The occurrences of T6 and T1 have been corrected. T6 now only occurs when a BRK
instruction is executing, it is a synonym for when the VEC1 node is logic high.
T1 now occurs when node 862 is logic high, which drives the SYNC pin, among
other control effects.
Formerly, T1 and T6 were displayed only when all the nodes that affect the PLA
were inactive. Node 1357's state was used in that case to choose between
displaying T1 (1357 high) or T6 (1357 low). That turned out to be incorrect
pair of inferences. The result was that T1 was absent when it should have been
present (when T+ was present without T0), and T6 was present when it should
have been absent (for instructions other than BRK). Among the corrective
changes, node 1357 is no longer consulted.
Expansion of state display adds V0, SD1, and SD2 indications. The last two are
in their own field (an eighth field). V0 is in the seventh field (square
bracketed, the same field occupied by T1 and T6). It is a two-character
representation of node VEC0 being at logic high. VEC0 high causes VEC1 to be
high one cycle later, which is T6. V0 is, like T6, activated only by execution
of a BRK instruction.
SD1 and SD2 are nodes 440 and 1258 respectively, which lie in the RCL block
outside of the timing generation (clock) block. They serve the needs of the
RMW (Read-Modify-Write) instructions for their addressing modes that use
external memory instead of the accumulator.
All of the corrected and new features has increased the total number of
displayed states to 24 from 10. Hopefully, this is the ultimate, final,
most fully comprehensive clock display possible, but we'll see.
The parameter to allTCStates() has had a name change from useNBSP to useHTML.
busToString() has had the TStateF pseudobus removed, and a pseudobus called
Phi added.
Phi pretty-prints the state of the cp1 node (the internal phase 1 node) as
Φ1 when high and as Φ2 when low. This is nicer than plain 0 or 1.
Logging the TState pseudobus followed by the Phi pseudobus creates the effect
intended by the removed TStateF pseudobus, keeping our collection of
operations orthogonal (no proliferation of trivially different operations
that duplicate each other diagonally).
'TState' returns the string returned by allTCStates().
'TStateF' returns the string returned by allTCStates() with phase indication
appended: "F1" or "F2" for phase1 or phase2 respectively. 'F' is chosen since
that's all that Phi really is: an "eff".
Perhaps another enhancement could call out an actual phi glyph for the HTML
(browser) context.
allTCStates() is also enhanced by an optional boolean parameter to control the
spacing of non-blank text within the string it returns.
* A true parameter puts HTML non-breaking spaces between non-blank text,
suitable for display on a browser page.
This is always used by busToString().
* A false or absent parameter puts ASCII spaces between non-blank text,
suitable for text logging.
that creates the text expression of the time codes seen in the documentation
to macros.js and expert-allinone.js.
Added documentation file: 6502timecodes.txt
Added function: allTCStates()
The comments with them indicate that the semantics of the node (high when branch
not taken) was already recognized.
pipeBRtaken in the expert version is left untouched, as it is opposite-valued
from #BRtaken, although its meaning is multiplexed by additional influences
(nodenames.js has it named pipeIPCrelated).
p6 and p7 were duplicates of their respective Pout-named nodes instead of
the phase 1 updated nodes of their storage complexes, as all the other
processor status register bits are (except for the B bit, of course).
Comment typo correction for cclk.
Patches for the general chipsim code and the 6502 emulation.
For the general chipsim code, allNodes() was constructing a list of
numeric strings for the node indexes instead of a list of numbers for them.
During the first iteration inside recalcNodeList(), the numeric string node
indexes would end up being the first elements of their respective node group
lists. This in turn would allow instances of a duplicate node index to be added
to the node group list. The duplicate in each case would be the numeric
equivalent of the initial numeric string index: indexOf() would not recognize
the string element as being the equivalent of the numeric node index, so the
numeric version would be added also.
For an example (from real log data), a node group list of: [ "49", 483 ] has a
node index of 49 tested against it. indexOf() says "49" is not an occurrence
of 49, so it allows 49 to be
Ed Spittles
mmfoerster