path: root/Master/tlpkg/bin/tlpkg-ctan-check

#!/usr/bin/env perl
# $Id$
# Public domain.  Originally written 2005, Karl Berry.
# Check if a package in TL has any changes on CTAN.

BEGIN {
  chomp ($mydir = `dirname $0`);  # we are in Master/tlpkg/bin
  unshift (@INC, "$mydir/..");
}

use TeXLive::TLConfig qw/$RelocPrefix $RelocTree/;
use TeXLive::TLPOBJ;
use TeXLive::TLPDB;

use File::Basename;

my $tlpdb;
my $Master;
our %OPT;

# toptesi - reinstate after next upload, generated files out of sync
# these packages should be up to date:
my @TLP_working = qw(
  12many 2up
  AkkTeX Asana-Math ESIEEcv HA-prosper IEEEconf IEEEtran
    SIstyle SIunits Tabbing Type1fonts
  a0poster a2ping a4wide a5comb
    aastex abbr abc abstract abstyles accfonts achemso
    acmconf acmtrans acronym acroterm
    active-conf addlines adforn adfsymbols adobemapping adrconv advdate
    ae aeguill afthesis
    aguplus aiaa aichej akletter
    alg algorithm2e algorithmicx algorithms alnumsec alterqcm
    altfont ametsoc amsaddr amscls amsfonts amslatex-primer amsldoc-it
    amsmath amsrefs amstex amsthdoc-it 
    animate anonchap antt answers ANUfinalexam anyfontsize anysize
    aomart apa apacite apalike2 appendix apprends-latex
    arabi arabtex arabxetex archaic arcs arev
    around-the-bend arphic arrayjobx arsclassica
    arydshln
    asaetr ascelike ascii ascii-chart assignment astro asyfig
    asymptote-faq-zh-cn asymptote-by-example-zh-cn asymptote-manual-zh-cn
    attachfile
    augie auncial-new aurical authoraftertitle authorindex
    auto-pst-pdf autoarea automata
  babel babelbib background bangtex barcodes barr bartel-chess-fonts baskervald
    bbcard bbding bbm bbm-macros bbold bbold-type1 bclogo
    beamer beamer-FUBerlin beamer-tut-pt beamerposter beamerthemenirma
    beebe begriff bera berenisadf betababel beton
    bez123 bezos bgreek
    bib-fr bibarts bibhtml
    biblatex biblatex-apa biblatex-chem biblatex-chicago
    biblatex-dw biblatex-historian biblatex-mla biblatex-nature
    biblatex-philosophy biblatex-science
    bibleref biblist bibtex bibtopic
    bibtopicprefix bibexport bibunits 
    bidi bigfoot bigints binomexp biocon bizcard
    blacklettert1 blindtext blkarray block blockdraw_mp blowup
    bodegraph boisik bold-extra
    boites boldtensors bookest booklet booktabs boolexpr bophook
    borceux bosisio
    boxedminipage boxhandler bpchem bpolynomial
    bracketkey braille braket breakcites breakurl
    bullcntr bundledoc bussproofs bytefield 
  c-pascal cachepic calctab calligra calrsfs cals calxxxx-yyyy cancel
    capt-of captcont captdef caption carlisle carolmin-ps cases casyl
    catchfilebetweentags catechis
    cbcoptic cbfonts
    cc-pl ccaption ccfonts ccicons cclicenses
    cd cd-cover cdpbundl
    cell cellspace censor cfr-lm
    changebar changelayout changepage changes chappg chapterfolder
    chbibref chem-journal chemarrow chembst chemcompounds chemcono chemfig
    chemstyle cherokee
    chess chess-problem-diagrams chessboard chessfss chextras
    chicago chicago-annote chletter chngcntr chronology
    circ circle circuitikz
    cite citeref cjhebrew cjk cjkpunct classicthesis
    clefval cleveref
    clock clrscode cm-super cm-unicode
    cmap cmcyr cmdstring cmdtrack cmll cmpica cmsd cmtiup
    codedoc codepage
    collref colordoc colorinfo colorsep colortab colortbl colorwav
    combelow combine combinedgraphics comfortaa comma commath comment
    compactbib
    complexity comprehensive computational-complexity
    concmath concmath-fonts concprog confproc constants
    context-account context-algorithmic context-bnf context-chromato 
    context-construction-plan context-degrade context-filter context-fixme
    context-french context-fullpage context-games context-gnuplot
    context-letter context-lettrine context-lilypond context-mathsets
    context-notes-zh-cn context-ruby
    context-simplefonts context-simpleslides
    context-top-ten
    context-typearea context-typescripts context-vim
    contour
    cooking cookybooky cool coollist coolstr cooltooltips
    coordsys courier-scaled courseoutline coursepaper coverpage covington
    cprotect
    crop crossreference crossword crosswrd cryst csbulletin csquotes
    csvsimple csvtools ctable ctex ctex-faq
    cursolatex cuisine
    currfile currvita curve curve2e curves
    custom-bib cutwin cv cweb-latex cyklop cyrillic
  dancers dashbox dashrule dashundergaps datatool
    dateiliste datenumber datetime
    dblfloatfix dcpic de-macro decimal dehyph-exptl
    delimtxt dfgproposal
    diagmac2 diagnose dichokey dictsym digiconfigs din1505
    dinat dingbat directory dirtree dirtytalk disser dk-bib dlfltxb
    dnaseq doc-pictex docmfp docmute doi doipubmed
    dot2texi dotarrow dotseqn dottex
    doublestroke dox dozenal dpfloat dprogress drac draftcopy
    draftwatermark dramatist dratex drawstack droid drs drv dtk dtxgallery
    dtxtut duerer duerer-latex duotenzor dvdcoll dviasm dviincl dvipdfmx-def
    dvipsconfig dyntree
  ean ean13isbn easy easylist
    ebong ebsthesis ecc ecclesiastic ecltree economic
    ecv ed edmac edmargin ednotes eemeir eepic egameps
    egplot eiad eiad-ltx eijkhout
    elbioimp electrum ellipsis elmath elpres elsarticle
    elteikthesis eltex elvish
    emptypage emulateapj emp encxvlna endfloat endheads endnotes
    engpron engrec engtlc enumitem enumitem-zref envbig environ envlab
    epigrafica epigram epigraph epiolmec eplain
    epsdice epsf epsincl epslatex-fr epspdf epspdfconversion epstopdf
    eqell eqlist eqname eqnarray eqparbox errata es-tex-faq
    erdc esdiff esint esint-type1 esk eskd eskdx eso-pic esvect
    estcpmm
    etaremune etex-pkg etextools ethiop ethiop-t1 etoolbox euenc eukdate
    euler eulervm euproposal euro europecv eurosans eurosym
    everyhook everypage
    exam examdesign examplep excludeonly exercise exp-testopt
    expdlist expl3 export expressg extarrows exteps
    extpfeil extract extsizes
  faktor fancybox fancyhdr fancynum fancypar fancyref fancytooltips fancyvrb
    FAQ-en fbs featpost fenixpar
    feyn feynmf fge fig4latex figbas figbib figflow figsize
    filecontents filehook fileinfo findhyph fink first-latex-doc
    fix2col fixfoot fixme fixpdfmag
    fjodor
    flabels flacards flagderiv flashcards flashmovie flippdf float floatrow
    flowfram fltpage fltpoint
    fmtcount
    fn2end fnbreak fncychap fncylab fnpara fntproof
    foekfont font-change fontch fontinst fontools
    fontspec fonttable fontwrap
    footbib footmisc footnpag forarray forloop formular
    fouridx fourier fouriernc
    fp fpl
    fragmaster fragments frame framed frankenstein frcursive
    free-math-font-survey
    frenchle frletter frontespizio ftcap ftnxtra
    fullblck functan fundus fwlw
  gaceta galois garrigues gastex gatech-thesis gauss
    gb4e gcard gchords gcite gene-logic
    genmisc genmpage gentium gentle geometry
    german germbib germkorr getfiledate
    gfsartemisia gfsbaskerville gfsbodoni gfscomplutum gfsdidot gfsneohellenic
    gfsporson gfssolomos
    gillcm ginpenc gloss glossaries gmdoc gmdoc-enhance gmeometric
    gmiflink gmutils gmverb gmverse gnu-freefont gnuplottex gost
    graphics graphics-pln
    graphicx-psmin greek-inputenc greekdates greenpoint grfpaste
    grid gridset grverb gu guitar guitlogo
  hands hanging har2nat hardwrap harmony
    harvard harvardkyoto harvmac hatching
    hc hep hepnames
    hepparticles hepthesis hepunits here hexgame
    hhtensor histogr historische-zeitschrift hitec
    hrefhide hrlatex hvfloat hvindex
    hypdvips hyper hypernat hyperref hyperxmp hyph-utf8 hyphen-base 
    hyphenat hyphenex hyplain
  ibygrk icsv idxlayout ieeepes
    ifmslide ifmtarg ifplatform iftex ifxetex
    ijmart ijqc
    imac image-gallery imakeidx impatient impatient-fr import imtekda
    inconsolata inlinebib inlinedef inputtrc insbox installfont
    interactiveworkbook interfaces intro-scientific inversepath
    ionumbers iopart-num iso
    iso10303 isodate isodoc isomath isonums isorot isotope itnumpar iwona
  jablantile jamtimes jeopardy jknapltx jmlr jneurosci jpsj junicode
    jura juraabbrev jurabib juramisc jurarsp js-misc
  kalender kastrup kerkis kerntest
    keycommand keystroke kix kixfont
    knitting knittingpattern knuth
    koma-moderncvclassic koma-script kpfonts kurier
  l2picfaq l2tabu l2tabu-english l2tabu-french l2tabu-it l2tabu-spanish
    labbook labelcas labels lastpage
    latex latex-course latex-doc-ptr latex-notes-zh-cn
    latex-referenz latex-tabellen
    latex-tds latex-veryshortguide
    latex2e-help-texinfo latex2e-help-texinfo-spanish latex2man
    latex4wp
    latexcheat latexcheat-esmx latexcheat-ptbr latexdiff latexmk latexmp
    lato layaureo layouts lazylist
    lcd lcg lcyw leading leaflet lecturer ledmac leftidx lettrine
    lewis lexikon lfb lgreek lh
    libertine librarian libris limap linearA linegoal lineno linguex
    lipsum listbib listing listings listings-ext listliketab listofsymbols
    lithuanian liturg lkproof lm
    locality logical-markup-utils logpap logreq lpic lps lsc
    lshort-bulgarian lshort-chinese lshort-dutch lshort-english
    lshort-finnish lshort-french lshort-german lshort-italian
    lshort-japanese lshort-korean lshort-mongol lshort-persian
    lshort-polish lshort-portuguese lshort-russian lshort-slovak
    lshort-slovenian lshort-spanish lshort-thai lshort-turkish lshort-ukr
    lshort-vietnamese ltabptch
    ltxdockit ltxindex ltxmisc ltxnew
    lua-alt-getopt luacode
    luainputenc lualatex-doc lualibs luamplib luaotfload
    luasseq luatexbase luatextra
    lxfonts ly1
  macqassign mafr magaz magyar mailing mailmerge
    makebarcode makebox makecell makecirc makecmds makedtx makeglos makeplot
    manuscript margbib
    marginfix marginnote marvosym math-e mathabx mathcomp mathdots mathexam
    mathmode mathspec mathspic mattens maybemath
    mcaption mceinleger mcite mciteplus
    mdframed mdputu mdwtools memdesign memexsupp
    memoir MemoirChapStyles mentis
    menu
    metafont-beginners metago metalogo metaobj metaplot metatex metauml
    method metre
    mff mfnfss mfpic mfpic4ode mftinc mh mhchem
    mhequ microtype midnight midpage miller
    minibox minipage-marginpar miniplot minitoc
    minted minutes
    mkgrkindex mkjobtexmf mkpattern
    mla-paper mlist mmap mnsymbol
    moderncv modref modroman mongolian-babel montex morefloats moresize
    moreverb movie15 mp3d mparhack
    mpman-ru ms msc msg mslapa msuthesis mtgreek
    multenum multibbl multibib multicap multirow
    multibox multido multiobjective munich musixtex muthesis mwcls mxedruli
    mylatexformat
  nag namespc natbib nath nature ncclatex ncctools
    needspace newcommand newfile newlfm newsletr newspaper newverbs nextpage
    nfssext-cfr niceframe nicetext nih nkarta nlctdoc
    noitcrul nolbreaks
    nomencl nomentbl nonfloat nopageno nostarch notes notes2bib notoccite
    nrc ntgclass ntheorem ntheorem-vn numericplots numname numprint
  oberdiek objectz ocgtools ocr-b ocr-latex octavo ofs ogham
    oldlatin oldstandard oldstyle
    onlyamsmath onrannual opcit optional
    ordinalpt orkhun ot-tableau othello oubraces outline overpic
  pagecont pagenote pagerange pagesLTS paper papercdcase papermas papertex
    paralist parallel paresse parrun parselines parskip passivetex 
    patch patchcmd patgen2-tutorial path pauldoc pax pbox pb-diagram
    pdf14 pdf-forms-tutorial-de pdf-forms-tutorial-en
    pdf-trans pdfcomment pdfcprot pdfcrop pdfjam pdfmarginpar
    pdfpages pdfscreen pdfslide pdfsync pdftex-def pdftricks pdfx
    pecha perception perltex
    permute persian-bib petiteannonce
    pgf pgf-soroban pgf-umlsd pgfopts pgfplots
    phaistos philex philosophersimprint
    phonetic photo picinpar pict2e pictex pictex2 piff pigpen
    pinlabel pitex pittetd pkfix pkfix-helper placeins placeins-plain
    plari plantslabels plates play
    plnfss plweb pmgraph pnas2009
    poemscol polski poltawski polyglossia polynom polynomial
    polytable poster-mac powerdot powerdot-FUBerlin
    ppr-prv pracjourn preprint prerex present presentations
    prettyref preview printlen proba probsoln procIAGssymp productbox program
    progress progressbar properties protex protocol
    psbao pseudocode psfrag psfrag-italian psgo pslatex psnfss pspicture
    pst-2dplot pst-3d pst-3dplot pst-abspos pst-am pst-asr pst-bar
    pst-barcode pst-bezier pst-blur pst-bspline
    pst-calendar pst-circ pst-coil pst-cox pst-dbicons pst-diffraction
    pst-electricfield pst-eps pst-eucl pst-eucl-translation-bg pst-exa
    pst-fill pst-fr3d pst-fractal pst-fun pst-func
    pst-gantt pst-geo pst-grad pst-infixplot pst-jtree pst-knot pst-labo
    pst-lens pst-light3d pst-magneticfield pst-math pst-mirror pst-node
    pst-ob3d pst-optexp pst-optic
    pst-osci pst-pad pst-pdgr pst-platon pst-plot pst-poly pst-pdf
    pst-qtree
    pst-sigsys pst-solides3d pst-soroban pst-spectra
    pst-slpe pst-stru pst-support pst-text pst-thick pst-tree pst-uml
    pst-vowel pst-vue3d
    pst2pdf pstool pstricks pstricks-add
    psu-thesis ptptex punknova purifyeps pxfonts
  qcm qobitree qstest qsymbols qtree quotmark
  r_und_s randbild randtext rccol rcs rcs-multi rcsinfo
    realscripts rec-thy recipe recipecard recycle rectopma
    refman refstyle regcount register relenc relsize resumemac revtex
    rjlparshap rlepsf rmpage
    robustcommand robustindex romande romanneg romannum rotating
    rotfloat rotpages
    roundbox rsc rsfs
    rtkinenc rtklage russ rvwrite ryethesis
  sageep sansmath sauerj sauterfonts savefnmark savesym savetrees scale
    scalebar schemabloc scientificpaper sciposter screenplay script sdrt
    secdot section sectionbox sectsty selectp semantic semaphor
    seminar semioneside sepnum
    seqsplit setspace seuthesis sf298 sffms sfg sfheaders
    sfmath sgame shade shadethm shadow shapepar
    shipunov shorttoc show2e showdim showexpl showlabels showtags shuffle
    sidecap sides silence simplecd simplecv simplewick simplified-latex siunitx
    skak skaknew skb skeycommand skeyval
    slantsc slashbox slideshow smalltableof smartref
    snapshot songbook sort-by-letters soton soul soyombo
    spanglish spanish spanish-mx sparklines spie
    sphack splines splitbib splitindex spot spotcolor spreadtab spverbatim
    srcltx stack stage standalone starfont statistik statex statex2 staves
    stdclsdv stdpage steinmetz
    stellenbosch stex stix stringstrings struktex
    sttools stubs suanpan subdepth subeqn subeqnarray
    subfig subfigmat subfigure subfloat substr subsupscripts sudoku
    sudokubundle suftesi sugconf
    supertabular susy svg-inkscape svgcolor svn svn-multi svn-prov svninfo
    swebib swimgraf syllogism syntax synproof syntrace synttree
  t-angles t2
    tableaux tablenotes tablists tablor tabls tabto-generic tabto-ltx
    tabu tabularborder tabularcalc tabularew
    tabulary tabvar talk
    tamethebeast tapir tdclock tdsfrmath
    technics ted templates-fenn templates-sommer tengwarscript
    tensor termcal termlist teubner
    tex-ewd tex-font-errors-cheatsheet tex-gyre tex-label tex-overview
    texapi texcount
    texdiff texdirflatten texilikechaps texilikecover
    texloganalyser texlogos texmate texments
    texpower texshade
    textcase textfit textmerg textopo textpath textpos tfrupee
    thailatex theoremref thesis-titlepage-fhac
    thinsp thmbox thmtools threeddice threeparttable threeparttablex
    thumb thumbpdf thumby thuthesis
    ticket tikz-3dplot tikz-inet tikz-qtree tikz-timing timetable timing tipa
    titlefoot titlepages titlepic titleref titlesec titling
    tkz-doc tkz-linknodes tkz-orm tkz-tab
    tocbibind tocloft tocvsec2 todo todonotes
    tokenizer toolbox tools topfloat totcount totpages
    trajan trfsigns trimspaces trivfloat trsym truncate
    tucv tufte-latex tugboat tugboat-plain turkmen turnstile turnthepage
    twoinone twoup txfonts
    txfontsb type1cm typehtml typogrid
  uaclasses ucdavisthesis ucs
    ucthesis uebungsblatt uiucthesis
    ulem ulqda
    umich-thesis uml umlaute umoline
    umthesis umtypewriter
    underlin underscore undertilde undolabl
    unicode-math units unitsdef universa
    uothesis uowthesis upmethodology upquote
    url urlbst ushort ut-thesis uwthesis
  vancouver variations varindex varisize
    varsfromjobname varwidth vaucanson-g
    vector velthuis venn venturisadf verbatimbox verbatimcopy verbdef
    verse version versions vertbars
    vhistory visualfaq vmargin vntex volumes vpe vruler vwcol vxu
  wadalab wallpaper warning warpcol widetable williams wordlike
    wrapfig wsuipa
  xargs xcolor xcomment xdoc
    xecjk xecolour xecyr xeindex xepersian xesearch
    xetex-def xetex-itrans xetex-pstricks xetexfontinfo
    xetexref xfor xgreek 
    xifthen xits
    xkeyval xlop xltxtra xmpincl xnewcommand xoptarg xpackages
    xq xskak xstring xtab xtcapts xunicode
    xwatermark xyling xypdf xypic xypic-tut-pt xytree
  yafoot yagusylo yax ydoc york-thesis youngtab yplan ytableau
  zed-csp ziffer zhmetrics zhspacing zwgetfdate zwpagelayout
); 


# these packages we do not expect to check.  Once this list is complete,
# we can start working on tlmgr list | grep shortdesc.
my @TLP_no_check = (
  "afm2pl",		# not on CTAN
  "aleph",		# binary
  "asymptote",		# binary
  "bibtex",		# binary
  "bibtex8",		# binary
  "bibtexu",		# binary
  "chktex",		# binary
  "cjkutils",		# binary
  "context",		# binary+taco
  "ctib",		# binary
  "ctie",		# binary
  "cweb",		# binary
  "detex",		# binary
  "devnag",		# binary
  "dvi2tty",		# binary
  "dvicopy",		# binary
  "dvidvi",		# binary
  "dviljk",		# binary
  "dvipdfm",		# binary
  "dvipdfmx",		# binary
  "dvipng",		# binary
  "dvipos",		# binary
  "dvips",		# binary
  "dvisvgm",		# binary
  "enctex",		# binary
  "etex",		# binary
  "finbib",		# missing on CTAN
  "fontname",		# tl-update-auto
  "groff",		# binary
  "gsftopk",		# binary
  "ifluatex",		# part of oberdiek
  "kpathsea",		# binary
  "latex-bin",		# binary
  "latexconfig",	# we maintain
  "lcdftypetools",	# binary
  "luatex",		# binary
  "makeindex",		# binary
  "metafont",		# binary
  "metapost",		# binary
  "mfware",		# binary
  "mltex",		# binary
  "omega",		# binary
  "omegaware",		# binary
  "patgen",		# binary
  "pdftex",		# binary
  "powerdot",		# stale generated files on CTAN
  "synctex",		# binary
  "t1utils",		# binary
  "tetex",		# our sources
  "tex",		# binary
  "tex4ht",		# binary
  "texconfig",		# our sources
  "texdoc",		# binary
  "texinfo",		# tl-update-auto
  "texlive.infra",	# binary
  "texware",		# binary
  "texworks",		# binary
  "tie",		# binary
  "ttfutils",		# binary
  "xdvi",		# binary
  "xetex",		# binary
  "xetexconfig",	# our sources
  "xindy",		# binary
);

exit (&main ());


sub main
{
  chomp ($Master = `cd $mydir/../.. && pwd`);
  $tlpdb = TeXLive::TLPDB->new ("root" => $Master);
  die "Cannot find tlpdb in $Master!" unless defined $tlpdb;

  $OPT{"verbose"} = 0;
  if ($ARGV[0] eq "--verbose") {
    $OPT{"verbose"} = 1;
    shift @ARGV;
  }
  
  $OPT{"all"} = 0;
  if ($ARGV[0] eq "--all") {
    $OPT{"all"} = 1;
    shift @ARGV;
  }
  
  if ($ARGV[0] eq "--list-not-treated") {
    my @not_checked = ();
    for my $b (&normal_tlps ()) {
      if (! grep ($b eq $_, @TLP_working, @TLP_no_check)) {
        push (@not_checked, $b);
      }
    }
    print "" . (@not_checked+0) . " TL packages not checked against CTAN:\n";
    print "@not_checked\n";
    @ARGV = (); # no normal checks

  } elsif ($ARGV[0] eq "--check") {
    # check all/only those packages we have actually run through this mill.
    @ARGV = @TLP_working;
  } elsif ($ARGV[0] eq "--check-all") {
    @ARGV = &normal_tlps ();
  }

  my $errcount = 0;
  for my $tlp (@ARGV) {
    print "checking $tlp..." if $OPT{"verbose"};
    $errcount += &do_tlp ($tlp);
  }
  return $errcount;
}

# gives a list with only the 'normal' packages,
# that is, excluding meta-packages and binary packages
# (and hyphen for the moment)
# 
sub normal_tlps
{
  my @normal;
  my $non_normal = `ls $Master/bin`;
  $non_normal =~ s/\n/\$|/g;
  $non_normal .= '^0+texlive|^bin-|^collection-|^scheme-|^texlive-|^hyphen-';
  foreach ($tlpdb->list_packages) {
    push (@normal, $_) unless (/$non_normal/);
  }
  return @normal;
}


# Return 1 if package needs updating, 0 if ok.
# 
sub do_tlp
{
  my ($tlpn) = @_;
  my $needed = 0;
  
  my $tlp = $tlpdb->get_package($tlpn);
  if (!defined($tlp)) {
    warn "$0: no package $tlpn\n";
    return 1;
  }

  chomp (my $ctan_dir = `$mydir/tlpkginfo --prepare '$tlpn'`);
  if (! $ctan_dir) {
    warn "$0: oops, no CTAN directory for $tlpn, fix fix\n";
    return ();
  }
  my @tl_files = ();
  push @tl_files, $tlp->runfiles;
  push @tl_files, $tlp->docfiles;
  push @tl_files, $tlp->srcfiles;
  if ($tlp->relocated) { for (@tl_files) { s:^$RelocPrefix/:$RelocTree/:; } }
  # we don't push bin files.
  my @tl_basefiles = ();  # compare with CTAN after the loop
    
  my @compared = ();
  for my $file (@tl_files) {
    (my $basefile = $file) =~ s,^.*/,,;
#warn "checking file $file -> $basefile\n";
    push (@tl_basefiles, $basefile);

    # no point yet in comparing our pdfs, we have generated too many.
    # revisit this someday to see how many of ours are actually different.
    # However, for lshort translations, the pdf is often the only thing
    # we have to compare.
    next if $file =~ /\.pdf$/ && $file !~ /short/;
    
    my $tl_file = "$Master/$file";
    if (! -e $tl_file) {
      warn "$tl_file: TL file missing\n";
      next;
    }
    
    chomp (my @ctan_files = `find $ctan_dir/ -name $basefile`);
#warn "ctan find $basefile: @ctan_files\n";
    # the trailing / is so if we happen to hit a symlink on CTAN, it'll
    # go through.
    next if @ctan_files > 1; # if more than one file by same name, skip

    my $ctan_file = $ctan_files[0];
#warn "ctan file is $ctan_file\n";
    if (! -e $ctan_file) {
      # maybe it'll be there with a case change in the name
      chomp (@ctan_files = `find $ctan_dir/ -iname $basefile`);
      next if @ctan_files > 1; # if more than one file by same name, skip
      $ctan_file = $ctan_files[0];
      
      if (! -e $ctan_file) {
        # we generate lots of files, eg perlmacros.sty, so might skip.
        warn "$ctan_file: CTAN file missing\n"
          if $ctan_file && $ctan_file !~ /(cfg|dvi|sty|tex)$/;
        next;
      }
    }

    push (@compared, $basefile);
    if (&files_differ ($tl_file, $ctan_file)) {
      print "# $tlpn\ndiff $ctan_file $tl_file\n";
      $needed = 1;
      last unless $OPT{"all"};
    }
  }
  
  # unfortunately, we cannot do this.  There are many PDF's on CTAN
  # which have no sources or otherwise problematic for TL.  Perhaps one
  # day we could use the %moreclean hash from ctan2tds as an additional
  # filter, i.e., put all those ctan2tds tables in an external file.
  # 
  # # check for PDF files on CTAN that we don't have.
  # my @ctan_pdf_needed = ();
  # chomp (my @ctan_files = `find $ctan_dir -name \*.pdf`);
  # for my $cfile (@ctan_files) {
  #   (my $base_cfile = $cfile) =~ s,^.*/,,;
  #   if (! grep { $_ eq $base_cfile } @tl_basefiles) {
  #     push (@ctan_pdf_needed, $base_cfile);
  #   }
  # }
  # if (@ctan_pdf_needed) {
  #   if (! $needed) {
  #     # if this is the first thing needed (no diffs), print package name.
  #     print "# $tlpn\n";
  #     $needed = 1;
  #   }
  #   print "# new on ctan: @ctan_pdf_needed\n";
  # }
  
  if (@compared == 0) {
    warn "\n$tlpn: no files to compare in $ctan_dir, fixme!\n";
    warn "  (tl_files = @tl_files)\n";
    warn "  (ctan_files = @ctan_files)\n";
  } elsif ($needed == 0) {
    print "ok\n" if $OPT{"verbose"};
  }
  print ((@compared + 0) . " compared (@compared)\n") if $OPT{"verbose"};

  # clean up the tmpdir possibly created from tlpkginfo --prepare.
  chomp (my $ctan_root = `$mydir/tlpkginfo --ctan-root`);
  if ($ctan_dir !~ m,^$ctan_root, && ! $OPT{"no-clean"}) {
    system ("rm -rf $ctan_dir");
  }

  return $needed;
}



# 0 if files are the same, 1 if they are different.
# 
sub files_differ
{
  my ($tl_file,$ctan_file) = @_;
#warn "comparing $ctan_file $tl_file\n";
  return system ("$mydir/cmp-textfiles $ctan_file $tl_file");
}

# vim: set ts=8 sw=2 expandtab: