Обсуждение: inline is not ANSI C
I've noticed that in postgresql 7.2b4 and in some contrib(I've seen dbf2pgsql) there are inline variables, I think this is not ANSI code. MIPS PRO compilers did not work with inline unless u use API specific tools hope it helps
Вложения
Luis Amigo <lamigo@atc.unican.es> writes: > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > dbf2pgsql) there are inline variables, I think this is not ANSI code. > MIPS PRO compilers did not work with inline unless u use API specific > tools configure arranges to #define inline as empty if your compiler doesn't take it. Is that not working as expected in your case? regards, tom lane
Luis Amigo <lamigo@atc.unican.es> writes: > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > dbf2pgsql) there are inline variables, I think this is not ANSI > code. Standards evolve. -- Trond Eivind Glomsrød Red Hat, Inc.
Tom Lane wrote: > Luis Amigo <lamigo@atc.unican.es> writes: > > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > > dbf2pgsql) there are inline variables, I think this is not ANSI code. > > MIPS PRO compilers did not work with inline unless u use API specific > > tools > > configure arranges to #define inline as empty if your compiler doesn't > take it. Is that not working as expected in your case? > > regards, tom lane No it is not working as expected, I'd to remove them at my own.
Вложения
Trond Eivind Glomsrød wrote: > Luis Amigo <lamigo@atc.unican.es> writes: > > > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > > dbf2pgsql) there are inline variables, I think this is not ANSI > > code. > > Standards evolve. > > -- > Trond Eivind Glomsrød > Red Hat, Inc I think u can not make a unix standard open source dbase if u don't respect standards, OS must evolve, not applications
Вложения
On Wed, 30 Jan 2002, Luis Amigo wrote: > Trond Eivind Glomsrød wrote: > > > Luis Amigo <lamigo@atc.unican.es> writes: > > > > > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > > > dbf2pgsql) there are inline variables, I think this is not ANSI > > > code. > > > > Standards evolve. > > I think u can not make a unix standard open source dbase if u don't > respect standards, OS must evolve, not applications The recent C standards have inline. -- Trond Eivind Glomsrød Red Hat, Inc.
> > > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > > > dbf2pgsql) there are inline variables, I think this is not ANSI > > > code. > > Standards evolve. > I think u can not make a unix standard open source dbase if u don't > respect standards, OS must evolve, not applications Per Tom Lane's response, there are provisions in the configuration step to define "inline" as an empty macro for the preprocessor. So it just goes away from the source code for compilers which do not support it. Tom inquired whether that configuration step seems to not produce the right result for you (it sounds like it doesn't). Actually removing "inline" from the source code is not required. Try running configure again and check the results. Let us know what you find. - Thomas
Luis Amigo <lamigo@atc.unican.es> writes: >> configure arranges to #define inline as empty if your compiler doesn't >> take it. Is that not working as expected in your case? > No it is not working as expected, I'd to remove them at my own. Would you look at configure's test for this and find out why it fails to detect that inline doesn't work? As the only one with access to the problem compiler, you cannot expect anyone else to fix this; it's your responsibility to improve the configure test so it gets the right answer on your platform. regards, tom lane
Thomas Lockhart wrote: > > > > I've noticed that in postgresql 7.2b4 and in some contrib(I've seen > > > > dbf2pgsql) there are inline variables, I think this is not ANSI > > > > code. > > > Standards evolve. > > I think u can not make a unix standard open source dbase if u don't > > respect standards, OS must evolve, not applications > > Per Tom Lane's response, there are provisions in the configuration step > to define "inline" as an empty macro for the preprocessor. So it just > goes away from the source code for compilers which do not support it. > > Tom inquired whether that configuration step seems to not produce the > right result for you (it sounds like it doesn't). Actually removing > "inline" from the source code is not required. > > Try running configure again and check the results. Let us know what you > find. > > - Thomas > > ---------------------------(end of broadcast)--------------------------- > TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org by now in 7.2b4 configure is not working as expected, I will try b5 when I can. It may be because MIPS Pro compilers have inlining tools, but in cannot be used as in gcc inline is an special type, can not be used as integer inline ... Regards