groff
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Groff] What does the "-u" in ".tmac-u" mean?


From: hj.oertel
Subject: Re: [Groff] What does the "-u" in ".tmac-u" mean?
Date: Sun, 05 Nov 2017 08:49:02 +0100

Am 05.11.2017 03:09 schrieb "G. Branden Robinson" <address@hidden>:
>
> At 2017-11-05T00:40:48+0100, Bertrand Garrigues wrote: 
> > > Am Donnerstag, 2. November 2017, 08:09:48 CET schrieb Werner LEMBERG: 
> > >> If my memory serves me well, the `u' stands for `uncompressed', i.e., 
> > >> without comments and indentation removed.  Bertrand removed the 
> > >> functionality to install stripped tmac files, IIRC — computers are 
> > >> much faster today, so this is (probably?) no longer needed. 
> > 
> > Correct, -u is for 'unstripped'.  Before the automake migration the the 
> > unstripped files were saved in the source tree without any extension, 
> > and the build system would first generate a stripped file in the build 
> > directory with a -s suffix and then install the file without this 
> > suffix.  I've simplified the build process by storing the unstripped 
> > file in the source tree with a -u suffix and generating the stripped 
> > version without suffix in the build directory.  The final result is 
> > unchanged though. 
>
> I move that we stop stripping the files.  Stripping them is saving us 
> only a few hundred kB out of 25 megs. 
>
> That's out of a full installation; in Debian, the "core" macro packages 
> are in groff-base, and only doc{,-old}.tmac are shipped there.  So for 
> at least some distributors, people only get the bulk of these stripped 
> macro packages when they're asking their packaging manager for something 
> non-minimal anyway. 
>
> As for groff-base and whatever its counterparts may be in other distros, 
> we should weigh the ~75kB cost of leaving it unstripped versus the 
> benefit of making the package more user-readable for the inspection of 
> prospective man page writers who'd like to write in mdoc instead of man, 
> so that they can learn and become more comfortable with it.  (This is me 
> shamelessly pandering to Ingo.  :P ) 
>
> The numbers below are in bytes. 
>
> $ for F in $(find -name "*.tmac-u"); do wc -c "$F" "${F%-u}"; done | awk 
> '!/total/ {printf "%6d %s\n", $1, $2}' 
>   9088 ./contrib/hdtbl/hdmisc.tmac-u 
>   3445 ./contrib/hdtbl/hdmisc.tmac 
> 26802 ./contrib/hdtbl/hdtbl.tmac-u 
> 17350 ./contrib/hdtbl/hdtbl.tmac 
> 689635 ./contrib/mom/om.tmac-u 
> 420776 ./contrib/mom/om.tmac 
> 34265 ./tmac/e.tmac-u 
> 22078 ./tmac/e.tmac 
> 150717 ./tmac/doc.tmac-u 
> 74050 ./tmac/doc.tmac 
> 36308 ./tmac/doc-old.tmac-u 
> 36446 ./tmac/doc-old.tmac 
>
> $ du -bs ~/local-groff 
> 25532687        /home/branden/local-groff 
>
> -- 
> Regards, 
> Branden

Reducing the size is not all what matters. Nobody so far has measured macro 
loading time.
Heinz

reply via email to

[Prev in Thread] Current Thread [Next in Thread]