Minor change to m4/argp.m4

2008-02-29 Thread Sergey Poznyakoff
Hello, I have installed the following patch: 2008-02-29 Sergey Poznyakoff [EMAIL PROTECTED] * m4/argp.m4 (gl_ARGP): Use AC_TRY_LINK to test if program_invocation_name and program_invocation_short_name are present. diff --git a/m4/argp.m4 b/m4/argp.m4 index

new module 'freadseek'

2008-02-29 Thread Bruno Haible
Here's the side-effecting companion of freadptr. 2008-02-29 Bruno Haible [EMAIL PROTECTED] New module 'freadseek'. * modules/freadseek: New file. * lib/freadseek.h: New file. * lib/freadseek.c: New file. * MODULES.html.sh (File stream based

closein, freadahead and ungetc

2008-02-29 Thread Bruno Haible
Hi Eric, closein.c assumes that if freadahead() == 0, the stream has no buffered contents, i.e. an fseek will not have to reposition the underlying file descriptor with lseek. This is not the case. Here is a test program: foo.c

__builtin_expect [was: __func__]

2008-02-29 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Bruno Haible on 2/28/2008 1:43 PM: | I was unsure where to document this, it isn't either a function or | header file, strictly speaking. | | I would put its documentation in chapter Particular modules. If/when | the 'inline' module gets

Re: closein, freadahead and ungetc

2008-02-29 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Bruno Haible on 2/29/2008 3:49 AM: | With strace you can see that there is only a single read() from the file | descriptor. I.e. at the moment when freadahead() returns 0, it is only | the ungetc-backup-buffer which has been emptied; the