From: Joshua Lock <josh@linux.intel.com>
To: bitbake-devel@lists.openembedded.org
Subject: Re: [PATCH 0/1][PULL] Hob2: A new implementation for Human Oriented Builder
Date: Thu, 23 Feb 2012 18:42:29 -0800 [thread overview]
Message-ID: <4F46F915.5010705@linux.intel.com> (raw)
In-Reply-To: <cover.1330004101.git.dongxiao.xu@intel.com>
On 23/02/12 05:48, Dongxiao Xu wrote:
> Hi Richard,
>
> This pull request is a new implementation for Human Oriented Builder, please help to review and pull.
Please, let's not call it the Human Oriented Builder. It was always just
Hob, not HOB - are we changing that now?
> Changes from previous pull requests:
> - Re-implemented a lot of code according to Belen's new GUI design.
This is really hard to review as a several thousand line patch... It's
difficult to know where what seems like an odd decision in the code was
made for a sane reason without any git history or code comments to
follow the logic of the development.
The fact that the patch was rejected by the mailing list is a good
indication that we need to separate into smaller logical commits.
I could review coding style, but that is probably pretty useless at this
point and so offer somewhat superficial comments below regarding UI
implementation with Gtk+.
First things first, I can't run the BitBake in the submitted branch:
joshual@shamshir:~/Projects/Yocto/poky/build [master *]
$ ../bitbake/bin/bitbake -u hob
Traceback (most recent call last):
File "../bitbake/bin/bitbake", line 258, in <module>
ret = main()
File "../bitbake/bin/bitbake", line 178, in main
ui_main = get_ui(configuration)
File "../bitbake/bin/bitbake", line 68, in get_ui
module = __import__("bb.ui", fromlist = [interface])
File "../bitbake/lib/bb/ui/hob.py", line 33, in <module>
from bb.ui.crumbs.hoblistmodel import RecipeListModel, PackageListModel
I found and ran a recent poky-contrib branch for Hob, though I've no
idea if it reflects what's in this patch, just so that I had something
to look at and offer feedback on.
The Build environment tab of the settings dialogue doesn't fit in the
allocated space - I had to resize the dialogue to see Save and Cancel
buttons. This is also true of the Others tab though less so.
These buttons don't follow the GNOME HIG in several ways, perhaps this
is intentional?
It's also worth noting that the visual style 'leaks' in the Settings
dialogue and I see standard Gtk light-bulb info icons rather than the
steely i displayed on the main page.
The 'Package list may be incomplete' dialogue is a very nice piece of
functionality but interface-wise I'd expect to see the buttons labelled
along the lines of 'Get full list' 'View existing'.
It seems like the Stop Build tracking isn't consistently reset between
builds so when I was playing around started & stopped a build then
played some more and started another build and tried to stop it I only
had the option to Force Stop or Cancel...
Taking a look at the code the first thing I notice is multiples of 5 for
spacing, why is that? I thought we'd agreed with Belen that we'd follow
the GNOME HIG for consistent visuals amongst Gtk+ apps?
Further, there are still hard-coded colour values. Are these issues on
the "to fix later" list?
I'm not suggesting these hold up merging - just asking the question. I
can play HIG zealot and write a patch once this is merged? I think it's
important we present a professional GUI which matches the visuals of the
host OS.
The buttons that act like tabs for changing between notebook pages (i.e.
on the 'Package Selection' view) took me a long time to figure out as
they look like they are depressed (mid-click) buttons for the inactive
tabs and ready to click buttons for the active tabs - confusing visual.
There's at least one instance of commented out code (a signal handler
connection in builder.py) - we can probably drop those, right?
Setting the size and position of windows on each launch (as in
builder.py) is generally considered to be hostile to users - it should
be sufficient to start with a sane default size and then as the UI
allows the user to resize it we should remember the size for the next
launch.
This UI has come on leaps and bounds and I look forward to seeing this
foundation polished over the coming weeks. Maybe I can even submit some
changes myself.
Hopefully this splash of feedback gives you something to work with...
Cheers,
Joshua
--
Joshua Lock
Yocto Project "Johannes factotum"
Intel Open Source Technology Centre
next prev parent reply other threads:[~2012-02-24 2:51 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-02-23 13:48 [PATCH 0/1][PULL] Hob2: A new implementation for Human Oriented Builder Dongxiao Xu
2012-02-23 13:51 ` Xu, Dongxiao
2012-02-24 2:42 ` Joshua Lock [this message]
2012-02-27 12:26 ` Wang, Shane
2012-02-27 23:15 ` Joshua Lock
2012-02-29 2:04 ` Wang, Shane
2012-02-24 18:17 ` Richard Purdie
2012-02-27 0:54 ` Xu, Dongxiao
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=4F46F915.5010705@linux.intel.com \
--to=josh@linux.intel.com \
--cc=bitbake-devel@lists.openembedded.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.