[Fontinst] Development history reconstruction

[Lars Hellström]

ulrik.vieth at arcor.de skrev:
> Hi Lars,
> 
> I don't have a complete collection myself, but I can point you
> to my collection on the TeX Historic archives:
> 
> ftp://ftp.tug.org/historic/fonts/utilities/fontinst/
> ftp://ftp.tug.org/historic/fonts/psfonts/tools/

Ah, this looks very good -- even some items before 1.335! I had a bit 
of trouble downloading the files, but think I've got it all now.

> Some of this material was reconstructed from old CTAN CDs 
> and some came from my personal collection.

Certainly saves me the trouble of digging through boxes on the garage 
attic in the hope of finding such an old CD myself.

> What I can recall about the version history is the following:
> 
> v1.335 was the last official version on CTAN from Alan Jeffrey
> in fonts/utilities/fontinst. It dates back to 1994, but stayed on 
> CTAN until it was replaced by 1.80x or 1.90x.

1.80x, I think.

> v1.500 and v1.504 came to CTAN as part of fonts/psfonts/tools
> and were maintained by Sebastian Rathz.  They back to 1995
> and were subsequently updated in 1996-1997 (v1.6.0, v.1.6.1).
> 
> v1.7 and v1.8 were short-lived fixes to Sebastian's distribution
> which came from me in 1998, before I started to rewrite fontinst 
> in ltxdoc. I'm not sure if they ever were submitted to CTAN.

It seems strange that there should have been an 1.8 which was different 
from 1.800, given that version numbers are compared as <factor>s.

> v1.800 and v.1801 were the first versions in fonts/utilities/fontinst
> that replaced Alan's version in 1.335.  These versions merged
> several changes from Alan's v1.511 and Sebastian's 1.6 and 1.7.

Indeed, 1.511 rather than 1.512.

> v1.90x were the first versions, when you took over from me
>  in 1999.  AFAIK, the first versions were kept on CTAN in 
> fonts/uitilities/fontinst-beta, before they replaced v1.801. 

An attempt of mine to use CTAN as home for delevelopment. Didn't work 
very well.

> In my archive, some the earlier verisions were distributed
> as SHAR files by mail on the fontinst mailing list.

Eh? The .mime.gz files appear to be mails I sent, that have the source 
files as MIME attachments. I certainly wouldn't have been able to 
create SHAR archives back then... Also they don't seem to be addressed 
to the list, but rather to you and Sebastian.

I think some bugfixes were distributed via the mailing list, but not 
fontinst as a whole.

> It is quite possible that some early versions didn't even 
> arrive on CTAN, before reaching v1.910 or some such.
> 
> So much for my take on history.
> 
> Regards, Ulrik
> 
> 
> P.S: If you want to upload additional versions to the archive,
> please contact me, so we can arrange for copying the files.

A historical repository would certainly have a place in that location.



Karl Berry wrote:
> Personally, I wouldn't worry too much about achieving a perfect vc
> history.  Having a sequence of release tarballs available (e.g., in
> ftp://ftp.tug.org/historic as Ulrik wrote about) seems like all the
> history that anyone could stand, anyway.

The thing is, that history gets a lot more accessible when put in a 
repository than when just a collection of tarballs. fontinst.sty has a 
pretty good history record in the \changes, but the various ETX and MTX 
files are another matter.

> Also, FWIW, I agree with Werner's suggestion to use savannah (or gna --
> anything but sourceforge, which is soooo slow and painful and full of
> advertising).

The catch there would be that savannah only supports CVS (AFAICT from 
the docs; they're not at all clear on the matter) and gna only CVS, 
SVN, and gnuarch. I really prefer to use git for version control, over 
the above because it's distributed and supports SSH.


Lars Hellström