The first version of \LUATEX\ only had a few extra primitives and it was largely
the same as \PDFTEX. Then we merged substantial parts of \ALEPH\ into the code
and got more primitives. When we got more stable the decision was made to clean
up the rather hybrid nature of the program. This means that some primitives have
been promoted to core primitives, often with a different name, and that others
were removed. This made it possible to start cleaning up the code base. In \in
{chapter} [enhancements] we discussed some new primitives, here we will cover
most of the adapted ones.
Besides the expected changes caused by new functionality, there are a number of
not|-|so|-|expected changes. These are sometimes a side|-|effect of a new
(conflicting) feature, or, more often than not, a change necessary to clean up
the internal interfaces. These will also be mentioned.
\stopsubsection
\startsubsection[title=Changes from \TEX\ 3.1415926]
\topicindex {\TEX}
Of course it all starts with traditional \TEX. Even if we started with \PDFTEX,
most still comes from the original. But we divert a bit.
\startitemize
\startitem
The current code base is written in \CCODE, not \PASCAL. We use \CWEB\ when
possible. As a consequence instead of one large file plus change files, we
now have multiple files organized in categories like \type {tex}, \type
{pdf}, \type {lang}, \type {font}, \type {lua}, etc. There are some artifacts
of the conversion to \CCODE, but in due time we will clean up the source code
and make sure that the documentation is done right. Many files are in the
\CWEB\ format, but others, like those interfacing to \LUA, are \CCODE\ files.
Of course we want to stay as close as possible to the original so that the
documentation of the fundamentals behind \TEX\ by Don Knuth still applies.
\stopitem
\startitem
See \in {chapter} [languages] for many small changes related to paragraph
building, language handling and hyphenation. The most important change is
that adding a brace group in the middle of a word (like in \type {of{}fice})
does not prevent ligature creation.
\stopitem
\startitem
There is no pool file, all strings are embedded during compilation.
\stopitem
\startitem
The specifier \type {plus 1 fillll} does not generate an error. The extra
\quote{l} is simply typeset.
\stopitem
\startitem
The upper limit to \prm {endlinechar} and \prm {newlinechar} is 127.
\stopitem
\startitem
Magnification (\prm {mag}) is only supported in \DVI\ output mode. You can
set this parameter and it even works with \type {true} units till you switch
to \PDF\ output mode. When you use \PDF\ output you can best not touch the
\prm {mag} variable. This fuzzy behaviour is not much different from using
\PDF\ backend related functionality while eventually \DVI\ output is
required.
After the output mode has been frozen (normally that happens when the first
page is shipped out) or when \PDF\ output is enabled, the \type {true}
specification is ignored. When you preload a plain format adapted to
\LUATEX\ it can be that the \prm {mag} parameter already has been set.
\stopitem
\startitem
When \type {\globaldefs} is positive while a local assignment is asked for,
\type {{\global enforced}} is shown in the log when \type {\tracingcommands}
is larger than one. When \type {\globaldefs} is negative and a global
assignment is requested by \type {\global}, \type {\gdef} etc.\ the log will
mention \type {{\global canceled}}.
\stopitem
\stopitemize
\stopsubsection
\startsubsection[title=Changes from \ETEX\ 2.2]
\topicindex {\ETEX}
Being the de factor standard extension of course we provide the \ETEX\
functionality, but with a few small adaptations.
\startitemize
\startitem
The \ETEX\ functionality is always present and enabled so the prepended
asterisk or \type {-etex} switch for \INITEX\ is not needed.
\stopitem
\startitem
The \TEXXET\ extension is not present, so the primitives \type
{\TeXXeTstate}, \type {\beginR}, \type {\beginL}, \type {\endR} and \type
{\endL} are missing. Instead we used the \OMEGA/\ALEPH\ approach to
directionality as starting point.
\stopitem
\startitem
Some of the tracing information that is output by \ETEX's \prm
{tracingassigns} and \prm {tracingrestores} is not there.
\stopitem
\startitem
Register management in \LUATEX\ uses the \OMEGA/\ALEPH\ model, so the maximum
value is 65535 and the implementation uses a flat array instead of the mixed
flat & sparse model from \ETEX.
\stopitem
\startitem
When kpathsea is used to find files, \LUATEX\ uses the \type {ofm} file
format to search for font metrics. In turn, this means that \LUATEX\ looks at
the \type {OFMFONTS} configuration variable (like \OMEGA\ and \ALEPH) instead
of \type {TFMFONTS} (like \TEX\ and \PDFTEX). Likewise for virtual fonts
(\LUATEX\ uses the variable \type {OVFFONTS} instead of \type {VFFONTS}).
\stopitem
\startitem
The primitives that report a stretch or shrink order report a value in a
convenient range zero upto four. Because some macro packages can break on
that we also provide \type {\eTeXgluestretchorder} and \type
{\eTeXglueshrinkorder} which report values compatible with \ETEX. The (new)
\type {fi} value is reported as \type {-1} (so when used in an \type
{\ifcase} test that value makes one end up in the \type {\else}).
\stopitem
\stopitemize
\stopsubsection
\startsubsection[title=Changes from \PDFTEX\ 1.40]
\topicindex {\PDFTEX}
Because we want to produce \PDF\ the most natural starting point was the popular
\PDFTEX\ program. We inherit the stable features, dropped most of the
experimental code and promoted some functionality to core \LUATEX\ functionality
which in turn triggered renaming primitives.
For compatibility reasons we still refer to \type {\pdf...} commands but \LUATEX\
has a different backend interface. Instead of these primitives there are three
interfacing primitives: \lpr {pdfextension}, \lpr {pdfvariable} and \lpr
{pdffeedback} that take keywords and optional further arguments (below we will
still use the \tex {pdf} prefix names as reference). This way we can extend the
features when needed but don't need to adapt the core engine. The front- and
backend are decoupled as much as possible.
\startitemize
\startitem
The (experimental) support for snap nodes has been removed, because it is
much more natural to build this functionality on top of node processing and
attributes. The associated primitives that are gone are: \orm
{pdfsnaprefpoint}, \orm {pdfsnapy}, and \orm {pdfsnapycomp}.
\stopitem
\startitem
The (experimental) support for specialized spacing around nodes has also been
removed. The associated primitives that are gone are: \orm
{pdfadjustinterwordglue}, \orm {pdfprependkern}, and \orm {pdfappendkern}, as
well as the five supporting primitives \orm {knbscode}, \orm {stbscode}, \orm
{shbscode}, \orm {knbccode}, and \orm {knaccode}.
\stopitem
\startitem
A number of \quote {\PDFTEX\ primitives} have been removed as they can be
implemented using \LUA: \orm {pdfelapsedtime}, \orm {pdfescapehex}, \orm
{pdfescapename}, \orm {pdfescapestring}, \orm {pdffiledump}, \orm
{pdffilemoddate}, \orm {pdffilesize}, \orm {pdfforcepagebox}, \orm
{pdflastmatch}, \orm {pdfmatch}, \orm {pdfmdfivesum}, \orm {pdfmovechars},
\orm {pdfoptionalwaysusepdfpagebox}, \orm {pdfoptionpdfinclusionerrorlevel},
\orm {pdfresettimer}, \orm {pdfshellescape}, \orm {pdfstrcmp} and \orm
{pdfunescapehex}.
\stopitem
\startitem
The version related primitives \orm {pdftexbanner}, \orm {pdftexversion}
and \orm {pdftexrevision} are no longer present as there is no longer a
relationship with \PDFTEX\ development.
\stopitem
\startitem
The experimental snapper mechanism has been removed and therefore also the
primitives \orm {pdfignoreddimen}, \orm {pdffirstlineheight}, \orm
{pdfeachlineheight}, \orm {pdfeachlinedepth} and \orm {pdflastlinedepth}.
\stopitem
\startitem
The experimental primitives \lpr {primitive}, \lpr {ifprimitive}, \lpr
{ifabsnum} and \lpr {ifabsdim} are promoted to core primitives. The \type
{\pdf*} prefixed originals are not available.
\stopitem
\startitem
Because \LUATEX\ has a different subsystem for managing images, more
diversion from its ancestor happened in the meantime. We don't adapt to
changes in \PDFTEX.
\stopitem
\startitem
Two extra token lists are provided, \orm {pdfxformresources} and \orm
{pdfxformattr}, as an alternative to \orm {pdfxform} keywords.
\stopitem
\startitem
Image specifications also support \type {visiblefilename}, \type
{userpassword} and \type {ownerpassword}. The password options are only
relevant for encrypted \PDF\ files.
\stopitem
\startitem
The current version of \LUATEX\ no longer replaces and|/|or merges fonts in
embedded \PDF\ files with fonts of the enveloping \PDF\ document. This
regression may be temporary, depending on how the rewritten font backend will
look like.
\stopitem
\startitem
The primitives \orm {pdfpagewidth} and \orm {pdfpageheight} have been removed
because \lpr {pagewidth} and \lpr {pageheight} have that purpose.
\stopitem
\startitem
The primitives \orm {pdfnormaldeviate}, \orm {pdfuniformdeviate}, \orm
{pdfsetrandomseed} and \orm {pdfrandomseed} have been promoted to core
primitives without \type {pdf} prefix so the original commands are no longer
recognized.
\stopitem
\startitem
The primitives \lpr {ifincsname}, \lpr {expanded} and \lpr {quitvmode}
are now core primitives.
\stopitem
\startitem
As the hz and protrusion mechanism are part of the core the related
primitives \lpr {lpcode}, \lpr {rpcode}, \lpr {efcode}, \lpr
{leftmarginkern}, \lpr {rightmarginkern} are promoted to core primitives. The
two commands \lpr {protrudechars} and \lpr {adjustspacing} replace their
prefixed with \type {\pdf} originals.
\stopitem
\startitem
The hz optimization code has been partially redone so that we no longer need
to create extra font instances. The front- and backend have been decoupled
and more efficient (\PDF) code is generated.
\stopitem
\startitem
When \lpr {adjustspacing} has value~2, hz optimization will be applied to
glyphs and kerns. When the value is~3, only glyphs will be treated. A value
smaller than~2 disables this feature. With value of 1, font expansion is
applied after \TEX's normal paragraph breaking routines have broken the
paragraph into lines. In this case, line breaks are identical to standard
\TEX\ behavior (as with \PDFTEX).
\stopitem
\startitem
The \lpr {tagcode} primitive is promoted to core primitive.
\stopitem
\startitem
The \lpr {letterspacefont} feature is now part of the core but will not be
changed (improved). We just provide it for legacy use.
\stopitem
\startitem
The \orm {pdfnoligatures} primitive is now \lpr {ignoreligaturesinfont}.
\stopitem
\startitem
The \orm {pdfcopyfont} primitive is now \lpr {copyfont}.
\stopitem
\startitem
The \orm {pdffontexpand} primitive is now \lpr {expandglyphsinfont}.
\stopitem
\startitem
Because position tracking is also available in \DVI\ mode the \lpr {savepos},
\lpr {lastxpos} and \lpr {lastypos} commands now replace their \type {pdf}
prefixed originals.
\stopitem
\startitem
The introspective primitives \type {\pdflastximagecolordepth} and \type
{\pdfximagebbox} have been removed. One can use external applications to
determine these properties or use the built|-|in \type {img} library.
\stopitem
\startitem
The initializers \orm {pdfoutput} has been replaced by \lpr {outputmode} and
\orm {pdfdraftmode} is now \lpr {draftmode}.
\stopitem
\startitem
The pixel multiplier dimension \orm {pdfpxdimen} lost its prefix and is now
called \lpr {pxdimen}.
\stopitem
\startitem
An extra \orm {pdfimageaddfilename} option has been added that can be used to
block writing the filename to the \PDF\ file.
\stopitem
\startitem
The primitive \orm {pdftracingfonts} is now \lpr {tracingfonts} as it
doesn't relate to the backend.
\stopitem
\startitem
The experimental primitive \orm {pdfinsertht} is kept as \lpr {insertht}.
\stopitem
\startitem
There is some more control over what metadata goes into the \PDF\ file.
\stopitem
\startitem
The promotion of primitives to core primitives as well as the separation of
font- and backend means that the initialization namespace \type {pdftex} is
gone.
\stopitem
\stopitemize
One change involves the so called xforms and ximages. In \PDFTEX\ these are
implemented as so called whatsits. But contrary to other whatsits they have
dimensions that need to be taken into account when for instance calculating
optimal line breaks. In \LUATEX\ these are now promoted to a special type of rule
nodes, which simplifies code that needs those dimensions.
Another reason for promotion is that these are useful concepts. Backends can
provide the ability to use content that has been rendered in several places, and
images are also common. As already mentioned in \in {section}
[sec:imagedandforms], we now have:
There are a few \lpr {pdffeedback} features that relate to this but these are
typical backend specific ones. The index that gets returned is to be considered
as \quote {just a number} and although it still has the same meaning (object
related) as before, you should not depend on that.
The protrusion detection mechanism is enhanced a bit to enable a bit more complex
situations. When protrusion characters are identified some nodes are skipped:
\startitemize[packed,columns,two]
\startitem zero glue \stopitem
\startitem penalties \stopitem
\startitem empty discretionaries \stopitem
\startitem normal zero kerns \stopitem
\startitem rules with zero dimensions \stopitem
\startitem math nodes with a surround of zero \stopitem
\startitem dir nodes \stopitem
\startitem empty horizontal lists \stopitem
\startitem local par nodes \stopitem
\startitem inserts, marks and adjusts \stopitem
\startitem boundaries \stopitem
\startitem whatsits \stopitem
\stopitemize
Because this can not be enough, you can also use a protrusion boundary node to
make the next node being ignored. When the value is~1 or~3, the next node will be
ignored in the test when locating a left boundary condition. When the value is~2
or~3, the previous node will be ignored when locating a right boundary condition
(the search goes from right to left). This permits protrusion combined with for
instance content moved into the margin:
Because we wanted proper directional typesetting the \ALEPH\ mechanisms looked
most attractive. These are rather close to the ones provided by \OMEGA, so what
we say next applies to both these programs.
\startitemize
\startitem
The extended 16-bit math primitives (\orm {omathcode} etc.) have been
removed.
\stopitem
\startitem
The \OCP\ processing has been removed completely and as a consequence, the
following primitives have been removed: \orm {ocp}, \orm {externalocp}, \orm
{ocplist}, \orm {pushocplist}, \orm {popocplist}, \orm {clearocplists}, \orm
{addbeforeocplist}, \orm {addafterocplist}, \orm {removebeforeocplist}, \orm
{removeafterocplist} and \orm {ocptracelevel}.
\stopitem
\startitem
\LUATEX\ only understands 4~of the 16~direction specifiers of \ALEPH: \type
{TLT} (latin), \type {TRT} (arabic), \type {RTT} (cjk), \type {LTL} (mongolian).
All other direction specifiers generate an error. In addition to a keyword
driven model we also provide an integer driven one.
\stopitem
\startitem
The input translations from \ALEPH\ are not implemented, the related
primitives are not available: \orm {DefaultInputMode}, \orm
{noDefaultInputMode}, \orm {noInputMode}, \orm {InputMode}, \orm
{DefaultOutputMode}, \orm {noDefaultOutputMode}, \orm {noOutputMode}, \orm
{OutputMode}, \orm {DefaultInputTranslation}, \orm
{noDefaultInputTranslation}, \orm {noInputTranslation}, \orm
{InputTranslation}, \orm {DefaultOutputTranslation}, \orm
{noDefaultOutputTranslation}, \orm {noOutputTranslation} and \orm
{OutputTranslation}.
\stopitem
\startitem
Several bugs have been fixed and confusing implementation details have been
sorted out.
\stopitem
\startitem
The scanner for direction specifications now allows an optional space after
the direction is completely parsed.
\stopitem
\startitem
The \type {^^} notation has been extended: after \type {^^^^} four
hexadecimal characters are expected and after \type {^^^^^^} six hexadecimal
characters have to be given. The original \TEX\ interpretation is still valid
for the \type {^^} case but the four and six variants do no backtracking,
i.e.\ when they are not followed by the right number of hexadecimal digits
they issue an error message. Because \type{^^^} is a normal \TEX\ case, we
don't support the odd number of \type {^^^^^} either.
\stopitem
\startitem
Glues {\it immediately after} direction change commands are not legal
breakpoints.
\stopitem
\startitem
Several mechanisms that need to be right|-|to|-|left aware have been
improved. For instance placement of formula numbers.
\stopitem
\startitem
The page dimension related primitives \lpr {pagewidth} and \lpr {pageheight}
have been promoted to core primitives. The \prm {hoffset} and \prm {voffset}
primitives have been fixed.
\stopitem
\startitem
The primitives \type {\charwd}, \type {\charht}, \type {\chardp} and \type
{\charit} have been removed as we have the \ETEX\ variants \type
{\fontchar*}.
\stopitem
\startitem
The two dimension registers \lpr {pagerightoffset} and \lpr
{pagebottomoffset} are now core primitives.
\stopitem
\startitem
The direction related primitives \lpr {pagedir}, \lpr {bodydir}, \lpr
{pardir}, \lpr {textdir}, \lpr {mathdir} and \lpr {boxdir} are now core
primitives.
\stopitem
\startitem
The promotion of primitives to core primitives as well as removing of all
others means that the initialization namespace \type {aleph} that early
versions of \LUATEX\ provided is gone.
\stopitem
\stopitemize
The above let's itself summarize as: we took the 32 bit aspects and much of the
directional mechanisms and merged it into the \PDFTEX\ code base as starting
point for further development. Then we simplified directionality, fixed it and
opened it up.
\stopsubsection
\startsubsection[title=Changes from anywhere]
The \type {\partokenname} and \type {\partokencontext} primitives are taken from
the \PDFTEX\ change file posted on the implementers list. They are explained in
the \PDFTEX\ manual and are classified as \ETEX\ extensions.
\stopsubsection
\startsubsection[title=Changes from standard \WEBC]
\topicindex {\WEBC}
The compilation framework is \WEBC\ and we keep using that but without the
\PASCAL\ to \CCODE\ step. This framework also provides some common features that
deal with reading bytes from files and locating files in \TDS. This is what we do
different:
\startitemize
\startitem
There is no mltex support.
\stopitem
\startitem
There is no enctex support.
\stopitem
\startitem
The following encoding related command line switches are silently ignored,
even in non|-|\LUA\ mode: \type {-8bit}, \type {-translate-file}, \type
{-mltex}, \type {-enc} and \type {-etex}.
\stopitem
\startitem
The \prm {openout} whatsits are not written to the log file.
\stopitem
\startitem
Some of the so|-|called \WEBC\ extensions are hard to set up in non|-|\KPSE\
mode because \type {texmf.cnf} is not read: \type {shell-escape} is off (but
that is not a problem because of \LUA's \type {os.execute}), and the paranoia
checks on \type {openin} and \type {openout} do not happen. However, it is
easy for a \LUA\ script to do this itself by overloading \type {io.open} and
alike.
\stopitem
\startitem
The \quote{E} option does not do anything useful.
\stopitem
In a previous section we mentioned that some \PDFTEX\ primitives were removed and
others promoted to core \LUATEX\ primitives. That is only part of the story. In
order to separate the backend specific primitives in de code these commands are
now replaced by only a few. In traditional \TEX\ we only had the \DVI\ backend
but now we have two: \DVI\ and \PDF. Additional functionality is implemented as
\quote {extensions} in \TEX\ speak. By separating more strickly we are able to
keep the core (frontend) clean and stable and isolate these extensions. If for
some reason an extra backend option is needed, it can be implemented without
touching the core. The three \PDF\ backend related primitives are:
\starttyping
\pdfextension command [specification]
\pdfvariable name
\pdffeedback name
\stoptyping
An extension triggers further parsing, depending on the command given. A variable is
a (kind of) register and can be read and written, while a feedback is reporting
something (as it comes from the backend it's normally a sequence of tokens).
\stopsubsection
\startsubsection[title={\lpr{pdfextension}, \lpr {pdfvariable} and \lpr {pdffeedback}},reference=sec:pdfextensions]
In order for \LUATEX\ to be more than just \TEX\ you need to enable primitives. That
has already been the case right from the start. If you want the traditional \PDFTEX\
primitives (for as far their functionality is still around) you now can do this:
The \prm {edef} can also be a \prm {def} but it's a bit more efficient to expand
the lookup related register beforehand.
The backend is derived from \PDFTEX\ so the same syntax applies. However, the
\type {outline} command accepts a \type {objnum} followed by a number. No
checking takes place so when this is used it had better be a valid (flushed)
object.
In order to be (more or less) compatible with \PDFTEX\ we also support the option
to suppress some info but we do so via a bitset:
In addition you can overload the trailer id, but we don't do any checking on
validity, so you have to pass a valid array. The following is like the ones
normally generated by the engine. You even need to include the brackets here!
Although we started from a merge of \PDFTEX\ and \ALEPH, by now the code base as
well as functionality has diverted from those parents. Here we show the options
that can be passed to the extensions. The \type {shipout} option is a compatibility
feature. Instead one can use the \type {deferred} prefix.
\starttexsyntax
\pdfextension literal
[shipout] [ direct | page | raw ] { tokens }
\stoptexsyntax
\starttexsyntax
\pdfextension dest
num integer | name { tokens }!crlf
[ fitbh | fitbv | fitb | fith| fitv | fit |
fitr <rule spec> | xyz [ zoom <integer> ]
\stoptexsyntax
The directional model in \LUATEX\ is inherited from \OMEGA|/|\ALEPH\ but we tried
to improve it a bit. At some point we played with recovery of modes but that was
disabled later on when we found that it interfered with nested directions. That
itself had as side effect that the node list was no longer balanced with respect
to directional nodes which in turn can give side effects when a series of dir
changes happens without grouping.
When extending the \PDF\ backend to support directions some inconsistencies were
found and as a result we decided to support only the four models that make sense
\type {TLT} (latin), \type {TRT} (arabic), \type {RTT} (cjk) and \type {LTL}
(mongolian).
\stopsubsection
\startsubsection[title={How it works}]
The approach is that we again make the list balanced but try to avoid some side
effects. What happens is quite intuitive if we forget about spaces (turned into
glue) but even there what happens makes sense if you look at it in detail.
However that logic makes in|-|group switching kind of useless when no proper
nested grouping is used: switching from right to left several times nested,
results in spacing ending up after each other due to nested mirroring. Of course
a sane macro package will manage this for the user but here we are discussing the
low level dir injection.
This is what happens:
\starttyping
\textdir TRT nur {\textdir TLT run \textdir TRT NUR} nur
\stoptyping
This becomes stepwise:
\startnarrower
\starttyping
injected: [+TRT]nur {[+TLT]run [+TRT]NUR} nur
balanced: [+TRT]nur {[+TLT]run [-TLT][+TRT]NUR[-TRT]} nur[-TRT]
result : run {RUNrun } run
\stoptyping
\stopnarrower
And this:
\starttyping
\textdir TRT nur {nur \textdir TLT run \textdir TRT NUR} nur
\stoptyping
becomes:
\startnarrower
\starttyping
injected: [+TRT]nur {nur [+TLT]run [+TRT]NUR} nur
balanced: [+TRT]nur {nur [+TLT]run [-TLT][+TRT]NUR[-TRT]} nur[-TRT]
result : run {run RUNrun } run
\stoptyping
\stopnarrower
Now, in the following examples watch where we put the braces:
\startbuffer
\textdir TRT nur {{\textdir TLT run} {\textdir TRT NUR}} nur
\stopbuffer
\typebuffer
This becomes:
\startnarrower
\getbuffer
\stopnarrower
Compare this to:
\startbuffer
\textdir TRT nur {{\textdir TLT run }{\textdir TRT NUR}} nur
\stopbuffer
We could define the two helpers to look back, pick up a skip, remove it and
inject it after the dir node. But that way we loose the subtype information that
for some applications can be handy to be kept as|-|is. This is why we now have a
variant of \lpr {textdir} which injects the balanced node before the skip.
Instead of the previous definition we can use:
Anything more complex that this, like combination of skips and penalties, or
kerns, should be handled in the input or macro package because there is no way we
can predict the expected behaviour. In fact, the \lpr {linedir} is just a
convenience extra which could also have been implemented using node list parsing.
Directions are complicated by the fact that they often need to work over groups
so a separate grouping related stack is used. A side effect is that there can be
paragraphs with only a local par node followed by direction synchronization
nodes. Paragraphs like that are seen as empty paragraphs and therefore ignored.
Because \type {\noindent} doesn't inject anything but a \type {\indent} injects
an box, paragraphs with only an indent and directions are handled as paragraphs
with content.
By default paragraphs before a display equation containing dir nodes are never ignored.
Changing that could break existing documents, but when you set \lpr {mathemptydisplaymode}
to~\type {1} empty paragraphs before a display equation will be ignored.
\stopsubsection
\startsubsection[title={Controlling glue with \lpr {breakafterdirmode}}]
Glue after a dir node is ignored in the linebreak decision but you can bypass that
by setting \lpr {breakafterdirmode} to~\type {1}. The following table shows the
difference. Watch your spaces.
\startsubsection[title={Controling parshapes with \lpr {shapemode}}]
Another adaptation to the \ALEPH\ directional model is control over shapes driven
by \prm {hangindent} and \prm {parshape}. This is controlled by a new parameter
\lpr {shapemode}:
\starttabulate[|c|l|l|]
\DB value \BC \prm {hangindent} \BC \prm {parshape} \NC \NR
\TB
\BC \type{0} \NC normal \NC normal \NC \NR
\BC \type{1} \NC mirrored \NC normal \NC \NR
\BC \type{2} \NC normal \NC mirrored \NC \NR
\BC \type{3} \NC mirrored \NC mirrored \NC \NR
\LL
\stoptabulate
The value is reset to zero (like \prm {hangindent} and \prm {parshape})
after the paragraph is done with. You can use negative values to prevent
this. In \in {figure} [fig:shapemode] a few examples are given.
Internally the implementation is different from \ALEPH. First of all we use no
whatsits but dedicated nodes, but also we have only 4 directions that are mapped
onto 4 numbers. A text direction node can mark the start or end of a sequence of
nodes, and therefore has two states. At the \TEX\ end we don't see these states
because \TEX\ itself will add proper end state nodes if needed.
The symbolic names \type {TLT}, \type {TRT}, etc.\ originate in \OMEGA. In
\LUATEX\ we also have a number based model which sometimes makes more sense.
We support the \OMEGA\ primitives \orm {textdir}, \orm {pardir}, \orm {pagedir},
\orm {pardir} and \orm {mathdir}. These accept three character keywords. The
primitives that set the direction by number are: \lpr {textdirection}, \lpr
{pardirection}, \lpr {pagedirection} and \lpr {bodydirection} and \lpr
{mathdirection}. When specifying a direction for a box you can use \type {bdir}
instead of \type {dir}.
\stopsubsection
\stopsection
\startsection[title=Implementation notes]
\startsubsection[title=Memory allocation]
\topicindex {memory}
The single internal memory heap that traditional \TEX\ used for tokens and nodes
is split into two separate arrays. Each of these will grow dynamically when
needed.
The \type {texmf.cnf} settings related to main memory are no longer used (these
are: \type {main_memory}, \type {mem_bot}, \type {extra_mem_top} and \type
{extra_mem_bot}). \quote {Out of main memory} errors can still occur, but the
limiting factor is now the amount of RAM in your system, not a predefined limit.
Also, the memory (de)allocation routines for nodes are completely rewritten. The
relevant code now lives in the C file \type {texnode.c}, and basically uses a
dozen or so \quote {avail} lists instead of a doubly|-|linked model. An extra
function layer is added so that the code can ask for nodes by type instead of
directly requisitioning a certain amount of memory words.
Because of the split into two arrays and the resulting differences in the data
structures, some of the macros have been duplicated. For instance, there are now
\type {vlink} and \type {vinfo} as well as \type {token_link} and \type
{token_info}. All access to the variable memory array is now hidden behind a
macro called \type {vmem}. We mention this because using the \TEX book as
reference is still quite valid but not for memory related details. Another
significant detail is that we have double linked node lists and that most nodes
carry more data.
The input line buffer and pool size are now also reallocated when needed, and the
\type {texmf.cnf} settings \type {buf_size} and \type {pool_size} are silently
ignored.
\stopsubsection
\startsubsection[title=Sparse arrays]
The \prm {mathcode}, \prm {delcode}, \prm {catcode}, \prm {sfcode}, \prm {lccode}
and \prm {uccode} (and the new \lpr {hjcode}) tables are now sparse arrays that
are implemented in~\CCODE. They are no longer part of the \TEX\ \quote
{equivalence table} and because each had 1.1 million entries with a few memory
words each, this makes a major difference in memory usage. Performance is not
really hurt by this.
The \prm {catcode}, \prm {sfcode}, \prm {lccode}, \prm {uccode} and \lpr {hjcode}
assignments don't show up when using the \ETEX\ tracing routines \prm
{tracingassigns} and \prm {tracingrestores} but we don't see that as a real
limitation.
A side|-|effect of the current implementation is that \prm {global} is now more
expensive in terms of processing than non|-|global assignments but not many users
will notice that.
The glyph ids within a font are also managed by means of a sparse array as glyph
ids can go up to index $2^{21}-1$ but these are never accessed directly so again
users will not notice this.
Single|-|character commands are no longer treated specially in the internals,
they are stored in the hash just like the multiletter csnames.
The code that displays control sequences explicitly checks if the length is one
when it has to decide whether or not to add a trailing space.
Active characters are internally implemented as a special type of multi|-|letter
control sequences that uses a prefix that is otherwise impossible to obtain.
\stopsubsection
\startsubsection[title=The compressed format file]
\topicindex {format}
The format is passed through \type {zlib}, allowing it to shrink to roughly half
of the size it would have had in uncompressed form. This takes a bit more \CPU\
cycles but much less disk \IO, so it should still be faster. We use a level~3
compression which we found to be the optimal trade|-|off between filesize and
decompression speed.
\stopsubsection
\startsubsection[title=Binary file reading]
\topicindex {files+binary}
All of the internal code is changed in such a way that if one of the \type
{read_xxx_file} callbacks is not set, then the file is read by a \CCODE\ function
using basically the same convention as the callback: a single read into a buffer
big enough to hold the entire file contents. While this uses more memory than the
previous code (that mostly used \type {getc} calls), it can be quite a bit faster
(depending on your \IO\ subsystem).
\stopsubsection
\startsubsection[title=Tabs and spaces]
\topicindex {space}
\topicindex {newline}
We conform to the way other \TEX\ engines handle trailing tabs and spaces. For
decades trailing tabs and spaces (before a newline) were removed from the input
but this behaviour was changed in September 2017 to only handle spaces. We are
aware that this can introduce compatibility issues in existing workflows but
because we don't want too many differences with upstream \TEXLIVE\ we just follow
up on that patch (which is a functional one and not really a fix). It is up to
macro packages maintainers to deal with possible compatibility issues and in
\LUATEX\ they can do so via the callbacks that deal with reading from files.
The previous behaviour was a known side effect and (as that kind of input
normally comes from generated sources) it was normally dealt with by adding a
comment token to the line in case the spaces and|/|or tabs were intentional and
to be kept. We are aware of the fact that this contradicts some of our other
choices but consistency with other engines and the fact that in \KPSE\ mode a
common file \IO\ layer is used can have a side effect of breaking compatibility.
We still stick to our view that at the log level we can (and might be) more
incompatible. We already expose some more details.
\stopsubsection
\startsubsection[title=Hyperlinks]
\topicindex {hyperlinks}
There is an experimental feature that makes multi|-|line hyper links behave a
little better, fixing some side effects that showed up in r2l typesetting but
also can surface in l2r. Because this got unnoticed till 2023, and because it
depends bit on how macro packages deal with hyper links, the fix is currently
under parameter control:
\starttyping
\pdfvariable linking = 1
\stoptyping
That way (we hope) legacy documents come out as expected, whatever those
expectations are. One of the aspects dealt with concerns (unusual) left and right
skips.
