[[email protected]: Re: Re: update / new features]

On Wed, Jul 29, 2026 at 18:40 (+0200), Hans Hagen wrote:

> On 7/29/2026 6:26 PM, Jim Diamond wrote:

>> My poor old laptop (Ryzen 4700U, but not quite as old as yours) took
>> context test-mp.tex  6.67s user 0.15s system 99% cpu 6.852 total

>>> (5.866 metapost) ..

>> Pray tell: where/what are you reading for that number?  When I
>> grep -i metapost the-log-file.log
>> I only find
>> mkxl run stats  > metapost: 0.014 seconds, loading: 0.030, execution: 0.=
005, n: 3, average: 0.016, instances: 1, luacalls: 16452 (file: 15976, text=
: 0, script: 456, log: 20), memory: 0.000 M

>> If you spent 5.866 in metapost, it seems extremely unlikely I finished up
>> in 14 millis.

> i bet you had a second run .. was cache set to true?

No, that was on the first run.

If I run it again, I get
        context test-mp.tex  0.36s user 0.07s system 99% cpu 0.436 total
and, indeed, it appears quite quickly.


However, if I copy my .tex file to a new file, it takes "quite a while":

...
mkxl run stats  > metapost: 0.015 seconds, loading: 0.031, execution: 0.005=
, n: 3, average: 0.017, instances: 1, luacalls: 16452 (file: 15976, text: 0=
, script: 456, log: 20), memory: 0.000 M
...
mkxl run stats  > runtime: 0.372 seconds, 1 processed pages, 1 shipped page=
s, 2.689 pages/second
system          | total runtime: 6.461 seconds of 6.525 seconds

context dog.tex  6.37s user 0.14s system 99% cpu 6.528 total


So the shell says I used 6.37 seconds, which seems about right.

But the two time reports from context say very little time was used.

Is it possible that there is some peculiarity in some library version on my
system which causes context's time reports to be incorrect?


>> Aside from that, you will be pleased to know that it shows up well (*) in
>> all of xpdf, evince and my venerable Acroread 9.5.5.

> oh, even the old acrobat!

You forgot "venerable".  ;-)
And I think JP would tell you the polite expression is "=E2g=E9e"
or "bien =E2g=E9e".


>> I saw a SGI machine on sale a number of years ago, at a price I could ha=
ve
>> afforded without crimping the budget substantially.  But I resisted the
>> temptation.  Which saved me some time today, because otherwise I would h=
ave
>> had to check your bet.

> even worse, make luametatex compile on it,

That might be a bit of a trick as well.

                                Jim

___________________________________________________________________________=
________
If your question is of interest to others as well, please add an entry to t=
he Wiki!

maillist : [email protected] / https://mailman.ntg.nl/mailman3/lists/ntg-c=
ontext.ntg.nl
webpage  : https://www.pragma-ade.nl / https://context.aanhet.net (mirror)
archive  : https://github.com/contextgarden/context
wiki     : https://wiki.contextgarden.net
___________________________________________________________________________=
________