Document that fs.futimes isn't supported on AIX

[gibfahn]

• Version: master
• Platform: AIX
• Subsystem: fs

fs.futimes isn't supported on AIX <7.1, see #12583 (comment)

This is mentioned in the libuv docs:

Note AIX: This function only works for AIX 7.1 and newer. It can still be called on older versions but will return UV_ENOSYS.

Do we normally duplicate such warnings in the Node docs?

> fs.futimesSync(fd, s2.atime, s2.mtime)
Error: ENOSYS: function not implemented, futime
    at Object.fs.futimesSync (fs.js:1216:11)
    at repl:1:4
    at ContextifyScript.Script.runInThisContext (vm.js:44:33)
    at REPLServer.defaultEval (repl.js:239:29)
    at bound (domain.js:301:14)
    at REPLServer.runBound [as eval] (domain.js:314:12)
    at REPLServer.onLine (repl.js:433:10)
    at emitOne (events.js:120:20)
    at REPLServer.emit (events.js:210:7)
    at REPLServer.Interface._onLine (readline.js:262:10)

[Trott]

@nodejs/documentation @nodejs/platform-aix

[gireeshpunathil]

I think we should document this in Node docs as well due to: i) The public API with the existing one line description makes it looks like platform agnostic, whereas AIX seem to be an exception. ii) There is no visible links which leads from node doc to libuv doc in this page.

[aaronbartell]

It should be noted the same error occurs for the IBM i operating system. The IBM i OS has an integrated environment named PASE which is essentially a port of AIX that uses IBM i as its kernel. I mention this to let the Node.js community know the AIX base is growing and will hopefully be considered when things like this come up.

[gireeshpunathil]

@aaronbartell - thanks for the information. Will you please also describe:

• What is the current process of building node on PASE?
• What amount of code changes are required for the port?
• What effort it takes to have node on PASE to be fully integrated into the community?
• Do you recommend the doc changes (doc: note that fs.futimes only works on AIX >7.1 #13659) to cover IBM i as well? In my opinion it will be bit awkward for documenting functional description of an API in a platform which the community does not cover at the moment, but consensus may be the key, with a one liner caveat on i as supported by IBM at present, not the community.

[aaronbartell]

What is the current process of building node on PASE?

IBM builds and distributes Node.js for IBM i. It is not currently open source but they have made declaration of intention to upstream their changes to this repo. I do not work for IBM though I collaborate with them a fair amount concerning Node.js on IBM i.

What amount of code changes are required for the port?

I am unsure of this per the above. I am mentioning @ThePrez (Jesse Gorzinski) who is the owner of Node.js on IBM i at IBM.

What effort it takes to have node on PASE to be fully integrated into the community?

Once the IBM i Node.js port is open sourced then any community member with access to an IBM i could contribute changes. Litmis Spaces is a free/hosted IBM i that operates on preconfigured chroot environments (i.e. Node.js, Ruby On Rails, Python) and that platform could be used by the community to participate in furthering Node.js on IBM i, including PRs for furthering this repo. DISCLAIMER: I work for Krengeltech, the company that hosts Litmis Spaces.

Do you recommend the doc changes (#13659) to cover IBM i as well? In my opinion it will be bit awkward for documenting functional description of an API in a platform which the community does not cover at the moment, but consensus may be the key, with a one liner caveat on i as supported by IBM at present, not the community.

I would defer to your judgement on this matter. IBM i folks currently have to supplement their education of Node.js by going to other sites, so it would be great if these things were formally documented by Node.js.

Let me know if there is other information I can provide.

[gireeshpunathil]

@aaronbartell - thanks for the detailed info. @mhdawson, @gib - FYI and consideration in terms of the doc changes.

[ThePrez]

@aaronbartell, have you verified this failure on IBM i 7.3?

I would like to first pursue this as a Node.JS defect before documenting a permanent restriction.

[gibfahn]

@ThePrez is IBM i 7.3 equivalent to AIX 7.3? If so, then as the original issue states:

Note AIX: This function only works for AIX 7.1 and newer. It can still be called on older versions but will return UV_ENOSYS.

so I'd expect it to work on that version.

[aaronbartell]

have you verified this failure on IBM i 7.3?

My v7.3 box is currently being updated. Will run test as soon as it is done.

is IBM i 7.3 equivalent to AIX 7.3?

No, IBM i v7.3 isn't equivalent to AIX v7.3. You can see the relationship here, or review the below screenshot. In short, IBM i adopts AIX into PASE one or two releases behind.

[aaronbartell]

@ThePrez

I have confirmed it doesn't work on IBM i v7.3

$ npm -v
3.10.8
$ node -v
v6.9.1
$ node futimes.js
{ Error: ENOSYS: function not implemented, futime
    at Error (native) errno: -109, code: 'ENOSYS', syscall: 'futime' }
$ uname -rv
3 7

Content of futimes.js

var fs = require('fs');
var fd = fs.openSync(__filename, 'r');
fs.futimes(fd, new Date(), new Date(), function(err) {
   console.log(err);
});

Thoughts?

[KerimG]

@aaronbartell @ThePrez

Familiar faces!

I just stumbled across this issue today. But for what it's worth:

gueney:~/code $ node fut
{ Error: ENOSYS: function not implemented, futime errno: -109, code: 'ENOSYS', syscall: 'futime' }
gueney:~/code $ node -v && npm -v && uname -a
v8.11.3
6.2.0
OS400 [redacted] 2 7 [redacted]

[KerimG]

FYI, running @ThePrez 's futimes.js on Windows causes the following error:

$ node fut
{ Error: EPERM: operation not permitted, futime errno: -4048, code: 'EPERM', sys
call: 'futime' }
$ node -v && npm -v && uname -a
v8.11.1
6.1.0
MINGW64_NT-10.0 DESKTOP-6UQOLB7 2.10.0(0.325/5/3) 2018-06-13 23:34 x86_64 Msys

Tried running it with Administrator but get the same error.