Package: tetex-base Version: 3.0.dfsg.3-5 Severity: wishlist Hi all,
once again to [EMAIL PROTECTED] I'm filing that as wishlist now, after asking for it a few month ago. Maybe you did change the code since then and I did not notice. If this is the case, please complain. I upgraded my tetex-base package and noticed the following: <snip> Setting up tetex-base (3.0.dfsg.3-1) ... mktexlsr: Updating /usr/local/share/texmf/ls-R... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFMAIN... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFDIST-TETEX... mktexlsr: Updating /var/lib/texmf/ls-R... mktexlsr: Done. Running updmap-sys. This may take some time... done. mktexlsr: Updating /usr/local/share/texmf/ls-R... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFMAIN... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFDIST-TETEX... mktexlsr: Updating /var/lib/texmf/ls-R... mktexlsr: Done. Running fmtutil-sys. This may take some time... done. mktexlsr: Updating /usr/local/share/texmf/ls-R... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFMAIN... mktexlsr: Updating /var/lib/texmf/ls-R-TEXMFDIST-TETEX... mktexlsr: Updating /var/lib/texmf/ls-R... mktexlsr: Done. </snip> i.e. we're calling mktexlsr 3 times. I understand that we have to call before updmap-sys and after fmtutil-sys, but do we really have to call between updmap-sys and fmtutil-sys? IMHO there is room for optimization. If not, please explain. Further additions coming from Frank and Florent: <Frank> > i.e. we're calling mktexlsr 3 times. I understand that we have to call > before updmap-sys and after fmtutil-sys, but do we really have to > call betwenn updmap-sys and fmtutil-sys? If a pdftex format includes font information, would it need a map file? I don't think so. </Frank> <Florent> > i.e. we're calling mktexlsr 3 times. I understand that we have to call > before updmap-sys and after fmtutil-sys, but do we really have to > call betwenn updmap-sys and fmtutil-sys? The first two calls to mktexlsr can be somehow justified by the recommended (at least by me!) sequence to use whenever you update map files (resp. format files): update-updmap mktexlsr updmap-sys resp. update-fmtutil mktexlsr fmtutil-sys But it's true that when both operations are done in a row, the second mktexlsr call is useless in most cases. The only case I see where it makes a difference is if the update-fmtutil call *creates* fmtutil.cnf, instead of updating it. Which can happen, though not very frequently. As for the third call, after fmtutil-sys, I think this one is really useless, because near the end of main(), fmtutil-sys has the following code: for i in *.fmt *.mem *.base; do test -f "$i" || continue rm -f "$destdir/$i" # We don't want user-interaction for the following "mv" command: if mv "$i" "$destdir/$i" </dev/null; then verboseMsg "$progname: $destdir/$i installed." $mktexfmtMode && echo "$destdir/$i" fi mktexupd "$destdir" "$i" done Just as with updmap(-sys), Thomas Esser was careful to have fmtutil-sys register the files it creates in the ls-R database. > IMHO there is room for optimization. If not, please explain. Yes, there is room IMHO too. However, calling mktexlsr does not take *very* long, so we won't gain that much. Still, 0.5 second or so on my system for non-first runs (later runs are faster than the first one due to the caching performed by the kernel). </Florent> <Frank> No, I don't think it has changed in teTeX. I think that TeXlive behaves more clever here, but I'm not sure. I don't think it's worth to fix this in teTeX - personally I do not plan any more uploads except those targetted at etch. But for the TeXLive packages, we should take care for optimization. </Frank> Thanks, Hilmar -- sigfault -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]