FC_WEIGHT didn't match

[jfbu]

Hi Uwe :

Le 5 mars 2019 à 16:40, Uwe Siart <usenet at siart.de> a écrit :

> jfbu <jfbu at free.fr> writes:
> 
>> Fwiw I have had that message with TL2018 for a long time
>> 
>> I mostly use dvipdfmx, but recently I started experimenting
>> with PsTricks hence latex+dvips+ps2pdf and I systematically
>> get that.
>> 
>> \documentclass{article}
>> \usepackage{pst-plot}
>> 
>> \begin{document}
>> 
>> \begin{pspicture}(-3,-7)(3,6)
>>  \psgrid[subgriddiv=1,gridlabels=0,gridwidth=0.4pt](-3,-5)(3,5)
>> \end{pspicture}
>> 
>> \thispagestyle{empty}
>> \end{document}
> 
> Same thing. I can't reproduce the error. Your MWE runs smoothly through
> latex/dvips/ps2pdf and produces correct results here.
> 
> Environment: Win7, TL19 pretest, Ghostscript 9.26


I have upgraded Ghostscript to 9.26 but the error persists

(after cleaning all files and starting from scratch)

$ ps2pdf testfcweight.ps
DEBUG: FC_WEIGHT didn't match
$ gs --version
9.26

MacOS 10.9.5 (a very good system still recent, 2014 I think)
TL 2018 fully updated

I get the same with the MacPorts TL2018 (december 2018) binaries.
(except for the kpse* binaries picked up from darwinlegacy, sorry
for complication).

I have not tested with TL2019-pretest.

Very strangely I *thought* having reproduced once the problem
with Bob example

\documentclass{minimal}
\usepackage{mathpazo}
\begin{document}
foo
\end{document}


on my TeXLive 2018 with 9.26 Ghostscript and darwinlegacy binaries

$ ps2pdf testfcweight.ps
DEBUG: FC_WEIGHT didn't match

but, I did not get it 

- not with the MacPorts binaries

- not with my own compiled binaries from long time ago

and now I can *not* reproduce with darwinlegacy either.

But I do confirm the problem on my PsTricks example

and **all three** sets of binaries.

(each time I do fmtutil-sys --all and updmap-sys)

Sorry not to report yet on pretest situation.

But this is to point out this issue was already there in some configurations of users
with TeXLive 2018.

Jean-François