This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Build failure on Linux - strnicmp undefined


muller@cerbere.u-strasbg.fr wrote:
> 
> At 15:03 01/12/00 -0500, Fernando Nasser wrote:
> >Thanks folks.  This works.
> >I will check it in momentarily.
> >
> >
> >* p-exp.y: Define strncasecmp as strnicmp for MSVC.
> >  (yylex): Use strncasecmp, not strnicmp.
> 
>   Thanks for checking it in. I only realized this after committing it
> when I tried to compile it also on a Linux machine.
> 
>   However, I also found strnicmp in
> src/gdb/nlm/gdserve.c
> so I wasn't sure that I should do the change.

I think this can go.

> On the other side you can find in src/gdb/ser-e7kpc.c
> 
> that strncasecmp is replaced by strnicmp by a macro  for MSVC
> but several other files using strncasecmp do not have this macro
> defined. Unless this is automatically created by the configure execution,
> in which case the defines inside the c source files
> would be unnecessary.

(The #ifdef MSVC can go.)

Yes, libiberty provides the function if it determines that the host
system doesn't have it.  GDB should always be using strncasecmp().

>   Concerning the date of the ChangeLog entry, it was the date of the
> patch RFA on gdb-patches list. As it was only delayed for
> formatting reasons I thought that I should better not modify
> the date.
>   I recieved no comment about strnicmp use when I submitted the RFA.
> 
>   One other strange fact is that the linux build failure was only on
> link time, is this normal ?
> Shouldn't the compiler complain about undefined function ?
> (maybe I didn't see a warning as I didn't look at the compilation progress
> on the linux box).

You'll only get a compile time warning if you configure with something
like:

	--enable-gdb-build-warnings=-Werror

>   I apologize deeply for the troubles I created, but
> couldn't realize before the strnicmp was not a standard
> function.

It is hard to avoid.  Your testing can only go so far :-(

>   Is there a way to avoid doing such errors again in the future ?

That is a more generic question, I've some thoughts but it is probably
better to start a new thread for that on gdb@sourceware.

	enjoy,
		Andrew

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]