This will make it more obvious that these files should not be edited directly, and provide a hint for finding the real source file.
Details
Diff Detail
- Lint
Lint Skipped - Unit
Tests Skipped
Event Timeline
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
60–61 | This is not exactly correct. For example, both sha256-x86_64.S and sha512-x86_64.S are generated from sha512-x86_64.pl. |
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
60–61 | Oh, indeed. But, do you think (something like) this is a reasonable change with that addressed? |
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
60–61 | Yes. Maybe "Please do not modify. This file is auto-generated." kind of comment is enough, I think. |
- use multiple echos as already done for i386 .S generation
- remove the .cmt intermediate step
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
60–61 |
Oh, perhaps. Although I think it's convenient to include the source file especially where the .S basename does not match the .pl. Anyway, I don't have a strong feeling either way. |
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
59 | *.cmt must be removed from the above three lines. |
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
71 | The x86-64 script requires the second argument for the output file because it needs to know whether it is creating sha256-x86_64.S or sha512-x86_64.S. That's why the intermediate files were created in the first place. :-( |
secure/lib/libcrypto/Makefile.asm | ||
---|---|---|
145 | Sorry for nit-picking but a blank is in the wrong place. :-) |