See kludge at bottom.

pdbHeader was gathered by messaging the header with
	beginFetchPDBHeader() (in msgcall.js)                                       #3:
		which runs in JSmol
		makeGetPDBFileHeaderSpt() (scripts.js) which
			collectMessages = true, then "show pdbheader" bracketed by messageMarkBeginHeader and
			messageMarkEndHeader.

When messageMarkEndHeader is received in jmolMessageCallback()                #34
	it calls processPDBHeader() (msgcall.js)                                    #7a
	which sets collectMessages = false, then using *** pdbHeader ***
	getPIPEBlock() which does lots of things if pipeBlock exists
	and if telAvivMode does lots more, then
		makeConsurfHelp()                      <<<<
	
	if !telAvivMode, setupPanels().          <<<<

------------------
RoseTTAFold detection -> pdbHeaderToAtom:
	(See comments in jgetfi() in moltab.js)
	In RoseTTAFold PDB files, MODEL 1 precedes the REMARK containing "RoseTTAFold".
	jgetfi(key) used jmolGetPropertyAsString("fileInfo", key) stops at MODEL 1,
	thereby preventing RoseTTAFold from being detected.

	This required getHeaderToAtom() (moltab.js) which gets *** pdbHeaderToAtom ***
	which stops at the first ATOM or HETATM record.

Thus, in FG 4.6, we had two header variables, pdbHeader and pdbHeaderToAtom.

---------------
startupPanels() (controls.js) calls getHeaderBlocks() (moltab.js)             #64
	which immediately calls getHeaderToAtom() (moltab.js)
		generating *** pdbHeaderToAtom ***
		and which decides whether pdbNoHeader.


---------------
In FG 4.7:

	In jgetfi(), jmolGetPropertyAsString("fileInfo", key)	FAILED with noFormulPDB.
		(It seemed to be looking for a file on disk via XMLHttpRequest.open.)
	Therefore I replaced it with getHeaderRecords(key) which uses pdbHeaderToAtom.

---------------
BUG in FG 4.7 (consurf in temperature):

1sva (not consurf) initially displays capsid with no bug.
	Then showing the AU also no bug.

1sva-1-consurf-colab.pdb initially displays the AU with no bug.

	BUG:
	Then showing the capsid MESSAGES pdbHeaderToAtom (WHERE? WHY?).
	Since REMARK 3 in the header has multiple lines containing "ERROR",
	this generates a msgcall alert.

	The bug occurs between msgcall #3 (beginFetchPDBHeader())
	and msgcall #34 (receipt of messageMarkEndHeader)
	thus while collectMessages is true.

	After spending several hours trying to find where and how pdbHeaderToAtom
	gets messaged to JSmol, I used a kludge:
	msgcall.js looks for "header to atom" and when present, suppresses the error alert.