RE: Building from source

--=-rinX1tIq/ACuMEHa+Nrw
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Fri, 2003-08-01 at 16:45, Dittmar, Daniel wrote:
> > sources from
> > sapdb.org ("Quick Download"), the development environment didn't build
> > because of missing PATH settings, ptoc etc. Well, this is=20
> > documented in
> > the mailing list archives (to use CVS sources), but could be improved
> > anyhow.
>=20
> We now have found someone with enough time to test the source packages
> regularly, so I hope this situation improves.

I'm sure this sounds great to all of us building preferably from source.

> > because of errors in dependencies made of 'sqlpy.mac',=20
> > command 'PYTHONPATH=3D$INSTMISC python -c "import sapdb.sql"...' not
>=20
> The important error comes before this one. It looks as if the python exte=
nsion
> could not be built.
>=20
> You get a protocol of all the errors in $WRK/fast/prot/all.mac.e0

This one is full of "dependency errors", "syntax errors" and "No such
file or dir...". I made it available (94k) on my web site at
http://www.antcom.de/download/all.mac.e0. Maybe this provides some help
for the debugging process.

> > Inbetween, there are some compile errors like the following, but the
> > build process doesn't stop, so one waits for hours (on a P4 2.8GHz) to
> > get the notice that there were some problems at the=20
> > beginning. I propose
> > a regular make behaviour aborting on errors.
>=20
> I've asked the vmake developer if he could add such an option and at leas=
t,
> he didn't say that it's impossible.

This would be very much appreciated!

> For our work, we prefer it of course the way it is. We either edit a smal=
l
> number of files, then the difference between 'make' and 'make -k' is negl=
igible.
> Or we make big changes, in which case we want to compile as much as possi=
ble
> while we're doing something else.

I once ran a project like this and won't do it again. While this is a
habit the original author can adapt to quite productively, it's
confusing to the rest of the community.

> And in the case of supporting external developers, I also prefer to get
> all the errors at once instead of one per day.

This could be done via an explicit argument, in turn. ;)

> > vls14_scan.cpp:68: syntax error before `[' token
>=20
> vls14_scan.cpp gets generated by flex from sys/src/ls/vls14_scan.l.
> Please look into $WRK/fast/tmp/ for this file and post this line and=20
> possible a few around it.

I guess you mean sys/wrk/fast/tmp/vls14_scan.save? This would be:

|   struct yy_trans_info
|           {
|           flex_int32_t yy_verify;
|           flex_int32_t yy_nxt;
|           };
68: static yyconst flex_int16_t yy_accept[738] =3D
|       {   0,
|           0,    0,  173,    2,    2,  171,  171,  165,   11,   10,
|           5,    6,   13,    7,   12,    8,  171,  164,    9,   15,
|          14,   16,  165,  165,  165,  165,  165,  165,  165,  165,
|         165,  165,  165,  165,  165,  165,  165,  165,  165,  165,
|         165,  165,  165,  165,  165,  165,    0,  165,    0,    0,
|           3,  170,    1,  166,    0,  164,   17,   19,   18,    0,


Roland Stigge

--=-rinX1tIq/ACuMEHa+Nrw
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQA/KoZacaH/YBv43g8RAsgKAJoDNAseNMA90x7VcqljIYg6a8KnyACgqDzh
608bRBpip2Muds48NIAGsgQ=
=xQag
-----END PGP SIGNATURE-----

--=-rinX1tIq/ACuMEHa+Nrw--