* EXTERNALFORMAT,UNICODE,XCCS: Consistent naming of \PEEKCCODE(.EOL), \BACKCCODE returns the code
It was confusing that the naming of the peek functions was inconsistent with the inccode and backccode functions with respect to the EOLC qualifier. Now uniformaly, the unqualified names (\PEEKCCODE, \BACKCCODE, \INCCODE) do not do any EOL interpretation. The qualified-name functions (\PEEKCCODE.EOLC, \INCCODE.EOLC, \BACKCCODE.EOLC) do EOL interpretation.
Also, the \BACKCCODE interface is changed so that it returns the code corresponding to the bytes that it passed over. This simplifies some of the callers.
* Clients of \PEEKCCODE \BACKCCODE and .EOL
Small changes either to fix inconsistencies (backing, reading, peeking with inconsistent EOL interpretation), updating the names in a few other cases, a few simplifications.
* glitches in ATERM, HARDCOPY updates
* TEDIT, TEXTOFD: Pass FORMAT in PROPS to OPENSTREAM
In TEDIT and OPENTEXTSTREAM. I don't think TEDIT interprets the external format, but at least the info is there
* WINDOW: CREATEW prompts on one line
It was calling PROMPTPRINT twice, which resulted in orphan-looking lines
* EXAMINEDEFS: Propagate textwidth to COMPARETEXT
* COMPARETEXT: propagate height/width of text windows, for linelength
#800
* GREP: Will work with external formats, after FILEPOS update
Line printout could be better, if e.g. it is a Lisp source file, but...
* EXAMINEDEFS: Propagate window width to COMPARETEXT, for LINELENGTH
* GITFNS: added fetch before git log
* PSEUDOHOSTS: Pseudohost-streams are now registered on the PH device, not the target device
* EXTERNALFORMAT, XCCS: Globalvar declaration in wrong place
Also updated \CHECKEOLC macro
* GITFNS: moved git fetch to beginning of prc
* GREP: Propagated OUTSTREAM
