Re: [PATCH] Silently lower nmerge; don't (sometimes incorrectly) range-check it.

> +/* Lower NMERGE if necessary, to fit within operating system limits. */ > +static void > +lower_nmerge_if_necessary (void) > +{ > + /* How many file descriptors are needed per input and output file. */ > + int fds_per_file = (compress_program ? 3 : 1); > + > + /* How many file descriptors w

[PATCH] Silently lower nmerge; don't (sometimes incorrectly) range-check it.

Here's a patch written by: Paul Eggert Nima Nikzad Max Chang Alexander Nguyen Sahil Amoli Nick Graham This patch fixes a bug in 'sort' where it incorrectly deduces the number of available file descriptors by doing the equivalent of 'ulimit -n'. This method is incorrect, since many file desc

root check failed for install-C-root

Every so often I run a make check, and (more eventually) a root make check. After a git pull & remake this morning, I ran a root check, and got a failure on install-C-root. I am unsure on what was intended, or if even this failure is real (I am not familiar, yet, with ginstall), but decided to pos

Re: [PATCH] build: allow ./bootstrap --srcdir=... to work with a git submodule

Pádraig Brady wrote: ... > Works fine here... > > $ git --version > git version 1.5.3.6 > $ ./bootstrap --gnulib-srcdir=/home/padraig/git/gnulib > $ git submodule status > 8a4dc70a0d4efd5f53abb1f7a1c24fe67915f96b gnulib (v0.0-1958-g8a4dc70) Thanks! __