TODO: move gse/itf_parser to python/itf_parser to have all scripts in one subdir.
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5342 bc5caf13-1734-44f8-af43-603852e9ee25
hetept_config.l3: no longer needs to be python-compatible
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5335 bc5caf13-1734-44f8-af43-603852e9ee25
hetept_calib.py: added a class to hold all names as variables...
hetept_dps.py: ... so that they can be loaded here
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5334 bc5caf13-1734-44f8-af43-603852e9ee25
both filenames on teh commandline and .include files are searched in .path.
option -Ipath
-q resets verbosity to 0
.compile() returns self
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5328 bc5caf13-1734-44f8-af43-603852e9ee25
Currently, the old one-pass mode is still default.
Use option -2 for 2-pass mode.
Use option -1 for 1-pass mode, when the default changed.
New option -F for the name of the forwardfile
if given, the name in the code is ignored with a warning.
In 2-pass mode, no forwardfile is read, nor written.
The compiler issues the second pass only when there are forwardconflicts.
In 1-pass mode, the forwardfile is written only when there are forwardconflicts.
All file output, except for the log, is now written at the end of the last pass.
Default output format is "none", i.e., no output will be generated.
Optiuon -g accepts an optional format string
-g 'hex,printf-format-string'
-g 'dis,l3dis-format-string'
-g hex requires a printf format for (addr,instr)
"@%03x %08x\n"
-g dis requires an l3dis format string.
"@%(addr)-3d %(cond)-3s %(mnem)-5s %(args)s\n"
-g verilog is the same as -g hex, with a different default format
" l3_code(10'h %03x, 32'h %08x);\n"
-g python is again the same as hex, but with an additinal file header.
" l3_code(0x%03x, 0x%08x)\n"
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5324 bc5caf13-1734-44f8-af43-603852e9ee25
class l3
class include_stack
TODO: output into self.l3code[]
TODO: class expression_parser_context
self.EC = expression_parser_context(self.names)
s,v = self.EC.parse_EXPR(s)
TODO? class lexer
those s become lexer instances
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5322 bc5caf13-1734-44f8-af43-603852e9ee25
It has a congiguration memory.
It can connect to units.
It can execute service (128,10?) telecommands.
It successfully runs the FM1 v0001->v0002 patch procedure.
It has not been tested with real units, yet.
See update_fm1_v1_v2.py for how it works!
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5312 bc5caf13-1734-44f8-af43-603852e9ee25
add .copy() to configtable; and itemplacement
add configtable.split()
allow configtable.patch_map() to split a message_index
add configtable.upload_index() to return a list of message_index
which write the table into EEPROM
all this makes no sense, because a table never needs to be patched to write itself :-)
TODO: read a .txt into a table and go from there.
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5308 bc5caf13-1734-44f8-af43-603852e9ee25
lex() returns a token() object now.
That makes for a lot more readable code.
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5297 bc5caf13-1734-44f8-af43-603852e9ee25
disallow ',' as separator for ADDI, ADD, SUB, HIST.
allow only matching operator with ADD and SUB
fix unary operators, allow unary +
documentation fixes
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5296 bc5caf13-1734-44f8-af43-603852e9ee25
catch all exceptions
filename and linennumber added in except clause
also print input line
-h usage()
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5292 bc5caf13-1734-44f8-af43-603852e9ee25
expression parsers:
EEXPR: [!] [-] [ (EXPR) | R[EXPR] | ID | $FUNC(EXPR) | int | float ]
MEXPR: EEXPR [ { * | / } EEXPR ]
AEXPR: MEXPR [ { + | - } MEXPR ]
EXPR: AEXPR [ { << | >> } AEXPR ]
Also fixed MULI so it accepts:
MULI R[x] * AEXPR
MULI R[x] * AEXPR >> AEXPR
In the second case both values have to be integers in the
correct ranges.
ep.l3: Changed the one instance in all our l3 files, where the above
change caused a problem
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5288 bc5caf13-1734-44f8-af43-603852e9ee25
MULI R[.-1] * 5 >> 5
i.e. where the factor is below 1, would result in a multipier
of 0, if shift value was explicitly given.
l3dis.py: changed output for GOTO to GOTO R[x]
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5276 bc5caf13-1734-44f8-af43-603852e9ee25
This allows the parser to distinguish between Registers and
values and thus detect errors.
Register objects are constructed by:
R[x], where x can be an ID, expression or Register
ID = [COND] INSTR
.
Register +- i, where i is an integer
In effect everything should work as before, only the PHA
and GOTO addresses need to be defined like:
HIT_0 = R[0xe0] or GOTO R[0x002]
step.l3, step_pha.l3, het_pen.l3, hetept_pha.l3:
made the (few) necessary changes as above;
The *_l3.py files are identical to before.
git-svn-id: svn+ssh://asterix/home/subversion/stephan/solo/eda/python@5275 bc5caf13-1734-44f8-af43-603852e9ee25
