On 2020-10-27 06:49, Michael Kerrisk (man-pages) wrote:
Hi Alex,
On 10/26/20 11:16 PM, Alejandro Colomar wrote:
Signed-off-by: Alejandro Colomar <colomar.6.4.3@xxxxxxxxx>
---
man7/system_data_types.7 | 39 +++++++++++++++++++++++++++++++++++++++
1 file changed, 39 insertions(+)
diff --git a/man7/system_data_types.7 b/man7/system_data_types.7
index 49ca2d8e1..4930aac8b 100644
--- a/man7/system_data_types.7
+++ b/man7/system_data_types.7
@@ -199,6 +199,45 @@ See also the
.I float_t
type in this page.
.RE
+.\"------------------------------------- error_t ----------------------/
+.TP
+.I error_t
+.RS
+Include:
+.IR <argz.h> ,
+or
+.IR <envz.h> .
+Alternatively,
+.IR <errno.h> .
+.PP
+An integer type
+used for error reporting.
+A value of 0 means success, and nonzero means failure.
+Portable programs should use
+.I int
+instead.
+.PP
+Conforming to:
+Present in glibc.
+It is not standardized by the C language standard nor POSIX.
+.PP
+Notes:
+C11 standardized a similar type as an optional extension:
+.IR errno_t .
+However,
+.I errno_t
+is not present in POSIX.
+.PP
+.I <errno.h>
+defines this type if the macro
+.B _GNU_SOURCE
+is defined before including the header.
+.PP
+See also:
+.BR argz_add (3),
+.BR envz_add (3),
+.BR feature_test_macros (7)
+.RE
.\"------------------------------------- fd_set -----------------------/
.TP
.I fd_set
So, I must say that I am skeptical about adding this type to the page.
There is a trade-off between having useful info and creating a page
that becomes so large that it becomes unwieldy.
error_t is an obscure [1], nonstandard type used by two API sets
(one of which is roughly a special case of the other). I'm so far
unconvinced of the value of adding this typr to the page.
Hi Michael,
Agreed.
That type should never be used IMO,
so I guess the line you added should be enough.
Cheers,
Alex
Cheers,
Michael
[1] Before Jonny Grant's mail yesterday, which I presume was
the inspiration for this patch, I had either neither heard of
the type or at least long ago forgotten about it.
