-----BEGIN PGP SIGNED MESSAGE----- Hello, There's two updates for the SunOS port here.
The first change is for the OpenSSL SqueakSSL case of openssl 1.0.2, For the second change this is a "documentation" update to the HowToBuild Thanks, -----BEGIN PGP SIGNATURE----- iQEcBAEBCAAGBQJes8ijAAoJEAwpOKXMq1MamV8H/2MrggX6VzRxGCdP40MN5vyE You can view, comment on, or merge this pull request online at:https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/496 Commit Summary
File ChangesPatch Links:
— |
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
with the definition the compile aborts with the SunPro C compiler. platforms/unix/plugins/SqueakSSL/sqUnixOpenSSL.inc", line 122: operands have incompatible types: With the definition compilation works. The code sqo_sk_GENERAL_NAME_pop_free(sANs, (sk_GENERAL_NAME_freefunc)sqo_sk_free); expanded - with the old defintion of sk_GENERAL_NAME_freefunc - to: sk_pop_free ( ( ( _STACK * ) ( 1 ? ( sANs ) : ( struct stack_st_GENERAL_NAME * ) 0 ) ) , ( ( void ( * ) ( void * ) ) ( ( 1 ? ( ( void ( * ) ( void * ) ) sk_free ) : ( void ( * ) ( GENERAL_NAME * ) ) 0 ) ) ) ); with the new definition of sk_GENERAL_NAME_freefunc it expands to :
which the SUNpro compiler accepts. — |
In reply to this post by David T Lewis
The header opensslv.h has a definition # define OPENSSL_VERSION_NUMBER 0x100020ffL In my case I observed that I had to change --- a/platforms/unix/plugins/SqueakSSL/openssl_overlay.h -#define sk_GENERAL_NAME_freefunc void()(void) — |
In reply to this post by David T Lewis
Then it is strange this did not come up earlier — |
In reply to this post by David T Lewis
I don't have the issue with "gcc" or with "clang". Only with the SunPro C compiler. Technically speaking I find it interesting to compile with the SunPro C compiler although that I can use gcc as an alternative. By the way, the "alternative" that I also have is to patch the source before compilation, which is also fine by me, then I patch it for the SunPro C specific case. However I submit the issue "upstream" so to speak to see what you think of it, maybe another solution exists. I think the change from point to void to point to GENERAL_NAME looked fine to me, by the way that fix also works for me in the gcc and clang cases. — |
In reply to this post by David T Lewis
yeah, you might be right… — |
In reply to this post by David T Lewis
@cstes pushed 1 commit.
— |
In reply to this post by David T Lewis
-----BEGIN PGP SIGNED MESSAGE----- This is a reminder about Currently I apply manually this small patch before building the My understanding from the feedback so far "let's keep it", That would be useful as I don't have to manually apply the patch any longer. It's not a disaster if it cannot be merged into Cog for some reason, Thanks, -----BEGIN PGP SIGNATURE----- iQEcBAEBCAAGBQJfIvoPAAoJEAwpOKXMq1Ma8KMIAIVrViG/cNkzqHRp7B0HVCvI — |
In reply to this post by David T Lewis
Reminder about this patch submitted a few months ago : Change void * to GENERAL_NAME* in : -#define sk_GENERAL_NAME_freefunc void()(void) — |
In reply to this post by David T Lewis
thanks. sorry… — |
In reply to this post by David T Lewis
Merged #496 into Cog. — |
Free forum by Nabble | Edit this page |