Thanks for your comments. I tried to follow them and updated the patch.
The tests (good gettimeofday, bad gettimeofday, no gettimeofday) still
work.
I have one question, though:
> where this code is removed from AC_FUNC_GETTIMEOFDAY_CLOBBER:
>
> if test $jm_cv_func_gettimeofday_clobber = yes;
I submitted a patch for glibc that, if installed, I believe would
allow us to sync with their md5 implementation. It include several
cleanups, some not already in gnulib (e.g., use of uint32_t instead of
md5_uint32). So I installed the below meanwhile.
FSF hasn't answered about the sha1 module..
I have installed this. It may look a bit redundant now, but will make
sense later on, compare how the file look and is used in GnuTLS.
Review most appreciated, as always.
2005-10-08 Simon Josefsson <[EMAIL PROTECTED]>
* gc.h: Add gc_hash and gc_hash_buffer.
* gc-gnulib.c (gc_
Ok to install?
Now that I read through the module again, I realize we have the same
'char*' vs 'void*' vs 'uint8_t*' issue here. To reiterate, 'uint8_t*'
causes problems when sooner or later the type ends up in public header
files. And 'char*' has the signed/unsigned trouble and the POSIX
assump
Simon Law <[EMAIL PROTECTED]> writes:
> Hello,
>
> I'm the maintainer for WvStreams, an LGPLed C++ networking library.
>
> I'd like to include gnulib's argp implementation, for portability
> reasons, but it depends on GPLed implementations of mempcpy, strchrnul,
> and sysexits.
>
> Looking at thos
Paul Eggert <[EMAIL PROTECTED]> writes:
> Simon Josefsson <[EMAIL PROTECTED]> writes:
>
>> I'm not sure whether the extract from RFC 1952 (a two line function)
>> warrant the big copyright comment, or whether borrowing that code is
>> OK.
>
> No, it's OK to borrow two lines of code. However it wo
Paul Eggert <[EMAIL PROTECTED]> writes:
> Thanks for explaining all that.
>
>> Is it ok to install the GC module as-is meanwhile?
>
> It's fine with me. One (hopefully easy) suggestion: perhaps your
> explanation could go into the source code somewhere, in a TODO area?
I added it to gc.h and ins
Bruno Haible <[EMAIL PROTECTED]> writes:
> Simon Josefsson wrote:
>> Ok to install?
>
>> +/* Compute CRC-32 value of LEN bytes long BUF, and return it. */
>> +extern unsigned long crc32 (char *buf, size_t len);
>> ...
>
> All 4 functions don't write into buf, therefore the 'buf' argument type shou
