<html><body><p class="p1">It looks like the new interpreter shaves about half a second off of parsing time for the full tlpdb. Of course, my comments in the code show a time of 1.7 seconds for TL 2016 (when I originally wrote the python-based parser). Ouch.<br></p><div><br></div><div><span class="s1">$ time ~/build/tlutility/parse_tlpdb.py /usr/local/texlive/2021/tlpkg/texlive.tlpdb -o /tmp/a.plist -f plist</span><br></div><p class="p1"><span class="s1">real<span class="Apple-tab-span"> </span>0m2.991s</span><br></p><p class="p1"><span class="s1">user<span class="Apple-tab-span"> </span>0m2.847s</span><br></p><p class="p1"><span class="s1">sys<span class="Apple-tab-span"> </span>0m0.127s</span><br></p><div><br></div><p class="p1"><span class="s1">time ~/build/tlutility/relocatable-python/<a rel="noopener noreferrer" href="http://Python.framework/Versions/3.9/bin/python3.9">Python.framework/Versions/3.9/bin/python3.9</a> ~/build/tlutility/parse_tlpdb.py /usr/local/texlive/2021/tlpkg/texlive.tlpdb -o /tmp/a.plist -f plist</span><br></p><p class="p1"><span class="s1">real<span class="Apple-tab-span"> </span>0m2.400s</span><br></p><p class="p1"><span class="s1">user<span class="Apple-tab-span"> </span>0m2.266s</span><br></p><p class="p1"><span class="s1">sys<span class="Apple-tab-span"> </span>0m0.102s</span><br></p><div><div><br></div><blockquote type="cite"><div>On Feb 8, 2022, at 6:53 AM, Bruno Voisin <<a href="mailto:bvoisin@icloud.com">bvoisin@icloud.com</a>> wrote:<br></div><div><br></div><div><br></div><div><div><blockquote type="cite"><div>On 8 Feb 2022, at 15:25, Herbert Schulz <<a href="mailto:herbs@wideopenwest.com">herbs@wideopenwest.com</a>> wrote:<br></div><div><br></div><div>Doing the same in TeXShop I see that dialog box with the progress bar too so it seems to be a standard Sparkle thing. The difference is that with TeXShop it's up long enough to see what is happening while things go so fast in TLU is flashed by so quickly it was disconcerting.<br></div></blockquote><div><br></div><div>It may be subjective, but I have the impression TLU is much faster now than it used to be, for example upon launch when checking for the availability of updates on the default repository.<br></div><div><br></div><div>If confirmed, could this be a side-effect of the switch to a built-in Python 3 framework?<br></div><div><br></div><div>Bruno<br></div></div></div></blockquote></div><div><br></div></body></html>