Re: Emacs time stamp hook

Fine with me, too.

Re: Alpine: useless-if-before-free: Exec format error

On 6/15/19 1:14 PM, Bruno Haible wrote: I found that they got in the way of my understanding the code Do you mean, we should highlight the comment about the program itself? Partly that - the first comment of a script should be very brief (typically one line) and say what the script does; it s

Re: Alpine: useless-if-before-free: Exec format error

Hi Paul, > One quibble: those long comments are > probably more trouble than they're worth. I heavily disagree: This piece of code needs to consider - the distinction between POSIX:2017 and earlier versions, - execlp/execvp vs. execl/execv, - GuixSD vs. traditional file system layouts, -

Re: Emacs time stamp hook

Paul Eggert wrote: > I've never liked those timestamps, and would like to see them go away. Fine with me: all of these files are in modules and therefore regularly update through gnulib-tool. Except for 'check-module', which is non-functional anyway for years (since the introduction of conditiona

Re: Emacs time stamp hook

Bruno Haible wrote: However, for files that are only ever copied through gnulib-tool's module system, I would not say that the time stamp is "really needed". I've never liked those timestamps, and would like to see them go away. I often try to avoid updating them even when I edit the files wi

Re: Alpine: useless-if-before-free: Exec format error

Thanks for looking into and fixing that. One quibble: those long comments are probably more trouble than they're worth. I found that they got in the way of my understanding the code, as my eyes glazed over when reading them. And when I finally forced myself to read one, I found that at least one

Re: Emacs time stamp hook

Darshit Shah wrote: > Do we really need these timestamp lines when everything is source-controlled > using git? They are useful in general for files that are not part of gnulib modules. For example, in some project I was using a config.guess+config.sub combination that was not updated in 8 years o

socket module and sc_prohibit_always_true_header_tests

Hi, at GNU Wget we are currently a bit puzzled about gnulib-tool's pro-tip #if HAVE_SYS_SOCKET_H # include #elif HAVE_WS2TCPIP_H # include #endif and 'make syntax-check' complaining about using HAVE_SYS_SOCKET_H: maint.mk: do not test the above HAVE__H symbol(s); with the corresponding gnul

Re: Emacs time stamp hook

* Bruno Haible [190615 14:02]: > Hi Darshit, > > > > my $VERSION = '2018-03-07 03:46'; # UTC > > > > Won't this break the emacs timestamp hook which apparently needs to be > > within > > the first eight lines of the file? Maybe either reduce the text, or move the > > hooks around. > > In some

Re: Emacs time stamp hook

Hi Darshit, > > my $VERSION = '2018-03-07 03:46'; # UTC > > Won't this break the emacs timestamp hook which apparently needs to be within > the first eight lines of the file? Maybe either reduce the text, or move the > hooks around. In some files, such as gnulib.html in the web pages, the time

Re: Alpine: useless-if-before-free: Exec format error

* Bruno Haible [190615 13:07]: > 2019-06-15 Bruno Haible > > Fix scripts to have valid executable format on Alpine Linux. > Reported by Tim Rühsen . > Idea by Paul Eggert. > * build-aux/useless-if-before-free: Use a prologue that starts with > '#!/bin/sh'. >

Re: Alpine: useless-if-before-free: Exec format error

Paul Eggert wrote: > suggests something like the > following instead. This would avoid having separate shell scripts and .pl > files, > which would be a plus: > > #!/bin/sh > #! −*−perl−*− > eval 'exec perl −x −wS $0 ${1+"$@"}' > if 0; This

Re: Alpine: useless-if-before-free: Exec format error

I wrote: > So, the ENOEXEC error is a bug in musl. But POSIX does not specify that the > command interpreter for scripts without shebang is /bin/sh; therefore IMHO > it would be good not to make this assumption. Correction: say

Re: Alpine: useless-if-before-free: Exec format error

Bruno Haible wrote: - Add a shell script useless-if-before-free that merely does this: #!/bin/sh exec perl "$0".pl "$@" suggests something like the following instead. This would avoid having separate shell scripts and .pl files, w