Name space collision

We've created in the openCryptoki project a software token which uses openSSL's crypto routines... When running under the iplanet regression test suite we get a core dump in the key generation code... Tracing it down, the crash occurs in the SHA1_Update... one of the community membe

Re: Name space collision

Steven Bade wrote: > We've created in the openCryptoki project a software token which > uses openSSL's crypto routines... When running under the iplanet > regression test suite we get a core dump in the key generation code... > Tracing it down, the crash occurs in the SHA1_Update... one

RE: Name space collision

-Original Message- From: Tom Wu [mailto:[EMAIL PROTECTED]] Sent: Fri 04/05/2002 6:25 AM To: [EMAIL PROTECTED] Cc: Subject:Re: Name space collision Steven Bade wrote: > We've created in the openCryptoki project a software token which > uses openSSL's

Re: Name space collision

> Thanks in advance > Fred > > > -Original Message- > From: Tom Wu [mailto:[EMAIL PROTECTED]] > Sent: Fri 04/05/2002 6:25 AM > To: [EMAIL PROTECTED] > Cc: > Subject: Re: Name space collision > > Steven Bade wrote: > >>We've creat

Re: Name space collision

Thanks, but I don;t think that in this case this will help since the particular use of the function is down in the deep libcrypto code itself so creating macros for our calls internaly would not do much good... I've been able to get past some of these types of problems in the past with -Wl,

Re: Name space collision

Here is a patch I created for the Makefile.org file to correct the problem we were having... Steven Bade wrote: > It seems that building with -Wl,-Bsymbolic on the linker options for > creating the libcrypto.so makes this issue go away... > > Tom Wu wrote: > >> Steven Bade wrote: >> >>>

Re: Name space collision

> > Any of the more experienced openSSL developers out there have any > > suggestions of things we may try with regard to linker options or other > > ideas that we could do to address this collision (I doubt that I;m > > going to be able to get netscape to chance their function names :)

Re: Name space collision

> Here is a patch I created for the Makefile.org file to correct the > problem we were having... I wouldn't like to see this in OpenSSL by default; there are times when you *want* to over-ride functions inside the library in your own program, and times when you don't. I'd advise you instead to

Re: Name space collision

Harald... We build our software token with -Bsymbolic. In fact all the shared objects we create are with -Bsymbolic. Because of licensing issues I can't staticaly bind with libcrypto (and even under debug scenarios, it turns out that binding static to libcrypto.a the problem occurs as well s

Re: Name space collision

From: Steven Bade <[EMAIL PROTECTED]> sbade> It seems that building with -Wl,-Bsymbolic on the linker options for sbade> creating the libcrypto.so makes this issue go away... Hmm, I guess that's a change to be made in Makefile.org. However, except for GNU ld, I have no idea what the option to

Re: Name space collision

From: Steven Bade <[EMAIL PROTECTED]> sbade> Here is a patch I created for the Makefile.org file to correct the sbade> problem we were having... Thanks. It will be applied immediately. -- Richard Levitte \ Spannvägen 38, II \ [EMAIL PROTECTED] Redakteur@Stacken \ S-168 35 BROMMA \ T: +4

Re: Name space collision

From: Harald Koch <[EMAIL PROTECTED]> chk> > Here is a patch I created for the Makefile.org file to correct the chk> > problem we were having... chk> chk> I wouldn't like to see this in OpenSSL by default; there are times when chk> you *want* to over-ride functions inside the library in your ow

Re: Name space collision

> Hmm, why would one want to do that? Out of curiosity... The first example that leaps to mind is the IPv6 bump-in-the-api shim library; See, for example: The idea is to pre-load a shared library that

Re: Name space collision

From: Harald Koch <[EMAIL PROTECTED]> chk> > Hmm, why would one want to do that? Out of curiosity... chk> chk> The first example that leaps to mind is the IPv6 bump-in-the-api shim chk> library; See, for example: chk> chk> chk>

Re: Name space collision

thanks for taking the patch into consideration... When I get back to work I'll see if i can;t find the AIX Linker options and pass them on... Richard Levitte - VMS Whacker wrote: > From: Steven Bade <[EMAIL PROTECTED]> > > sbade> It seems that building with -Wl,-Bsymbolic on the linker option

AW: Name space collision

e [mailto:[EMAIL PROTECTED]] > Gesendet: Freitag, 5. April 2002 16:05 > An: [EMAIL PROTECTED] > Betreff: Re: Name space collision > > > Thanks, but I don;t think that in this case this will help since the > particular use of the function is down in the deep libcrypto > cod