however still missing the symbols from the first attempt.
See, for example:

C:\Scanline-Rendering\ScanLineRender\ScanLineRender>..\toolchain\CEdev\bin\make.exe
ael.c
DEBUGCONFIG.H   (12,8) :        WARNING (3) Illegal redefinition of macro "NDEBUG"
ael.src
edge.c
DEBUGCONFIG.H   (12,8) :        WARNING (3) Illegal redefinition of macro "NDEBUG"
edge.src
point.c
point.src
primitive.c
DEBUGCONFIG.H   (12,8) :        WARNING (3) Illegal redefinition of macro "NDEBUG"
primitive.src
scanlinerenderer.c
DEBUGCONFIG.H   (12,8) :        WARNING (3) Illegal redefinition of macro "NDEBUG"
scanlinerenderer.src
transformation.c
transformation.src
main.c
DEBUGCONFIG.H   (12,8) :        WARNING (3) Illegal redefinition of macro "NDEBUG"
main.src
Linking...
WARNING (724) --> Symbol _RBMapPut (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBDelete (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBMapRemove (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBTreeDestroy (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBMapNodeAlloc (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _mergeSort (AEL.obj) is not defined.
WARNING (724) --> Symbol _RBSetContains (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBTreeInsert (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBNodeAlloc (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _SafeMalloc (AEL.obj) is not defined.
WARNING (724) --> Symbol _RBSetAdd (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBTreeMapInit (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBExactQuery (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _freeLink (AEL.obj) is not defined.
WARNING (724) --> Symbol _removeLink (AEL.obj) is not defined.
WARNING (724) --> Symbol _linkFront (AEL.obj) is not defined.
WARNING (724) --> Symbol _RBTreeInit (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _RBTreeClear (ScanlineRenderer.obj) is not defined.
WARNING (724) --> Symbol _TreeSuccessor (AEL.obj) is not defined.

SPACE ALLOCATION:
=================

Group: group->name                          Base         Top      Size      Used    Unused
------------------------------------ ----------- ----------- --------- --------- ---------
Space: RAM                              00D031F6    00D1DCF9    65801H     352FH    622D2H
                                                            (   415745     13615    402130)

Space                     Base         Top      Size      Used    Unused
------------------ ----------- ----------- --------- --------- ---------
RAM                   D:D031F6    D:D1DCF9    65801H     352FH    622D2H
                                          (   415745     13615    402130)

ERROR: Object file(s) deleted because of option unresolved=fatal.
Consult scanln.map for more information.
make: *** [scanln.hex] Error -1

C:\Scanline-Rendering\ScanLineRender\ScanLineRender>..\toolchain\CEdev\bin\make.exe

Input File: scanln.hex
Output File: scanln.8xp
Output Calculator Name: SCANLN
Mark archived: No
Success!

Program Size: 13450 bytes

Can we please have some sort of common hook handling system for the CE? Please? So that way people don't break everything all the time?

ti_DetectVar takes as the third argument the file type, but that isn't correctly loaded in the file type for the searching routine, see https://github.com/CE-Programming/toolchain/blob/master/src/fileioc/fileioc.asm#L836-L843. It now loads the LSB of the second argument in the file type ;)

Yep this things perhaps a part of graphx.asm? 👍

Personally, I find having to do this ugly; as it should only be used at the end of a program. Please put it in the cstartup.asm file, and add a rule to the makefile to allow it in order to allow for a custom cleanup if need be.

The instructions make reference to a non-existent bin folder.

[thomas@Mithlond] ScanLineRender $ wineconsole ../toolchain/CEdev/bin/ez80cc.exe
err:winediag:SECUR32_initNTLMSP ntlm_auth was not found or is outdated. Make sure that ntlm_auth >= 3.0.25 is in your path. Usually, you can find it in the winbind package of your distribution.
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a060, {485e7de8-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a068,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a080, {485e7de9-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a088,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0a0, {485e7dea-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0a8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0c0, {485e7deb-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0c8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0e0, {485e7dec-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0e8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a100, {485e7ded-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a108,): stub
fixme:win:RegisterDeviceNotificationW (hwnd=0x170f00, filter=0x53f56c,flags=0x00000001) returns a fake device notification handle!
wine: Unhandled page fault on read access to 0x45435c6e at address 0x45435c6e (thread 0031), starting debugger...
[thomas@Mithlond] ScanLineRender $ wine ../toolchain/CEdev/bin/ez80cc.exe
err:winediag:SECUR32_initNTLMSP ntlm_auth was not found or is outdated. Make sure that ntlm_auth >= 3.0.25 is in your path. Usually, you can find it in the winbind package of your distribution.
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a060, {485e7de8-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a068,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a080, {485e7de9-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a088,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0a0, {485e7dea-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0a8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0c0, {485e7deb-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0c8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a0e0, {485e7dec-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a0e8,): stub
fixme:advapi:RegisterTraceGuidsW (0x100778a, 0x100a100, {485e7ded-0a80-11d8-ad15-505054503030}, 1, 0x33fe40, (null), (null), 0x100a108,): stub
fixme:win:RegisterDeviceNotificationW (hwnd=0x170f00, filter=0x53f56c,flags=0x00000001) returns a fake device notification handle!
wine: Unhandled page fault on read access to 0x45435c6e at address 0x45435c6e (thread 0009), starting debugger...
Unhandled exception: page fault on read access to 0x45435c6e in 32-bit code (0x45435c6e).
Register dump:
CS:001b SS:0023 DS:0023 ES:0023 FS:1007 GS:000f
EIP:45435c6e ESP:0032f96c EBP:0032fed0 EFLAGS:00010202( R- -- I - - - )
EAX:00000001 EBX:00000000 ECX:ffff0000 EDX:00010000
ESI:00000001 EDI:00530e70
Stack dump:
0x0032f96c: 5c766564 5c6e6962 30387a65 632e6363
0x0032f97c: 02006e6f 02100210 02100210 02100210
0x0032f98c: 02100210 02100210 02100210 02840284
0x0032f99c: 02840284 02840284 02840284 02840284
0x0032f9ac: 02100210 02100210 02100210 03810210
0x0032f9bc: 03810381 03810381 03010381 03010301
0200: sel=1007 base=7ffc0000 limit=00000fff 32-bit rw-
Backtrace:
=>0 0x45435c6e (0x0032fed0)
1 0x7b846d7c in kernel32 (+0x36d7b) (0x0032fee8)
2 0x7b8494fe in kernel32 (+0x394fd) (0x0032ff28)
3 0x7bc5d35c (0x0032ff38)
4 0x7bc5f501 (0x0032ffb8)
5 0x7bc5d322 (0x0032ffd8)
6 0x7bc3c879 (0x0032ffe8)
0x45435c6e: -- no code accessible --
Modules:
Module Address Debug info Name (3 modules)
PE 400000- 417000 Deferred ez80cc
PE 7b810000-7b9ad000 Export kernel32
PE 7bc10000-7bc14000 Deferred ntdll
Threads:
process tid prio (all id:s are in hex)
00000008 (D) D:\Documents\sc2.11_eclipse_workspace\Scanline-Rendering\ScanLineRender\toolchain\CEdev\bin\ez80cc.exe
00000009 0 <==
0000000e services.exe
00000027 0
00000026 0
0000001e 0
00000018 0
00000016 0
00000014 0
00000010 0
0000000f 0
00000012 winedevice.exe
0000001d 0
00000019 0
00000017 0
00000013 0
0000001a plugplay.exe
00000021 0
00000020 0
0000001b 0
00000022 wdfmgr.exe
0000002d 0
0000002c 0
0000002b 0
0000002a 0
00000028 0
00000023 0
00000024 explorer.exe
0000002e 0
00000025 0

It would be more useful to have dbg_printf be an actual function, mimicking printf usage. Keep dbgout as a global variable, and provide a definition that sprintfs to dbgout.

I have a code in a function, with functions call on tice.h and/or graphx.h, and just after that, I have a char pointer declared and a int, like that :
os_EnableCursor(); char *Str = CheckAlloc(calloc(1, MAX_STR_SIZE)), key; unsigned short xPos=0;

That will never be compiled, because the compiler make somes errors, but, when i put os_EnableCursor(); after the pointer and unsigned short declaration, the code will compiling and work perfectly.
I think there is a bug in the compiler or an error in the libs for graphics calculator (but every functions in thoses lib make this bug)
Thanks you

If I #include <ti84pce.h> in multiple files, at link time, the object files for some of them emit duplicate symbol warnings for __frameset:

WARNING (909) (00000000 rb_tree\misc.obj)--> Public symbol "__frameset" already externally defined.
WARNING (909) (00D1D3EE rb_tree\misc.obj)--> Public symbol "__frameset" already externally defined.

If I rewrite these later files with #pragma asm to provide equates for only the labels I need, the warning is not emitted.

These could be useful, though I'm not sure where to put them, so I'll let someone who knows take care of that :P

char* strndup(const char* s, size_t n)
{
  char* result;
  size_t len = strlen(s);

  if (n < len) {
    len = n;
  }

  result = (char*)malloc(len + 1);
  if (!result) {
    return NULL;
  }

  result[len] = '\0';
  return (char*)memcpy(result, s, len);
}

double round(double x) {
    return (x >= 0.0) ? floor(x + 0.5) : ceil(x - 0.5);
}

I'm not so sure about that one (well, it doesn't handle all cases...), but it's fine for most.

Other additions that would be nice: log2, tgamma.

Linking...
WARNING (724) --> Symbol _kb_Scan (obj\main.obj) is not defined.

More and more people are reporting issues on their recent CE (for instance on revisions I, manufactured at least in March 2017, but possibly earlier ones, after E)
For instance in Oiram, the keypad doesn't respond to keypresses anymore ; the toolchain's keypad interrupt demo doesn't work either... , and the interrupt counter demo also freezes.

Needless to say, this breaks a whole lot of things, and it may be hard to do anything about it until we get our hands on a recent CE...

What the heck were you thinking dude.

When ymax and ymin both are 0 in the clip region; and a sprite is drawn with a y value < 0, a single line of pixels is drawn. No pixels being drawn is the desired result.

This should build the shared libraries and link user programs. Convhex can stick around for compression uses I guess.

<debug.h> should provide differing definitions of the functions defined, depending on whether the NDEBUG preprocessor flag is defined. In the case where it is, provide no function prototype, but instead a function-like macro that inserts ((void)0) on use (cf. the definition of assert in <assert.h>)

On Windows, the makefile is not prepared for spaces in names

Solution: surround variable names with quotation marks for the commands/locations

Workaround: clone it into the Public user or create a new user account with no spaces in it.

Thank You
-David M.

I'd prefer having some opinions/reviews (@jacobly in particular?) before committing something or making a PR about all this...

Taking this Makefile as reference:

• Line 28 For Mac compatibility (someone else should try...), and possibly Linux, I had to change it to: WINPATH = $(subst \,\\,$(shell winepath --windows $(1))) otherwise ZDS thought paths with slashes were switches, and with backslashes ones, it was standard escapes. Neither are what we mean.
• Line 34 should be RM = rm --force instead, for more compatibility
• Line 66: $(realpath .) can be simply replaced by . (well, it works fine for me anyway)
• Lines 57, 59, 62, 67: I'm pretty sure the NATIVEPATH should be WINPATH - at least on Mac I had to replace those
• Line 83: I guess || @$(RM) $(OBJECTS) $(TARGET).hex should not remove the objects. Probably only the .hex if you want to clean things up in case of failure
• Line 84-85: such lines shouldn't be in the Makefile at all? (Well, maybe in the clean rule only, but they're there already)
• Line 102: deleting $(TARGET).* is way too dangerous. If a source file has the same name as a target, for instance... $(TARGET).map $(TARGET).hex sounds a bit safer.

Also:

• I couldn't get it to work under Linux with an apparently standard wine setup, it gave me weird paths whatever I tried, which ZDS refused...
• ... but that's not so bad, because I can use ZDS's make.exe under wine, on which it then works fine (and only one call to wine is made, not several as it would apparently be from Linux/Mac)
• For some reason, make -jX is now broken. When trying that (which would be greatly beneficial for multi-files projects), it errors on missing files (related to cstartup.obj, IIRC)

abort() is part of <stdlib.h>, so it would be wise not to redefine it in a separate header.

• macOS 10.12.1-beta (16B2338c),
• latest version of the repository
• GNU Make 3.81
• $PATH: /Users/alexgustafsson/Documents/Apps/Calculator/CEdev/bin
• $CEDEV: /Users/alexgustafsson/Documents/Apps/Calculator/CEdev

Running makeyields various (similar) errors:

Graphics demo 3:

/bin/sh: @\Users\alexgustafsson\Documents\Apps\Calculator\CEdev\bin\eZ80asm: command not found
/bin/sh: src;UsersalexgustafssonDocumentsAppsCalculatorCEdev.: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevinclude: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludeasm: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibce: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibstd: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevlibcegraphx: command not found
make: *** [obj/cstartup.obj] Error 127

Graphics demo 2:

/bin/sh: @\Users\alexgustafsson\Documents\Apps\Calculator\CEdev\bin\eZ80asm: command not found
/bin/sh: src;srcgfx;UsersalexgustafssonDocumentsAppsCalculatorCEdev.: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevinclude: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludeasm: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibce: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibstd: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevlibcegraphx: command not found
make: *** [obj/cstartup.obj] Error 127

Keypad demo 0:

/bin/sh: @\Users\alexgustafsson\Documents\Apps\Calculator\CEdev\bin\eZ80asm: command not found
/bin/sh: src;UsersalexgustafssonDocumentsAppsCalculatorCEdev.: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevinclude: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludeasm: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibce: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevincludelibstd: command not found
/bin/sh: UsersalexgustafssonDocumentsAppsCalculatorCEdevlibcekeypadc: command not found
make: *** [obj/cstartup.obj] Error 127

ti_Write doesn't check if the size of the data = 0, which means it ldir's with BC=0, causing most likely a crash.

It's not great form to have weird include-order issues for client files, especially since <assert.h> is defined by the standard. It should probably include debug.h or something along those lines.

firestar115/TIWaifu

I don't know why it is too big (probably because of my image size), but how in the future (with the changes mentioned in the readme) can I keep it at an acceptable size? Currently the calculator runs out of RAM.

My goal is to recreate my BASIC program but without the small size and refresh delay.

WAIFU.zip

(CEmu was acting up and wouldn't give me my GIF)

It is quite annoying the verification steps that executable must have to run smoothly on windows. Perhaps sign the installation executable with http://www.cacert.org/ or something like that?

It's annoying because I make everything lowercase now. Thoughts?

Following the on-screen instructions for the installer (just selecting a path, I chose the default C:\CEdev), I went to the wiki and to the "Building your first program" section. All I have done is clicked install from the v7.4.1 installer. Running make would produce this error:

C:\CEdev\examples\hello_world>make
makefile:33: /include/.makefile: No such file or directory
make: *** No rule to make target '/include/.makefile'.  Stop.

I was unsure of why $(CEDEV) was undefined from the makefile, so I manually went in to change that. Changing $(CEDEV)/include/.makefile) to /c/CEdev/include/.makefile produced this error:

C:\CEdev\examples\hello_world>make
%%I was unexpected at this time.

and would proceed to open up a new instance of cmd.exe. Other than edit the makefile, I have done nothing to anything in C:\CEdev so the cause of this error alludes me other then it was not setup correctly. Please look into!

Implementing text wrapping would be neat, as would adding commands such as \r and \n to graphx.

I'm thinking a separate function to define the wrap left and right boundaries.

Need to track down exactly what the issues are and fix them.

GRAPHX4 will not run as a result of this
Replacing sin(ANG); and cos(ANG) with 1 in lines 58-59 allows the program to run

One comment says the clip region is inclusive, however the default clip region (0, 0, lcdWidth, lcdHeight) is exclusive.

To make the examples more attractive, let's add some .gif/.png screenshots to the example readmes :)

stddef.h incorrectly defines ptrdiff_t as unsigned int, when in fact it should be be int.

Executing ez80cc.exe from the bin directory, either directly, or through the makefile results in one of Windows "This program has stopped working" dialogs with the following problem details:

Problem signature:
Problem Event Name: APPCRASH
Application Name: ez80cc.exe
Application Version: 0.0.0.0
Application Timestamp: 4ee77814
Fault Module Name: StackHash_0a93
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 00000000
Exception Code: c0000005
Exception Offset: 65644543
OS Version: 6.1.7601.2.1.0.256.48
Locale ID: 1033
Additional Information 1: 0a93
Additional Information 2: 0a93d4dc8b52e4293c6412941ec699fe
Additional Information 3: a466
Additional Information 4: a46638d0902048a75b3c1538b3e88c07