From: Friedrich Weber <f.weber@proxmox.com>
To: Stefan Sterz <s.sterz@proxmox.com>,
Proxmox VE development discussion <pve-devel@lists.proxmox.com>
Subject: Re: [pve-devel] [PATCH widget-toolkit 3/3] window: edit: avoid shared object for extra request params
Date: Thu, 4 Apr 2024 12:10:12 +0200 [thread overview]
Message-ID: <147ee23c-97ae-4cc5-8c1f-ffc2ad4d6773@proxmox.com> (raw)
In-Reply-To: <D0B89ODDN2EF.32UYX8ZQQZHU5@proxmox.com>
On 04/04/2024 11:23, Stefan Sterz wrote:
> -- >8 snip 8< --
>>>
>>> i did a quick an dirty test and using a constructor like this seems to
>>> rule out this class of bug completelly:
>>>
>>> ```js
>>> constructor: function(conf) {
>>> let me = this;
>>> me.extraRequestParams = {};
>>> me.initConfig(conf);
>>> me.callParent();
>>> },
>>> ```
>>>
>>> basically it configures the edit window as usual, but overwrites the
>>> `extraRequestParams` object for each instance with a new empty object.
>>> so there are no more shared objects :) could you check whether that also
>>> fixes the other instances?
>>>
>>> [1]: https://docs-devel.sencha.com/extjs/7.0.0/classic/Ext.window.Window.html#method-constructor
>>
>> Nifty, didn't think about a constructor solution. Such a general
>> solution would be way more elegant, thanks for suggesting it!
>>
>> However, this particular constructor seems to break the pattern of
>> defining `extraRequestParams` in the subclass properties, as done by
>> `PVE.Pool.AddVM` [1]. With the constructor above, the API request done
>> by `AddVM` seems to be missing the `allow-move` parameter.
>>
>> Looks like once `PVE.Pool.AddVM` is instantiated and the constructor is
>> called, `extraRequestParams` with `allow-move` is only defined in
>> `me.__proto__`, so `me.extraRequestParams = {}` essentially shadows it
>> with an empty object, losing the `allow-move`.
>>
>
> not sure what you mean by that, if an `PVE.Pool.AddVM` is instantiated,
> the `extraRequestParams` is already set, so it isn't just in `__proto__`
> for me. but yeah, the problem is correct as `me.extraRequestParams = {}`
> overwrites the field.
I agree it doesn't matter here, but just for completeness, I meant that
if I set a breakpoint before line 2, so before the overwrite:
```js
constructor: function(conf) {
let me = this;
=> me.extraRequestParams = {};
me.initConfig(conf);
me.callParent();
},
```
... `extraRequestParams` is not a property of `me`, but inherited from
its prototype:
```
>> me.extraRequestParams
Object { "allow-move": 1 }
>> "extraRequestParams" in me
true
>> Object.hasOwn(me, "extraRequestParams")
false
```
Doesn't make a difference for the overwrite, though.
>> Do you have an idea how to fix this? Maybe making a copy of
>> `extraRequestParams` would work (I suppose the overhead of creating a
>> new object for all edit window (subclass) instances is negligible).
>>
>> [1]
>> https://git.proxmox.com/?p=pve-manager.git;a=blob;f=www/manager6/grid/PoolMembers.js;h=75f20cab;hb=4b06efb5#l9
>
> this worked for me, can you confirm that this also does what it should
> for you?
>
> ```js
> extraRequestParams: undefined,
>
> constructor: function(conf) {
> let me = this;
> if (!me.extraRequestParams) {
> me.extraRequestParams = {};
> }
> me.initConfig(conf);
> me.callParent();
> },
> ```
It works in the sense that it fixes the bug mentioned in my patch 1/3,
and fixes the lost `allow-move` issue from the previous constructor. But
with this constructor, all instances of `AddVM` share the same
`extraRequestParams` (the body of the `if` never gets executed for
`AddVM` instances), which is the condition that my patch 2/3 tries to
avoid (even though it is currently not buggy).
Maybe we could do:
```js
extraRequestParams: {},
constructor: function(conf) {
let me = this;
me.extraRequestParams = Ext.clone(me.extraRequestParams);
me.initConfig(conf);
me.callParent();
},
```
... which, if I'm not missing anything, *should* cover everything (with
the cost of allocating unnecessary empty objects)?
next prev parent reply other threads:[~2024-04-04 10:10 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-03 9:10 [pve-devel] [PATCH manager/widget-toolkit 0/3] ui: avoid UI bugs due to shared " Friedrich Weber
2024-04-03 9:10 ` [pve-devel] [PATCH manager 1/3] ui: pool members: avoid setting request parameter for all edit windows Friedrich Weber
2024-04-03 9:10 ` [pve-devel] [PATCH manager 2/3] ui: pool members: avoid sharing object for extra request parameters Friedrich Weber
2024-04-03 9:10 ` [pve-devel] [PATCH widget-toolkit 3/3] window: edit: avoid shared object for extra request params Friedrich Weber
2024-04-04 8:22 ` Stefan Sterz
2024-04-04 9:01 ` Friedrich Weber
2024-04-04 9:23 ` Stefan Sterz
2024-04-04 10:10 ` Friedrich Weber [this message]
2024-04-04 10:54 ` Stefan Sterz
2024-04-04 11:04 ` Stefan Sterz
2024-04-04 10:59 ` Thomas Lamprecht
2024-04-04 11:28 ` Friedrich Weber
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=147ee23c-97ae-4cc5-8c1f-ffc2ad4d6773@proxmox.com \
--to=f.weber@proxmox.com \
--cc=pve-devel@lists.proxmox.com \
--cc=s.sterz@proxmox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox