Re: acl, copy-file: First step towards more consistent function names

2024-08-25 Thread Pádraig Brady
On 25/08/2024 11:48, Bruno Haible wrote: Pádraig Brady wrote: I always thought the x...() variants indicate they might eXit(). Are there any other x...() variants that just output a diagnostic, and don't exit()? Indeed, we currently have three categories of x...() variants: - Most of them

Re: acl, copy-file: First step towards more consistent function names

2024-08-25 Thread Bruno Haible
Pádraig Brady wrote: > I always thought the x...() variants indicate they might eXit(). > > Are there any other x...() variants that just output a diagnostic, > and don't exit()? Indeed, we currently have three categories of x...() variants: - Most of them emit diagnostics and call exit (EXIT_

Re: acl, copy-file: First step towards more consistent function names

2024-08-25 Thread Collin Funk
Hi Pádraig, Pádraig Brady writes: > I always thought the x...() variants indicate they might eXit(). > > Are there any other x...() variants that just output a diagnostic, > and don't exit()? Originally I felt the same way. Since I generally think about xmalloc, xstrdup, which 99% of the time j

Re: acl, copy-file: First step towards more consistent function names

2024-08-25 Thread Pádraig Brady
On 24/08/2024 17:49, Bruno Haible wrote: In most Gnulib modules, when we have a function that does something and another function that does the same thing with diagnostics, the common convention is that the latter has an 'x' prefix. Such as for malloc — xalloc vasprintf — xvasprintf getc

Re: acl, copy-file: First step towards more consistent function names

2024-08-24 Thread Collin Funk
Hi Bruno, Bruno Haible writes: > In the name of consistency of the API, I'd like to migrate 'acl' and > 'copy-file' > to the common conventions: In a first step > qcopy_acl — xcopy_acl > > and then later (in a year or two): > copy_acl — xcopy_acl > > At some point, we can then also rename t

acl, copy-file: First step towards more consistent function names

2024-08-24 Thread Bruno Haible
In most Gnulib modules, when we have a function that does something and another function that does the same thing with diagnostics, the common convention is that the latter has an 'x' prefix. Such as for malloc — xalloc vasprintf — xvasprintf getcwd — xgetcwd striconv — xstriconv etc. Th