#!/bin/bash
## webput: does my at-website-upload chores then invokes WEBPUT-LL -- E Reimer 2001-Oct;
##
## USAGE:  
##	webput [--copyonly]
## --copyonly -- do only the copying, replicating, etc, skipping sitemap-gen & WEBPUT-LL;
## can be invoked from any of or any subdir of  /er/  /noci/  /debwendon/;
##
## most everything will need customization before being of use to anyone else;


if	[[ $(pwd) == /er*        ]];then	DIR=/er/website
##elif	[[ $(pwd) == /debwendon* ]];then	DIR=/debwendon/website					##2010-10: now done by Will
##elif	[[ $(pwd) == /noci*      ]];then	DIR=/noci/website					##2010-10: now done by Will
else  echo "webput lacks support for dir $(pwd)";  exit 8;
fi


##================================
##  PROLOG-part for all websites
##================================
cd $DIR || exit 9
clean													##get rid of "noise" files, tilde-backups etc
chgsed -q -n "s|site updated:[^;<]*|site updated: $(date +%Y%b%d)|"  $DIR/index.htm			##revise site-updated date


if [[ $DIR == /er* ]];then
##================================
##  webput for ER-website
##================================
webcp /home/ereimer/tools/bird-puzzle.icn					$DIR/programs		##2004-12 for bird-puzzle.htm
webcp /etc/sbin/{mkfavicon,mkfavicon-transparent,img2transparent,jsminify}	$DIR/programs		##2007-04-10 for favicon-making page;  2007-08-23 jsminify
webcp /pix/er-HermanRempelDictionary/{aa-intro-etc,endefns,pddefns}.htm		$DIR/plautdietsch	##2007-06-09 for Plautdietsch-Dictionary+Spellchecker page
webcp /pix/er-Preservings-Experiments/PreservingsIndex-1-13.htm			$DIR			##2007-09-06 for Preservings-INDEX
webcp /pix/pkg/Business-CanadaPost-1.04/lib/Business/CanadaPost.pm		$DIR/programs		##2009-04-04 for SellOnline page
webcp /etc/sbin/CanadaPost-SellOnline-get-quotes				$DIR/programs		##2009-04-04 for SellOnline page
webcp /etc/sbin/{cvt-,}tesseract-*[^~]						$DIR/programs		##2009-08-02 for tesseract-training page - scripts
webcp /pix/pkg/tesseract-Box-TrainingPages-???/???.SOURCEFILE{,.latin1}		$DIR/programs		##2009-08-06 for tesseract-training page - sourcefiles OBSOLETE??
webcp /pix/pkg/tesseract-Box-TrainingPages-???/*zip				$DIR/programs		##2009-08-06 for tesseract-training page - zipfiles
webcp /etc/sbin/{cvt-utf8-to-latin1,cvt-latin1-to-utf8}				$DIR/programs		##2010-10-30: for general-purpose-scripts + tesseract-training
webcp /etc/sbin/find-{anomalous-textfiles,unprintable-meaning-nonASCII}		$DIR/programs		##2009-08: converting-to-cp1252/utf8 tools
webcp /etc/sbin/{diff-by-charset,cvt-textfiles-to-utf8-charset}			$DIR/programs		##2009-08: converting-to-cp1252/utf8 tools
webcp /home/ereimer/.Xmodmap							$DIR/programs		##2009-08: converting-to-cp1252/utf8 tools
webcp /pix/pkg/charsetdetective/charsetdetective-*tar.gz			$DIR/programs		##2009-08: for charsetdetective.htm
webcp /pix/pkg/charsetdetective/ChangeLog		$DIR/programs/charsetdetective-ChangeLog.txt	##2009-08: for charsetdetective.htm;  note the name-change
webcp /etc/sbin/{rename,chg,chgsed,unchg,fullname*SLASH,arr}			$DIR/programs		##2009-09: for general-purpose-scripts page
webcp /etc/sbin/{weblog-search-strings-report,url-decode}			$DIR/programs		##2009-10: for webalizer + general-purpose-scripts
webcp /etc/sbin/{mk-calendar-8up-wallet-sized,mk-calendar-covers}		$DIR/programs		##2009-11: for mk-calendar page
webcp /pix/er-Calendar-Making-2009/defscover.txt				$DIR/programs		##2009-11: for mk-calendar page;  may rethink??
webcp /etc/sbin/{bc,scalepix,annotatepix,datecvt}				$DIR/programs		##2009-11: for general-purpose-scripts + mk-calendar
webcp /beans/beans.zip /books/books.zip						$DIR/programs		##2009-12: for beans.htm + books.htm;  each has 01-EXPORT
webcp /beans/00-README.htm						$DIR/programs/beans-README.htm	##2009-12: for beans.htm;  note the name-change
webcp /books/00-README.htm						$DIR/programs/books-README.htm	##2009-12: for books.htm;  note the name-change
webcp /usr/share/lifelines-ER-reports/related.ll					$DIR/programs	##2010-02: for related.htm + genealogy/index.htm
webcp /usr/share/lifelines-ER-reports/related-NOTES-on-Literature+Algorithm.txt		$DIR/programs   ##2010-02: for related.htm + genealogy/index.htm
webcp /etc/sbin/llexec									$DIR/programs	##2010-02: for programs-index + genealogy/index.htm
webcp /etc/sbin/{WEBLINKCHECK-LL,weblinkcheck.cron}					$DIR/programs	##2010-04: for weblinkcheck.htm
webcp /etc/sbin/{webcommentcheck,SetIntersection,SetSubtract,SetUnion}			$DIR/programs	##2010-04: for general-purpose-scripts + weblinkcheck.htm
webcp /home/ereimer/tools/LatLong-UTM.cpp						$DIR/programs	##2010-08: for LatLong-UTM.htm
webcp /usr/local/bin/LatLong-UTM							$DIR/cgi-bin	##2010-08: for LatLong-UTM online-converter cgi-bin/coordinatecvt
webcp /etc/sbin/{NEWER,OLDER,clean,SubsetOrEqual,chgtr,tolower,todosfile,find-dirs}	$DIR/programs	##2010-10: for general-purpose-scripts + webmaster-tools
webcp /etc/sbin/{bcf,dif,difsed,ATTRIBS,fullname,dateplusdays,dateplusmonths,dateERW}	$DIR/programs	##2010-10: for general-purpose-scripts + webmaster-tools
webcp /etc/sbin/{datecvt-from-Apache,list,counted-uniquify-ordered,counted-uniquify}	$DIR/programs	##2010-10: for general-purpose-scripts + webmaster-tools
webcp /etc/sbin/{webput,WEBPUT-LL,webclean,webmksubsets,webnavreplicate,sitemap-gen}	$DIR/programs	##2010-10: for webmaster-tools page; BEWARE:webput in $DIR/programs==!!==
webcp /etc/sbin/{genByCaption-ER,genOrchidsBySpecies-ER,genFrontpageCandidates-ER}	$DIR/programs	##2010-10: for webmaster-tools page
webcp /etc/sbin/{webcp,INSTALL-new-version-of-nopercart,mk-webalizer-ER-distribution}	$DIR/programs	##2010-10: for webmaster-tools page
webcp /etc/sbin/{WEBGETLOGS.cron,FTP,404report}						$DIR/programs	##2010-10: for webmaster-tools page
webcp /etc/sbin/{webalize-auto,webalize}						$DIR/programs	##2010-10: for webmaster-tools page
webcp /etc/sbin/{bk-auto.cron,bkall,bk,bk-restore,webmv}				$DIR/programs	##2010-10: for webmaster-tools page;  2011-01:bk-restore,webmv
webcp /etc/sbin/WEBVARS-SAMPLE							$DIR/programs/WEBVARS	##2010-10: for webmaster-tools page;  note the name-change!!
webcp /etc/sbin/{nl2sp,sp2nl}								$DIR/programs	##2010-11: for general-purpose-scripts page
webcp /etc/sbin/{sendtoElist,generateElistFromHISTORY+MBRS,ListSubtract-e,fixPeggyMBRS}	$DIR/programs	##2010-11: for webmaster-tools page
webcp /etc/sbin/{subsetMBRStoSMALL,countMBRS,harvest-emailids,uniquify-emailidlist}	$DIR/programs	##2010-11: for webmaster-tools page;  2011-01:countMBRS
webcp /etc/sbin/mk-greetingcard-from-photo						$DIR/programs	##2010-12: for mk-greetingcard.htm page
webcp $DIR/photo-upload/photo-upload.cgi				$DIR/programs/photo-upload.txt	##2010-12: for photo-uploader page;  source is photo-upload dir
D=$DIR/programs; Z=ertools.zip;  if NEWER $D $D/$Z;then							##2010-10: remake ertools.zip zipfile, if needed
  (cd $D; L=$(for F in *;do file $F |egrep -q 'script text executable' && echo "$F"; done)		##2010-10: in SUBSHELL, get L list of program names and
  echo "webput: making $Z"; rm -f zip $Z; zip $Z $L README.htm &>/tmp/web-mk-ertools)			##2010-10:   (still in SUBSHELL) remake zipfile, msgs to tmpfile
fi
(cd $DIR/genealogy; for F in LLrelated*txt;do webcp /pix/er-grandma/$F .;done)				##2010-02: in SUBSHELL, copying for genealogy/index.htm
if [ /pix/er-NOPERCART-ShoppingCartSoftware/nopercart.js -nt $DIR/NOPERCART-DEMO1/nopercart.js ];then	##2008-01-10 install new nopercart, if needed
  echo "webput: running INSTALL-new-version-of-nopercart";  INSTALL-new-version-of-nopercart		##2008-01-10 install nopercart (redo clean??)
fi;  ##chg -n --cmd "expand|todosfile"  $DIR/{,NOPERCART-DEMO*/}{shop*htm,noper*js}			##2008-03-10 shopping-cart files in Windozer-style?? (YANKED)
if NEWER /pix/pkg/webalizer-2.20-03-ER  $DIR/programs/webalizer-ER.zip;then				##2008-12-30: install new webalizer-ER.zip, if needed
   echo "webput: running mk-webalizer-ER-distribution";  mk-webalizer-ER-distribution			##2008-12-30: install webalizer-ER.zip
fi
H=/er/website/nav1-head.txt;  if [[ $H -nt $DIR/programs/navER.css || $H -nt $DIR/programs/navER.js ]];then	##2009-03-22: copy navER.{css,js} from nav1.txt if needed
  echo "webput: replacing $DIR/programs/navER.{css,js}"								##2009-12-24: source $H was /noci, now using /er
  cat $H |sed '0,/Nav-CSS:/d; /^@media/,$c@media print {\n#nav\t{display:none;}\n}' >$DIR/programs/navER.css	##2009-03-22: CSS for navER page
  cat $H |sed '0,/<script>/d; /<\/script>/d'					    >$DIR/programs/navER.js	##2009-03-22: JS  for navER page
fi
D=/usr/lib/locale; L=en_US.CP1252; Z=$DIR/programs/$L.zip; [[ $D/$L -nt $Z ]] && (rm -f $Z; cd $D; zip $Z $L/*)	##2009-08: in SUBSHELL, mk en_US.CP1252.zip w $L/* names
D=/pix/er-Calendar-Making-2009; H=calendar_ca.txt; if [[ $D/$H -nt $DIR/programs/$H ]];then		##2009-08: calendar_ca.txt (pcal) Canadian-Holiday definitions
  echo "webput: replacing $DIR/programs/$H w/o NOCI-specific parts"					##2009-08: msg;  2009-11:added "w/o NOCI-specific parts"
  cat $D/$H |sed '/NOCI-calendar:/,$s|^|##|'					>$DIR/programs/$H	##2009-11: yank NOCI-specific parts while copying...
fi
[[ $1 == --copy* ]] && exit										##support --copyonly option, for ER also skip gen* routines??
genByCaption-ER												##needs clean beforehand
genOrchidsBySpecies-ER
[ /etc/sbin/genFrontpageCandidates-ER -nt $DIR-ls ] && genFrontpageCandidates-ER			##2010-01-05: recreate frontpage-images if needed??


elif [[ $DIR == /debwendon* ]];then
##================================
##  webput for Debwendon-website
##================================
true


elif [[ $DIR == /noci* ]];then
##================================
##  webput for NOCI-website
##================================
for SS in about projects pictures shopping;do								##split about.htm into about-Joining.htm etc;  ditto for...
  for S in $SS-*;do        if [[ $SS.htm -ot $S ]];then echo -n "==WARNING: update to $S will be lost;  Continue??";  read; fi;  done
  REDO=; for S in $SS-*;do if [[ $SS.htm -nt $S ]];then REDO=1; fi; done				##==navreplication to use keep-TS, to avoid spurious REDO==??==
  if [ $REDO ];then echo "running webmksubsets $SS.htm";  webmksubsets $SS.htm; fi			##==webmksubsets splits into subfiles, each gets $SS-Timestamp
done
for NEWS in news20????.htm;do   true; done								##get name of CurrentNewsletter
for FT   in fieldtrips??.htm;do true; done								##get name of CurrentFieldtripSchedule
chgsed "/Current/{s|news20....\.htm|$NEWS|g; s|fieldtrips..\.htm|$FT|g;}"  nav2-body.txt		##revise Current-links in nav2-body.txt;  ==use -p for DEBUG
fi  ##==ENDIF er/debwendon/noci


##================================
##  EPILOG including WEBPUT-LL for all websites
##================================
webnavreplicate												##==replicate Nav-code from index.htm to other webpages
[ -d Thiessen ] && webnavreplicate Thiessen								##(for ER) replicate Thiessen nav-code separately
[[ $1 == --copy* ]] && exit										##support --copyonly option to skip sitemap-gen & WEBPUT-LL
sitemap-gen  $DIR											##was sitemap-gen-Debwendon|sitemap-gen-ER|sitemap-gen-NOCI
WEBPUT-LL  $DIR												##WEBPUT-LL
##==Ping Google etc about newly revised sitemap;  see notes in sitemap-gen==!!==



exit
==================
== OBSOLETE PARTS:
==================
==2009-12-19: YANKED this cvt-to-htm-per-img code on going to CGI-script to generate those pages:  was in ER-section just before genByCaption-ER;
for D in $(find-dirs);do										##2008-11-18: convert to webpage-per-image format
  if [[ $D -ot $DIR-ls ]];then continue; fi								##skip $D if olderthan previous upload;  YANK to reconvert all
  for H in $D/*htm;do  [[ -e $H ]] || continue;
    if [[ $H -ot $DIR-ls ]];then continue; fi								##skip $H if olderthan previous upload;  YANK to reconvert all
    if ! grep -q '<a href=.*<img src=.*[^0-9]1[0-9][0-9].jpg' $H;then continue; fi			##skip $H if not a "thumbnails-to-image" page
    echo "webput: running cvt-to-htm-per-img-ER $DIR/$H";  cvt-to-htm-per-img-ER $DIR/$H		##2008-11-18: convert or re-convert to webpage-per-image format
  done
done


================================
NOTES ABOUT ER-website:
================================
2006may09: changed from only ALT= to only TITLE= in my local copy of HTML files (so Firefox shows them), continuing to have both on the uploaded copy;
	see the /er/chg-alt-to-title script for details & the resulting changes elsewhere.
2006may20: the ereimer directory got big enough to get "disk quota exceeded" response on an upload to shaw;
	the new limit appears to be 25MB;
	da shows the size as 28MB, but had just added 33 photos (06may19) and upload failed beginning with the 16th, of the sz400 images.
2006may20:  BEWARE:  on reimer, am getting strange behaviour: only fails on every 7th, or so, but seem to be back to a 10MB limit??

2007mar27:  for NEW WEBHOSTING at HostExcellence IP:76.162.133.212:
	now upload from dir:website (single dir, but with subdirs);
	use webput-Shaw to upload to Shaw;  07aug31:  used it to upload redirect pages;  07sep02: redid after  mk-new-redirects-ShawSite-to-ereimerDOTnet;
	OLD LOOP:  for DIR in ereimer irhymer reimer reimer? rhj00?; do  ##used to have adames too
	NEW LOOP:  for DIR in website; do
2007mar30:  added sitemap-gen-ER
2007-05:    revised genByCaption & genOrchids scripts for /er/ereimer-->/er/website  (calls to them had been yanked since 07mar27);
2008-09-21: rewrote genByCaption & genOrchids, to use shared Part1;  both now in /etc/sbin, with names ending in "-ER";

index.html-->index.htm  once I'm willing to do do With-Deletion
add an ALT-string onto INPUT TYPE=IMAGE tags -- see NOCI notes below for examples;


================================
NOTES ABOUT Debwendon-website:
================================
2008-05: copy "NAV-CODE" from index.htm to each other webpage;

index.html-->index.htm  once I was willing to do With-Deletion
==UNDOING NAV-SCREWUPS:  for F in /tmp/bknav-[PROJ*htm;do /bin/cp -f $F ./${F##*[};done	##LogoDesignContest-20071217.htm LogoDesignContestWinner.htm TrailDesign.htm Vid01-EKO.htm aboutdebwendon.htm documents.htm donation.htm donationthanks.htm links.htm videos.htm;
==CONSIDER: generating the news+articles.htm page, from the NAV-stuff on frontpage;  since adding in 2 places 'tis painful...
==VICE-VERSA IS BETTER:  (1) can mechanically abbreviate when required;  (2) can keep old stuff on that page, but only the last 10 or so items in the NAV-stuff;


================================
NOTES ABOUT NOCI-website:
================================
2005aug26:  added sitemap-gen;  still needed: notifying google that sitemap has changed (notes are in sitemap-gen)
2008-06-14:  had a nasty accident, doing a WITH-DELETION without being ready for such;
	lost some webalizer stats;  now added the clean-ignoring code, as in {/debwendon,/er}/webput...


================================
NOTES ABOUT ALL websites:
================================
2008-10-30:  combined the separate per-website scripts;
	USAGE + WEBCLEAN:
	debwendon + er versions on $1==clean invoked webclean;  all 3 handled $1==clean by touching $DIR-clean-needed, which is now obsolete;
	consider: all 3 support $1==webclean to invoke webclean;  NOTE: was via /PROJ/webclean (there was no /debwendon/webclean)  which used /etc/sbin/WEBCLEAN;
	have renamed WEBCLEAN-->webclean;  revised it to accept $DIR param;  scrapped /er/webclean;  ==revise webclean to traverse subdirs;
2008-11-01:
	have removed $DIR-clean-needed touching, which was obsoleted by WEBPUT-LL-rewrite to always use With-Deletion;
	have removed NEWERTHAN loop, which was only useful with Multi-Site-ER-at-Shaw;
	have renamed index.html-->index.htm under both /er and /debwendon;
	have added  "|| exit 9"  to cd-commands,  to prevent nasty surprises when a cd-cmd fails  ==good practice to do so in any script making changes...
	WEBPUT-LL-interface:  now using fully-qualified $1 param;  note: WEBPUT-LL now traverses subdirs, as needed for the ER-site;
	sitemap-gen:  have revised interface so can be done in EPILOG part;  ==make it faster by using ls+sed;
	revising site-updated-date in index.htm also now done in EPILOG part;
	clean|clean-quickly:  now have fewer calls (and use only clean-quickly),  made possible by tempfiles for "gen"-routines being under /tmp,  and -n on chg|chgsed;
	TITLE+ALT fixups:  moved into WEBPUT-LL, since needed on subdirs too -- see notes there for more info;
	lftp-mirror-e mystery:  it refuses to clean-up .htaccess~  also reports uploading .htaccess without first reporting that it's removing it;
		is ER-site different?  or are dot-names always a problem for lftp?  (/noci/website-ls also still has a .htaccess~  but debwendon does not);
		To get around hidden-dot-file problem use:  set ftp:list-options -a  (from an online man-page, is also in mine altho less klar);
		TRIED in WEBPUT-LL;  see WEBPUT-LL;  was forced to abandon it!!
	NAV-CODE was crashing Netscape-4.7:  have solved in /debwendon/website/index.htm;  more detail:
		removing nested-ul is one solution, see:  /debwendon/nav-demo-that-works-in-ns47.htm
		tip from http://www.aaelectron.co.uk/desig/part_02.html:
		   replace <link rel=stylesheet href=nav.css> with <style>@import url("nav.css");</style>  <--to hide it from 4.7;
		   ==Hurrah!! that alone solves the 4.7-crash;  still works as intended in Opera, Firefox, IE6;  have so revised  /debwendon/website/index.htm
		   ==appearance in 4.7 is better if nested-ul removed, but don't care to go to that much trouble for an obsolete browser...
	TESTING:  did bkall;  used webput on /debwendon (see notes in WEBPUT-LL), on /noci, and on /er;
	Debwendon-site:  added an ALT-string onto INPUT TYPE=IMAGE tags (see NOCI-notes for examples);
		on ER-site that had already been done;  note: such ALT-strings had been lost due to faulty ALT-removal;
	genByCaption-ER:  now ANON-pixsearch.txt's mod-time unchanged when contents unchanged, to avoid slow upload when unchanged;
	REDIRECTS:  have switched to .htaccess-redirects;  see notes in webmv;
	NAV-CODE:  now supported for all 3;
2008-11-18:
	moved NAV-fixup from prolog to epilog;  moved clean + site-updated-revision from epilog to prolog;  yanked cd $DIRWEBPUT;  seems ok;

==ER-site NEEDS:  reorg photos (htm+jpg) with a dir for each day;  the per-page-Shaw-redirects now obsolete, enough that members.shaw.ca/ereimer has one??
==ER-site NEEDS:  all htm files in subdirs need to be re-uploaded, if they are to have ALT-strings;
==ER-site NEEDS:  auto-revise of "site updated" in /er/website/Thiessen/index.htm -- perhaps any capitalized subdir??
==NOCI+Debwendon-sites NEED:  padding|margin|spacing on table-with-border items -- way too little padding in ie6|ie7 -- answer: specify margin too?
==NOCI-site NEEDS:  rename UNINDEXED, etc (all robots-suppressed names) to ANON-style, in order to simplify tests here, and in sitemap-gen;
	can then simplify /noci/website/robots.txt;  all 3 robots.txt actually;

2009-02-19:  ==SURPRISE, after minor playing with flash-players under rants, next webput redid cvt-to-htm-per-img-ER everywhere==??==
	had done webput a few hours earlier after revising the HostExcellence-rant;
	on preceding day (2009-02-18) after recaptioning one photo - webput did cvt-to-htm-per-img-ER on that htm-file as expected;
	NOTE: the webput failed - am being forced to change FTP-password==!!==
	AHA, there is no /er/website-ls;  do have /er/website-ls-before with ts: Feb 18 21:10
	Suspect I'd already had a failed upload, leaving /er/website-ls nonexistent...
	NOTE: my cron-initiated download of logfiles (see WEBGETLOGS) failed for 20090219+20+21+22+23;  ==HostExcellence's webalized stats also have gap;

2009-02-27:  support Nav-code replication pages having their own STYLE parts;
	shortly after, while adding nav-code to NOCI, made webnavreplicate a separate script;
	on NOCI, the webnavreplicate has been (largely) obsoleted by the use of SSI-includes  (of nav1-head.txt, nav2-body.txt);
	==do likewise on Debwendon + ER==??==
	==Testing is now a pain on NOCI site;  Consider (A) using Apache on my Linux-box;  (B) script to construct navtest-index.htm--??--

2009-07-02: I revised captions in /er/website/20090701/index.htm during an upload, but not in time for the running cvt-to-htm-per-img-ER step;
	then couldn't get that cvt-to-htm-per-img-ER to be redone;
	AHA, need to TOUCH 20090701  as well as 20090701/index.htm==!!==

2009-08-27:  on an "er" upload, the copy of programs/find-anomalous-textfiles was done,  BUT no upload done for programs dir??
2009-08-31:  on an "er" upload, the copy of programs/charsetdetective-2.00.tar.gz + find-anomalous-textfiles was done,  but no upload for programs dir;
	Timestamps:  on programs dir: Aug 31 15:48,  on charsetdetective-2.00.tar.gz: Aug 31 19:37,  on find-anomalous: Aug 31 17:36;
	clearly directory-timestamp-updating is not happening as I've been thinking==!!==  AND YET manpage for open states:  
	If the file is newly created, its atime, ctime, mtime fields are set to the current time, and so are the ctime and mtime fields of the parent directory.
	--that certainly sounds like the way I thought it worked;  ==wrote testing-directory-timestamp to test mtime-updating on parent directory;
2009-08-31:
	wrote NEWER script, and modified WEBPUT-LL to use it for directory timestamp-comparisons;
	also revised "webalizer" test above to use NEWER -- and discovered it had been broken  (-nt gives no errmsg for nonexistent file)
	was surprised by result from NEWER, when I thought I'd used touch to get into a state that would avoid re-installing webalizer:
	-rw-r--r--  1 ereimer 558662 2009-01-06 04:04:28.000000000 -0600 /er/website/programs/webalizer-ER.zip
	drwxr-xr-x  3 ereimer   4096 2009-01-06 04:04:28.000000000 -0600 /pix/pkg/webalizer-2.20-03-ER/
	--AHA: newest subfile was:   2009-01-06 16:50:59.000000000 -0600 Copyright
	--timestamps:  /pix/pkg/webalizer-2.20-03-ER is still: 2009-01-06 04:04:28.000000000;  zipfile is now: 2009-09-01 21:35:27.000000000;
2009-08-31:
	yanked the cater-to-Windozers chg-cmd that revised NOPERCART-DEMO*/{shop*htm,noper*js} files -- expect I'll go to: webpages are Unix, zipfiles msdos...
2009-12-14:
	webcp:  replaced "cp -fpuv" by webcp to WARN if web-version newer;  got 5 such warnings & fixed problems, all cvt-to-utf8 related  (notes are in webcp);
	have long contemplated such a script to: add copyright-notice, check PREREQS (bash only), remove ER-only parts -- detailed notes are now in webcp;
	==needed: automate calling /{beans,books}/01-EXPORT to make beans.zip books.zip when needed -- presently need to run those 01-EXPORT manually;
	==have mixed feelings about automated-installs:  can lead to accidental install when in the midst of nontrivial mods...
	==may need prompts??  eg: my nopercart-install already has prompts to allow certain steps (copy to NOCI) to be skipped;

2009-12-15:  for faster upload (ER site) when doing quicky tests, eg: .htaccess for pixpage,  can yank the sitemap-gen step;
2009-12-19:  YANKED the cvt-to-htm-per-img code (ER-website), on going to pixpage CGI-script to generate those pages;
2010-02-02:  ER-site new file programs/cp1252-er.enc has no "original" (the copy under ~/tools is obsolete);  could extract it from tesseract-training-from-source--??--
2010-03:  removed comment about running webtestcopy (was in EPILOG before sitemap-gen), because it's now totally OBSOLETE;
2010-04-24:
	==NEEDED: add <META NAME="description" CONTENT="..."> to every important webpage;
	otherwise Googlebot uses the menu-entries;  only started very recently: any crawled on 2010-04-14 or earlier didnot have that problem;
	eg: when doing google "site:ereimer.net" see many pages described as:
		ereimer.net. Frontpage · Photos · Recent photos · Search photos · Orchid photos · Rants · Programs · Plautdietsch · Genealogy · About ...
	==webmake: add such META-tag algorithmically -- use the first "Photos from ... on ..." line;
	==webnavreplicate: add algorithmically; and/or grumble if missing;
	==semi-manual revision to all existing pages lacking such META-tag;  on all 3/4 websites;
	==also check for pages lacking <title> -- DONE for ereimer.net where to my surprise nopercart.htm lacked title;
2010-05-08:
	checking bash-scripts being published for missing PREREQS;  eg: 2010-05-08 nichols.m.eric@gmail.com email about scalepix using /noci/com2both;
	(A) for each script, make list of its PREREQS by parsing bash-program -- a lot of work??
	(B) test each script in a shell having std PATH (ie: w/o /etc/sbin, but with /er/website/programs) -- an easier solution==!!==
	(C) make a list of script-names that are in /etc/sbin but not in /er/website/programs, then grep for any occurence of any such script;
	note: plan-B is easy, but results will be incomplete since I won't test every combination of options;  plans-A+C would be complete in that regard;
	note: only plan-A will catch things like /noci/com2both used as a cmd;  hopefully very rare??
	note: none of these plans will catch reference to ER-only datafile eg: /{pix,noci,er}/*;
	note: plan-C is easiest of all, and is more complete than plan-B wrt my own scripts, tho not for non-standard "linux" tools;
		note: comments saying "see also..." -- are easily handled by ignoring entire-line comments, and everything after entire-line "exit";
		to handle all non-standard-Linux tools, include all programs from all PATH-dirs, removing those that are included in my reference distro==??==
		however that sounds near hopeless -- Fedora vs Ubuntu etc -- consider Windows + Cygwin, and additions to my list of Cygwin-Recommendations==??==
	--webtest-bashscripts-ALL:  try a simple test-case for each published script, using:
	--webtest-bashscript: for plan-B, sets PATH excluding /etc/sbin but including /er/website/programs, then executes rest of cmdline;
	--webcheck-bashscripts:  plan-C,  ==DONE!!  checks all files in {general-purpose-scripts,webmaster-tools}.htm for occurances of an unpublished /etc/sbin filename;
2010-10-22:
	BEWARE: webput is now one of the scripts being published;  ergo using webput when in /er/website/programs will use that possibly out-of-date copy!!
2010-10-22:
	==FIXME: a subtle flaw in the copying-from-source strategy, used for programs on ER-website:  when link-revision (301) is done by weblinkcheck.cron, 
	a link-revision comment will be added to the "web" copy, but not to the "source" copy;
	eg: have such link-revision comments in:  programs/beans-README.htm  programs/bird-puzzle.htm  programs/books-README.htm;
	am handling semi-manually, by updating the "source" copy when warning received and no other differences present;
2010-10-22:
	==FIXME:  NEWER $DIR/programs $DIR/programs/webmaster-tools.zip  -- not working, webmaster-tools.zip being remade on every upload==??==
	2010-11-13: has been working recently;  erratic because remade zipfile gets same timestamp as containing dir??  consider: newer-than last upload==??==
2010-10-22:
	discarded the $1==webclean option;  old notes and code for it:
		USAGE:  supports $1==webclean to invoke webclean;  ==!!==AVOID using on NOCI until first saving some old images...
		if [[ $1 == webclean ]];then webclean $DIR;  else clean $DIR; fi  ##webclean or clean;  webclean => remove unreferenced jpg files
	then added --copyonly option to skip sitemap-gen & WEBPUT-LL, ie: do only the copying, replicating, etc;
2010-11-07:
	rewrote the ertools.zip-making code, going from files linked-to from general-purpose + webmaster-tools pages to all bash-files in programs dir:
	from:  W=webmaster-tools; Z=$W.zip; ... L=$(cat {general-purpose-scripts,$W}.htm |sed -n '/^<tr/{s|.*href=||; s|[> \t].*||; p}')
	to:    Z=ertools.zip;               ... L=$(for F in *;do file $F |grep -q 'shell script' && echo "$F"; done)
	zipfile-name:  webmaster-tools.zip-->ertools.zip
	next webput, to my surprise, remade the frontpage-images;  not a mystery after all:  genFrontpageCandidates-ER had been updated, I'd added Author+Date comment;
2010-11-11:
	ertools.zip: also include all Perl scripts in programs dir, eg: CanadaPost-SellOnline-get-quotes;  L=...egrep -q '(perl|shell) script'...
2010-11-19:
	ertools.zip: serious flaw in the making of ertools.zip in that bc was NOT being included--!!--
	(A) consider just checking for first line being a shebang-line??  however shebang-line is NOT required in Linux (in Cygwin?) & file-cmd can diagnose w/o one??
	--Note:
	file bc produces:	a /usr/bin/bc -l script text executable		<--for a "bc-script";
	file webput produces:	Bourne-Again shell script text executable	<--for a bash-script WITH /bin/bash shebang-line;
	file CanadaPost.pm:	a /usr/bin/perl -w script text executable	<--for a Perl-script;
	--Other examples from /etc/sbin, though none of these are "up":
	file redline produces:	POSIX shell script text executable		<--for a shell-script WITH /bin/sh shebang-line;
	file tmp produces:	ASCII text  OR  diff output text  OR  ...	<==for a shell-script lacking shebang-line;  ==file-cmd looks to be useless here==??==
	file encoding.py:	a python script text executable			<--for python script;
	==(B) revising the test, using file-cmd, to be 'script text executable' appears to be the answer;  it was '(perl|shell) script';
	NOTE: using the file-cmd appears to be entirely useless;  it is incapable of recognizing a shebang-less shell-script, which was my only reason for using it:-(
2010-11-23:
	ertools.zip:  now including README.htm;  need a navbar-free subset of it??  Nope, no need, as those SSI-Includes are just HTML-comments w/o Apache;