Roland Becker | 28 Aug 10:04 2014
Picon

Re: Introducing MantisBT Modern UI

Victor,
maybe you can have a look at the other open PR's where we voted +1 and merge
them to master if they are ok for you.
If not, there might be another round of time consuming tasks for Rafik.

> Victor Boctor <victor <at> mantishub.net> hat am 28. August 2014 um 05:26
> geschrieben:
>
>
> When you get to it, I would try to different approaches:
>
> - Squash all your changes into a new branch, then try merging with that.  That
> will remove the spacing issue.  I wonder if having the early commits with
> these changes would cause more conflicts compared to the squashed changes
> which reverted these spaces manually.
>
> - Try merging to the current branch as is.
>
> The result of this experiment will help guide the approach that we should use
> when getting this into master.  The goal is to simplify the process of merging
> between 1.3 and 2.0 as much as possible.
>
> On Aug 27, 2014, at 8:18 PM, Rafik Robeal
> <rafik <at> mantishub.net<mailto:rafik <at> mantishub.net>> wrote:
>
> Hi Roland,
>
> I saw you comment and I noticed that differences with master while doing the
> cleanup.
>
(Continue reading)

Robert Munteanu | 25 Aug 22:28 2014
Picon

Alternative UI proposal (was: Introducing MantisBT Modern UI)

Hi Paul,

On Mon, Aug 25, 2014 at 8:17 PM, P Richards <paul-P/naLZWKUS8A3RpLMO19ww@public.gmane.org> wrote:

I’m going to spent an hour now and see how easy it is to theme the same page with uikit, to get an idea of file size.

 


So here's your own thread :-) Please don't hijack the discussion about Rafik's proposal.

Robert
------------------------------------------------------------------------------
Slashdot TV.  
Video for Nerds.  Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
Robert Munteanu | 19 Aug 14:38 2014
Picon

Using bootstrap in MantisBT ( was: Introducing MantisBT Modern UI )

Hi,

I'd like to split the discussion into its own separate thread, both
for keeping the discussion on the new theme separate and for having a
clear link to point to the results of the discussion on including
Boostrap.

Here's the 'checklist' that I proposed when reviewing external libraries [1]

- Technical fit

I believe boostrap is one of the most used and productive frontend
frameworks. As such, it is a good technical fit for MantisBT.

- License compatibility

License is MIT [2], I see no conflicts here.

- Recent development activity

9k+ commits, 500+ contributors according to [3]

- Community size/support

Community support via irc and stackoverflow, accoring to [4] .

I think all four criteria are met, and boostrap can be included in
MantisBT, so here's my '+1'.

On a side note, I think that if Rafik is more familiar and confident
using bootstrap, we should follow his approach. I don't see any value
in debating alternative frontend frameworks ; we don't have a terribly
complex use case, and any of them would do frankly. If Rafik chose
bootstrap, so be it.

Robert

[1]: https://www.mantisbt.org/wiki/doku.php/mantisbt:development_process#changing_external_libraries
[2]: https://github.com/twbs/bootstrap/blob/master/LICENSE
[3]: https://github.com/twbs/bootstrap
[4]: https://github.com/twbs/bootstrap#community

On Tue, Aug 19, 2014 at 8:25 AM, Rafik Robeal <rafik <at> mantishub.net> wrote:
> With respect to Bootstrap vs. other frameworks:
>
> There are a couple of popular web frameworks: Bootstrap & Foundation
>
> + If you scan both bootstrap & foundation docs, you will see the same
> features, similar markup and same web components.
>
> + Bootstrap has a much bigger ecosystem. It came first and was backed by
> twitter. It captured developers mindshare early on and became the de-facto
> standard web framework as these recent stats show:
> http://techcrunch.com/2013/07/28/bootstrap-3-goes-mobile-first-now-reportedly-powers-1-of-the-web/
> http://blog.meanpath.com/4-7m-sites-using-bootstrap-vs-334k-on-foundation/
>
> As such, bootstrap has the lion share of docs, plugins, free and commercial
> themes (like the one used with Mantis) … etc
>
> + Other than that, it is hard to differentiate between both frameworks. Here
> are couple of different opinions to chew on:
> https://medium.com/ <at> felippenardi/top-5-core-differences-between-bootstrap-3-and-foundation-5-8b3812c7007c
> http://blog.teamtreehouse.com/use-bootstrap-or-foundation
>
>
> What about JQuery Mobile?
> It is a great framework for building mobile user interfaces. I started using
> it when it was in early alpha and was a big fan. That’s until Bootstrap v3.0
> came along. Bootstrap v3 is far superior framework for building responsive
> interfaces that works on desktop, tablet and mobile.
>
>
> What about Theme Roller?
> That’s not a framework. It is a feature available in jqueryui.com to allow
> users to customize the look and feel of jqueryui components. The output of
> the theme roller is a custom css file to apply alongside the core jqueryui
> library. If you want the same for mantis, you need to build it. It won’t
> matter if the underlying framework is Bootstrap or Foundation.
>
>
>
> So far, the works been done by someone that has never done a commit to
> mantis before
>
>
> There is always a first time
>
>
> The new layout would look great for a service (like mantis hub),
>
>
> Nothing in this work was customized for MantisHub.
> The sidebar design is increasingly popular choice (Github, Jira, BitBucket)
> because of the popularity of widescreen monitors. With widescreen, you have
> more horizontal space to work with and less vertical space. Using vertical
> space to show a navigation menu with 5 links is just a waste of precious
> user work area.
>
>
> -Rafik
>
>
> On Aug 18, 2014, at 1:33 PM, P Richards <paul <at> mantisforge.org> wrote:
>
> My main concern is that we evaluate options properly - any decision we take
> now is something that the whole dev-team and any plugin author will be
> using.
>
> So far, the works been done by someone that has never done a commit to
> mantis before - whilst it's fairly clear that they are not stupid, at the
> same time, there's historical decisions and historical roadmap items and
> plans etc that may not have been considered. It may be that after this work,
> we never see the author again. Our published contribution guidelines state
> to talk to the dev list first etc etc etc.
>
> Hopefully Rafik will submit a PR with whitespace tidied up and splitting out
> some of the non-design changes (i.e. jquery) so we can actually have a
> proper look at the work, but there's decisions in that PR that we've already
> discussed and decided on a path 'for now'.
>
> For example,
>
> In october 2013 Victor send a mail about upgrading JQuery from 1.7.2 to 1.9
> or 2.0
> Robert replied stating that moving to 2.0 would rule out IE8
> Etc
> I replied with what I deemed a favourite question: "what browsers do *we
> support*? :)"
> A Mantis user replied to the converstation stating about being careful about
> browser support as they only have IE8
> Robert replied stating IE8 is here to stay.
>
> I might have got the order slightly wrong in the above, but the end outcome
> I believe was we updated master from jquery 1.7.2 to 1.9.1
>
> If I google for "IE market share" - the first hit I get is
> http://www.netmarketshare.com/browser-market-share.aspx?qprid=2&qpcustomd=0,
> which states IE8 as having 22% share (not sure how they generate these stats
> as that seems nut ;)) IF I go for the 2nd hit
> (http://www.w3counter.com/globalstats.php), that shows chrome with 20% etc,
> IE9 6.5%, IE11 5.6%, and IE8 4%
>
> My understanding is bootstrap 3 would support IE 8 for *most* things, and
> that jquery 2.x does not support IE8 - hence my point about breaking the PR
> up a bit, and also it needing to have a discussion.
>
> The other part is ensuring that we map design changes to something that is
> flexible for as many existing users to use and fit into their existing setup
> as possible. The new layout would look great for a service (like mantishub),
> it may not look as good for people that are trying to top/tail it to use it
> in existing environments. We need to ensure that our approach provides
> flexibility for both - this is something that mantis arguably does well for
> other things due to our complexity of configuration options.
>
> Paul
>
>
>
> -----Original Message-----
> From: Roland Becker [mailto:roland <at> atrol.de]
> Sent: 18 August 2014 21:01
> To: developer discussions; Paul Richards
> Subject: Re: [mantisbt-dev] Introducing MantisBT Modern UI
>
> Paul,
>
> it seems your main concern is using Bootstrap
>
> We set a policy in january that before we accept any external library
> changes etc, we need to have a discussion on the mailing list.
> And that’s aside from the fact that we need to have a discussion
> around whether to use bootstrap.
>
>
> Let's start the discussion and get a decision: Use Bootstrap or not.
>
> I can't contribute that much to this decision as I have no own experience
> using Bootstrap or any other comparable component.
> I don't want to vote just because I read a few articles and comparisons
> about it.
>
> What I noticed is a big amount of transferred CSS and JavaScript compared to
> the current version. (about additional 700KB when opening for the first
> time, but also more traffic for following requests)
>
> Are there any known issues using Bootstrap?
>
> Roland
>
> Paul Richards <paul <at> mantisforge.org> hat am 18. August 2014 um 09:16
> geschrieben:
>
>
> On 18 Aug 2014, at 01:00, Victor Boctor <vboctor <at> gmail.com> wrote:
>
> Paul,
>
> On Aug 17, 2014, at 2:23 AM, P Richards <paul <at> mantisforge.org> wrote:
>
> Rafiq,
>
> As i’ve seen on the PR, and when you originally said you were looking
> at this a month ago, and asked you to mail the Mailing List, we need
> to have a discussion about the design and the approach.
>
>
> [Victor] Correct.  The request from you when Rafik sent out a preview
> was to publish the code and the discussion to the DL, which he just
> did.  Your words were that even though the decision to use bootstrap
> is likely to be a short discussion, it is worth having anything.
>
>
> [Victor] The bootstrap decision really started when the website was
> developed.  At this time, bootstrap was used and the request by the
> team from the Rafik was to do the same modernization to MantisBT.
> Hence, he started doing the work.  As for plugins, they are typically
> easy to adapt to whatever framework MantisBT is using.
>
>
> Actually, I said we should have a be having a discussion on the
> mailing list and before writing code:
>
> "Whilst I suspect that is probably a good choice, we should have been
> having a discussion on whether bootstrap is the way to go with Mantis
> or something else before writing any code.”
>
> We set a policy in january that before we accept any external library
> changes etc, we need to have a discussion on the mailing list.
>
> And sorry, the website was something that was also done without
> discussion initially - so that’s not really a valid point. The website
> has always been a separate code base to Mantis.
>
>
> One of the main requests from users in the past, is that users want
> slightly different thing’s - i.e. people want to be able to customise
> the look and feel of mantis in different ways.
>
>
> [Victor] I’m not personally in favor of too much customization to the UI.
> In my opinion, a lot of the users that heavilty themed Mantis in the
> past  were motivated to improve the look.  However, what we really need
> to focus  on is the ability for users to continue to be able to extend
> MantisBT via  plugins.  If we get into theming, we can support color
> shade (e.g. green  template, blue template, etc) rather than complete
> customization.  Complete  customization adds little value and can
> significantly increase complexity  of code and test matrix.  I believe
> users evaluate MantisBT evalute it  based on functionalty, look, and
> extensions to fill gaps that we don’t have  in core or integrate their own
> custom functionality.
>
>
> We’ve had a lot of questions over the years to allow alteration of
> bits of Mantis. What we’ve always done is been flexible with the
> approach - hence the fact today we have configuration variables for
> everything.
>
> If you consider the design of Mantis and the number of pages, we have
> 7 main things users use / want to customise:
>
> a) header
> b) menu bar
> b) page content -> view issue
> c) page content -> view issues
> d) page content -> report issue
> e) page content -> my view page
> f) footer
>
> When people have gone in the past “it should be possible to
> template/theme whole of mantis’, personally - I can’t see people doing
> that, or it being sustainable for us.
>
> However, I can see companies that want to brand mantis replacing A+F
> in the list above, and potentially B
>
> In terms of B/C/D/E - I suspect you would not touch the design of C/E,
> but the ability to tweak B+D is where most drivers come from.
>
> For example, I scrapped the report issues page at work a long term ago
> as it was not suitable for non-technical users. And within Mantis at
> the time, the flexibility we provide for customising it does not
> exist. This is actually one block of Mantis where i’d agree with the
> people wanting support for templates.
>
> View Issue I didn’t touch as it’s mainly technical staff that use it,
> however if the ability existed to template it properly, would be tweaked.
>
>
>
>
> In it’s current form, this PR isn’t suitable for this - For example -
> whilst it’s nicer, the reason I use mantis is due to the way it can be
> integrated into other systems - so the side bar is a no go.
>
>
> [Victor] I’m not sure what you mean by integrating MantisBT in other
> systems and why the side bar blocks that?
>
>
> Mantis has always been a top down design - if you replace the
> header/footer you can embed the remaining content within another page,
> leaving the menu at the top for navigation.
>
>
>
> For most end users, Mantis consists of 4 pages:
>
> 1) My View Page
> 2) View Issues Page
> 3) Report Issues Page
> 4) View Issue Page
>
> I think we can do a better job then the current theme if we focus on
> looking at the design and the ability to customise these 4 pages.
>
>
> For example, in a modern system, the “My View” page I would have
> expect to be a ajax dashboard.
>
>
> I sent an email a few weeks ago now for feedback on using DataTables
> to ajaxify the View Issues list.
>
>
>
> [Victor] This is just a scoping exercise.  The first iteration’s goal
> is to improve the look and use a more modern code styling.  Once this
> is in, we can follow up by enriching specific page to make it use
> ajax.  Good examples are “My View” and filter page.  Once Rafik’s
> change is checked in, we can all collaborate on enriching specific pages as
> necessary.
>
> For 3/4, the requests tend to be to allow fields to be re-ordered,
> turned on/off.
>
> At the moment, the work doesn’t help any of these requests from users.
>
>
> [Victor] Agreed on these features, but think of Rafik’s change as a
> foundation work for these features.  I don’t see how there is a
> conflict between these features and styling of pages.
>
> Whilst I did try to have a look at the Pull Request, there seems to be
> a number of Whitespace changes, and non-theme/design changes that make
> it impossible to review the patches.
>
>
> [Victor] I agree that the change has some white spacing issues that
> cause more changes than actually needed.  Maybe Rafik can explore of
> this issue can be fixed with reasonable effort.
>
> Back in January, we agreed as a Team to get 1.3 shipped, and then look
> at UI and email notifications changes. Victor knows I’ve got some work
> on templating emails and pages within Mantis from back then, which I
> said I’d get a PR up for once the Database changes had been done -
> which we agreed as a team to do immediately after 1.3. So until we’ve
> done this, compared the two approaches and decide which route we want
> to go, i’m strongly against this getting merged.
>
>
> [Victor] Here is a case where you also have gone and implemented some
> change without discussion with the team.  At that point I had
> discussed email templating (not UI tempting) and proposed an approach,
> some previews, and you complained and stopped the work.  I’ve asked
> you then to share your code / approach and you haven’t. You just
> killed the momentum and said that you will publish your code by end of
> January then after 1.3.  At that point we also agreed that code that
> is not published doesn’t exist and can’t be used to block other work.
> Rafik has shared a preview of this work with the team 1-2 month ago
> and everyone except you on the core team was excited and gave him UI
> level feedback which he addressed.  So, I expect this code to be
> merged.  The discussion is when exactly it is going to be merged based on
> the 1.3 plan.
>
>
> We agreed to ship 1.3 first so…
>
> When rafiq shared a preview with people 1-2 months ago was when i
> jumped in immediately and said:
> a) we should be discussing this on the dev list.
> b) we need to have a discussion on whether to go with bootstrap.
> c) I believe I said in the same post, we need to have a discussion on
> sidebar on the dev list.
>
> It’s really disappointing actually for me that Rafik has refused for 2
> months to email the development list as it’s not enabled a discussion
> to take place properly. I think I said in the same email that I quite
> liked the design.
>
> Rafik is someone that’s never contributed to Mantis before, has never
> been introduced to us apart from dumping a website re-design and now
> has added a
> 10,000 line commit, which is basically impossible to review.
>
> My purpose for asking him to communicate with the development list 2
> months ago was so that we could have a discussion at an EARLY stage.
>
> As it is, until we get a PR that doesn’t contain X000 lines of
> incorrect whitespace changes in the first commit, it’s actually
> impossible to even attempt to review and look at what has been done.
> And that’s aside from the fact that we need to have a discussion around
> whether to use bootstrap.
>
> Paul
>
>
> On Mon, Aug 18, 2014 at 1:11 AM, Victor Boctor <vboctor <at> gmail.com> wrote:
>
> Paul,
>
> On Aug 17, 2014, at 2:23 AM, P Richards <paul <at> mantisforge.org> wrote:
>
> Alain,
>
> It’s even more painful, when it’s a big change, and changes the
> approach we go in Mantis - hasn’t been discussed on the mailing
> list, and there’s work in progress (well, on hold until after 1.3)
> to allow greater flexibility of theming mantis already. We agreed
> back in January to use the following approach regarding external libraries:
>
>
> [Victor] When Rafik contributed the new website in January, he was
> asked  by several on the team to help out with doing the same for MantisBT
> itself.
> Hence, his contribution.  He has used Bootstrap on the website and
> carried  that forward to MantisBT itself.  Given that he has followed
> through with  this and put this hard word, I would expect that we
> would motivate him for  that and appreciate the good work, rather than just
> being negative.
>
> Now, in terms of Rafiq’s work, I don’t necessarily see it as being a
> major problem (i.e. that we’ve got other work on theming) as I think
> it covers two different things - i.e. Rafiq’s trying to do a visual
> design, anything I’ve looked at tends to be working on allowing
> flexibility/functionality rather then visual effects, so the two probably go
> together in the end.
>
>
> [Victor] Exactly, the visual design is orthogonal to allows users to
> extend MantisBT through plugins.  The believe Rafik’s suggested
> change was focused on revamping the visual design for all features.
> But to scope the change, it doesn’t try to go further than that.
> Once it is in, we can add more flexibility and modernize further.
>
> My main issue is we have plugin’s that add functionality, there’s
> multiple CSS frameworks out there now - Jquery UI, foundation,
> bootstrap etc, that we do need to consider our approach within
> Mantis. I’m pretty confident we could make a similar style using
> query ui, foundation or bootstrap - without two much work, so we need to
> decide what works best long term.
>
>
> [Victor] Bootstrap is a very popular and respected library.  I think
> given that we have this used in the website and the proposed pull
> request now, I don’t see why we would go for something else.
> Bootstrap also provides a good story for responive design allowing
> us to have great native support for phones over time.
>
> The reason I started looking at using Jquery UI in the past (we
> actually moved the Jquery UI javascript from a plugin to the Mantis
> Core) was due to it being used by some plugins, and having the
> visual theme roller - which struck me as something that would allow
> an end user without HTML or CSS experience to brand Mantis into
> their company colours without very much work. Given that Mantis has
> found a whole bunch of uses - for both software development and
> non-software development workflows, this struck me originally as an easy win
> for end users and plugin author’s for styling.
>
>
> [Victor] From memory the dependency we took was mainly on jquery and
> not jquery mobile.  With bootstrap you can also have the equivalent
> of colored themes.
>
> Thanks,
> -Victor
>
>
> --------------------------------------------------------------------
> ----------
>
> _______________________________________________
> mantisbt-dev mailing list
> mantisbt-dev <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
>
>
> ----------------------------------------------------------------------
> -------- _______________________________________________
> mantisbt-dev mailing list
> mantisbt-dev <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
>
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> mantisbt-dev mailing list
> mantisbt-dev <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> mantisbt-dev mailing list
> mantisbt-dev <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
>

--

-- 
http://robert.muntea.nu/

------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
P Richards | 18 Aug 01:21 2014

Graphviz Replacement

Something that I looked at a while back was ideas for replacing ‘graphviz’.

 

Graphviz integration provides relationship graphs for issues and the workflow config page – historically, it’s been problematic on windows (wingraphviz COM object) that used to crash – although that situation has been improved in newer IIS’s as you can run the graphviz.exe I believe now.

 

However, across linux and windows, I’m not sure how many people go to the steps of setting up graphviz to enable this feature.

 

I’ve just pushed a branch <at> https://github.com/grangeway/mantisbt/tree/graphviz containing some mock-up code that attempts to “replace” graphviz with d3.js i.e. http://d3js.org/ . It makes use of another javascript library called dagra-d3

 

This is something that I looked at a long while ago, and I’ve recently updated to newer javascript graph libraries.

 

At the moment, this is a mock up, to get an idea of what it would look like – it would be useful to get some feedback on:

 

a)      Whether it makes sense to replace graphviz with javascript

b)      Whether anyone knows of any better graphing libraries /alternative graphing libraries we could use then d3.js for this

c)       Whether if we were to look at dropping the graphviz stuff, whether the preferred approach would be to move graphviz to a plugin, and then add a javascript alternative – so we get 2 plugins or what

d)      Whether to just leave as is, as graphviz handles complicated graphs better etc

 

At the moment, the code’s not complete in the branch, and hasn’t been in any way tidied up for any review/PR stage – but has been pushed so we can get an idea of what the d3 library could provide.

 

The existing graphviz code provided an image map for links to the bugs – the version in the attached branch doesn’t do that atm. My understanding is that d3 is generating SVG, and that you can add formatting/html etc to these files. However I believe that IE may not support this.

 

My goal here would be to make Mantis easier to setup for end users to get access to all features, i.e. not require hosting company to have graphviz (which may be fairly obscure) installed?

 

The reason I originally looked at this many months ago was due to wingraphviz crashing and graphviz being problematic with php/iis– which isn’t an issue in modern PHP versions for me at least  anymore.

 

I’m fairly easy between any of the options I list in A-D above, equally, if someone has some suggestions for alternatives in B, I don’t mind taking a stab at a similar mock-up branch to this one so we can all get to see what it could look like.

 

Once we’ve agreed any approach between implementing this, implementing something else or doing nothing, depending on the approach, I’ll try and get a PR generated for review.

 

Paul

 

 

 

------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
P Richards | 18 Aug 00:56 2014

FW: [mantisbt] MantisBT Modern UI (#272)

Victor,

 

And regarding your point on the comment, in master, this comment is there, and is indeed invalid – there’s a Pull Request to remove it if you really want it gone <at> https://github.com/mantisbt/mantisbt/pull/273

 

So fixing that isn’t really something related to ‘theming’/design of mantis…

 

Paul

 

From: Victor Boctor [mailto:notifications-9UaJU3cA/F/QT0dZR+AlfA@public.gmane.org]
Sent: 17 August 2014 23:37
To: mantisbt/mantisbt
Cc: grangeway
Subject: Re: [mantisbt] MantisBT Modern UI (#272)

 

In account_prefs_inc.php:

> -     # prefix data with u_

> -     $t_pref = user_pref_get( $p_user_id );

> +    # prefix data with u_

Is this comment out of date?


Reply to this email directly or view it on GitHub.

------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
P Richards | 18 Aug 00:49 2014

FW: [mantisbt] MantisBT Modern UI (#272)

OK – I’ve marked this PR request as closed for now, as we need to have a discussion (as per our agreed policy that victor proposed IIRC) for two reasons:

 

1)      Adding or removing an external library should be discussed first on the developer mailing list.

The following must be considered when reviewing an external library

§  Technical fit

§  License compatibility

§  Recent development activity

§  Community size/support

Changing versions of a library does not need to be discussed, unless there are major changes in the review criteria, e.g. change of license

2)      There’s a lot of what appears to be Whitespace changes in the first commit – I’m thinking/guessing tab/spaces. There’s no point us doing a line-by-line review of each commit if the linefeeds/tabs have been incorrectly changes on the first file

As that approach will get really spammy. The PR needs to be discussed and fixed up first before we can start a review process.

Paul

 

From: Victor Boctor [mailto:notifications-9UaJU3cA/F/QT0dZR+AlfA@public.gmane.org]
Sent: 17 August 2014 23:37
To: mantisbt/mantisbt
Cc: grangeway
Subject: Re: [mantisbt] MantisBT Modern UI (#272)

 

In account_prefs_inc.php:

> -     # prefix data with u_

> -     $t_pref = user_pref_get( $p_user_id );

> +    # prefix data with u_

Is this comment out of date?


Reply to this email directly or view it on GitHub.

------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
Paul Richards | 17 Aug 12:19 2014
Picon

Wiki: how to get involved page

I think we need to look at updating our wiki steps on ‘how to get involved’ to make them clearer, so that people post their ideas to the mailing list *before* starting work.

We’ve got comments in the http://www.mantisbt.org/wiki/doku.php/mantisbt:development_process page like “For non-trivial work items: If you need immediate feedback from the core developers, email the dev mailing list with a pointer to your issue. The feedback should be captured directly in the issue rather than the mailing list.” and “Adding or removing an external library should be discussed first on the developer mailing list.”, but they are hidden away in the text of the page.

If you look back over time, we’ve had a number of people come along with different ‘patches’ to Mantis to enhance something that have never been merged as they’ve not been discussed/thought out first. Equally, we’ve got stuff that’s been done as a ‘plugin’ that I’d probably argue should have been a patch to the core project.

I think we need to create a clear wiki page that tries to outline when to create a plugin, when to create a patch, and how to discuss a new feature/work item with the list – so it can be decided whether it should go into a plugin, or be done as part of the core.

If you take the design’s we’ve had over the years – we have had patches to enhance the design and users releasing their own themes/’forks’ - some of which are pretty good, others pretty poor – so the quality varies. In all cases, they’ve been done without people going to the development list first, and nothing has ever been merged. And in some cases, never hitting the bug tracker, forums or mailing list.

For example Mantis Themes - On github, https://github.com/bueltge/MantisBT-Colorized / https://github.com/J-Rabe/MantisBT-Colorized and https://github.com/TimPietrusky/mantisbt-is-a-rockstar, and you can find similar patches on bugs on the bug tracker.<sidenote>I actually prefer some of these older theme’s over the current ‘modern’ theme (and prefer some parts of the new modern theme over the older attempts)</sidenote>

I suggest we create a single page for new developers with the following content points expanded on:

·         Thank people for their interest in mantis

·         Explain that mantis has both core code + plugins

·         State if making a plugin to have a think whether what you are implementing is something generic (that might make sense to put in core) or something specific that should likely be a plugin. If plugin, code it. If unsure, email dev list for input.

·         If implementing a trivial bug fix for core code, it’s probably just worth implementing.

·         If planning a new feature, contact the dev list so advice on historical information (relating to that feature) can be provided, and equally, we can decide if it’s a feature that would be better implemented by adding some plugin hooks etc

·         If planning to use a new library, email dev list early on so a decision can be made.

·         Provide a single patch at the end that can be reviewed easily i.e. PR should really only have a single DIFF in it.

Let me expand on the last point briefly, when it comes down to reviewing a PR, commits should really be collapsed in a sensible way. As I think this probably needs some thought.

When we used to have SVN/GIT before we started doing PR’s, I’d always review a patch on its own merit. So for an example, if I look at https://github.com/mantisbt/mantisbt/pull/235/commits - that’s started off as 1 Patch that needed a review, and now contains 3 follow up patches. In that case all the patches are unique.

I’ve been tending to rebase existing patches when updating patches (in some cases), so that I can run git format-patch on the patch, and have a proper look through at what I’m changing as a final step (to ensure there’s no ‘vardump’s etc left in).

And here comes before a benefit and negative with git on this – when reviewing changes on their web interface, a single diff on a PR is the ideal starting point. In theory, that single diff should be the complete bugfix/feature (that should have already been discussed on the mailing list if needed), and therefore the branch feeding it in can be collapsed into a single commit before making the initial Pull Request. If the feature is complex enough to warrant it, it may be worth keeping around a copy of the branch before the commits get merged in case it aids the reviewer to be able to look back and see how the commit got to its final point, but we don’t need to be reviewing the history on the PR.

It also allows a quick check to ensure that no whitespace changes etc have crept in to the commit.

I’d then argue that depending on the complexity of the patch, a decision should be made whether to do any follow up fixes as amendment to the original patch or as a separate commit – and depending on how many people have reviewed it. So for example, a missing semi-colon spotted by the first reviewer should probably be an amendment. A new config option that’s been added because people figured it would be necessary should probably be a separate commit. This allows people to review the new change only rather than going over old work.

Thoughts?

Paul

 

 

------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
Rafik Robeal | 17 Aug 05:00 2014
Picon

Introducing MantisBT Modern UI

Hello Everyone,

For people who don't know me, my name is Rafik Robeal and I am working on building MantisHub.

I've experienced MantisBT first hand working with many users who love and admire MantisBT. I can simply say that MantisBT has proven itself to be powerful feature-rich bug tracker. One thing remains though, its UI is behind the times. To establish MantisBT as the best modern open source bug tracker for many years to come, a modern UI is needed.

Over the past few months I've been working on building a modern UI for MantisBT. The main drivers for this effort are: (1) a request from MantisBT core team back in January when I delivered the mantisbt.org modern website that you see live today, (2) a loud and clear feedback from MantisBT/MantisHub customers who publicly and privately spoke against the current UI , (3) a personal desire to contribute to MantisBT open source effort with a lasting positive impact on both end users satisfaction and product adoption.

Enough introduction ... what is this all about? In four words: freaking awesome user interface

Before reading any further, fire up your browser and go to: http://mantishub.com/modern/index.php
You are accessing the preview version on MantisBT Modern UI. Use username: reporter & password: demo

I've already submitted the pull request for this work which includes:
  • Integrate with Bootstrap v3.2 framework as the foundation for all Mantis UI elements.
  • Use Fontawesome for vector based icons.
  • Redesign Layout & Navigation
  • Update all pages with the new style, look and feel
  • Improved experience on mobile devices
The goal is to share the code and get feedback on it. We need to decide on when this code can be merged specially in relation to the 1.3 release. Decide whether it is 1.3 or 2.0 and figuring out when the 1.3 branch will be created so that the pull request can eventually be integrated.

I'd love to hear from everyone and address any potential issues you see with the new design or code.

Thanks,
Rafik Robeal
------------------------------------------------------------------------------
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev
Damien Regad | 11 Aug 17:28 2014

Github / Pull request spam

Paul,

Can we please have a SINGLE branch/pull request for a feature/fix, with
several commits in it, instead of a series of pull requests each having
a single commit ?

I'm referring to:
- "Following merge of #216 - update calls to db_get_table()"
  ==> 14 pull requests (not counting #260 which is more of the same)
- "Remove use of AS keyword from sql query"
  ==> 3 pull requests

This basically multiplies the effort for you to generate the individual
branches, and later on merge them, as well as for the reviewers: 17
distinct reviews and "+1" posts - assuming there are no issues - instead
of 2 in this case.

Please reword also the commit messages:

- it is useless to have each commit begin with "Following merge of #216"
(no added value in in terms of information on what the commit is about,
and "#216" will be interpreted as a Mantis bug ID when in fact it's a
Github pull request)
- please indicate *which table is being updated* - just think for a
second at how the git log will look like...

It would be much better instead, to have each commit's summary be
"update db_get_table() calls for table XXX"

Also, can we please have meaningful descriptions for the pull requests?
"Backport <SHA>" is useless really, particularly when the SHA is from
your own fork and does not exist in the mantisbt repository.

Finally, how can something be a "backport" when master is the latest
head (and has never been released) ?

Kindly fix the above, I'll review things then.

Thanks for your understanding.

---
This email is free from viruses and malware because avast! Antivirus protection is active.
http://www.avast.com

------------------------------------------------------------------------------
Roland Becker | 10 Aug 15:15 2014
Picon

Unnecessary Commits

Sorry, for the noise
https://github.com/mantisbt/mantisbt/commit/ad94ddb293deeddd46a8c332dcaf525855284f76
What are the right steps to have no spaghetti history and to have no unnecessary
merge commits?

What I did
Created a clone of https://github.com/mantisbt/mantisbt

git checkout -b atrol-fix13-test_langs master
git pull git@...:atrol/mantisbt.git fix13-test_langs
git checkout master
git merge --no-ff atrol-fix13-test_langs
git branch -d atrol-fix13-test_langs
git push origin master

Roland

------------------------------------------------------------------------------
且行且歌 | 4 Aug 14:09 2014

Hosting a new plugin request



Hi, guys

    I am developing a new plugin for mantis right now and I want to add it to the list of mantis plugins. Can you help me?

    Here is the basic info of my plugin:
    repository: https://github.com/aaronsh/MantisAcra.git
    plugin's name: MantisAcra
    plugin's description: The plugin collect ACRA crash reports from android devices and save the reports as issues in Mantis. Later, user can review the reports with all ACRA features, stack trace, logcat and so on. Note: ACRA is a great crash report tool for android.
    plugin status: It runs, but need more test and remove some duplicated files.
    owner list:

    Thanks and best regards!

    Sam Liu

------------------------------------------------------------------------------
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls. 
Build a bridge from your legacy apps to the future.
http://pubads.g.doubleclick.net/gampad/clk?id=153845071&iu=/4140/ostg.clktrk
_______________________________________________
mantisbt-dev mailing list
mantisbt-dev@...
https://lists.sourceforge.net/lists/listinfo/mantisbt-dev

Gmane