Re: [Qucs-devel] Closed bugs

[Frans S. <fra...@gm...>]

Hello Guilherme,

I think rolling back will make things worse now, if you have found these 
bugs it may be good to fix them in stead.
I think it is also good to have some people with mac's test the package 
before we release 0.0.17.
We can sure wait one or two weeks, as Mike Brinson also wants to put his 
BSIM4 model in the new release.
I have filed 3 new bugs with the exact text from your points here.
What do you think, will it take a lot of time to fix those bugs? I 
assigned one to myself, but we can see who is fixing which one.
Thanks for addressing those.

Frans


On 7-6-2013 10:28, Guilherme Brondani Torri wrote:
> On 06/06/13 22:57, Frans Schreuder wrote:
>> Hello everyone,
> Hi!
>> I have closed most of the bugs. Some of them dated back from 2006 and
>> some of them clearly had been fixed.
>> If there are any bugs that need to be reopened, please do so.
> Nice cleanup! Thanks!
>> I guess all other bugs have been fixed by now!
>> Anything else blocking the release?
>>
>
> Well I believe I introduced/found at least 3 bugs into master. That is 
> because I started pushing to master instead of a branch. My bad! I was 
> used with pull-requests... Now I'm using the branch 
> post-release-0.0.17 (created by Clemens) and checking things more 
> thoroughly before landing changes into master.
>
> By the way, I'm putting together a qucs_testset_prj with a variety of 
> small examples very helpful in 'exercising' or 'covering' and 
> debugging the code execution. I'll share it later.
>
> Now back to the issues, we can either roll those commits back or merge 
> the code on post-release and fix the issues for the release.
>
> Let me mention that, as it is now, most of the qucs/qucs/dialogs and 
> the subtools (except qucs-help and qucs-filter-v2) are mostly free of 
> Qt3Support classes. I say mostly because once you remove the flag 
> -DQT3_SUPPORT other mismatches in methods' signatures will show up.
>
> The whole dockable dialog containing Projects/Content/Components was 
> not yet ported (I couldn't get my head around its implementation either).
>
> 1)
> The dialog/librarydialog still lacks a field to insert the subcircuit 
> description. I'm planing to use a QStacks to avoid creating another 
> dialog.
>
> 2)
> The dialog/qucsettingsdialog is incomplete on the table of file types. 
> I started placing a QTableView, but perhaps a QListView is better.
>
> 3)
> I was testing master for digital simulation and it is now crashing with:
> ASSERT: "uint(i) < uint(size())" in file 
> /usr/local/Cellar/qt/4.8.4/include/QtCore/qstring.h, line 699
>
> The crash report point to these changes :
> https://sourceforge.net/p/qucs/git/ci/6db21755d115c770b3aba4e61102fc6fd9085ccf/tree/qucs/qucs/schematic_file.cpp?diff=fd43f78cb5965a622aa1975af0a96dc0a0e367a7
>
> It is about the section symbol (§), not paragraph as I wrote on the 
> comments.
>
> I shall mention that the schematic_file.cpp was encoded as ISO-8859-1 
> which was choking Qt Designer because it expected a Utf-8 file. I 
> converted it with iconv. Now it is crashing during a digital 
> simulation, the simple example from the 'Getting started' help (yes I 
> have freehdl installed).
> For that matter, there is a variety of file encoding across the code 
> base... Shall we fix that?
> There were similar issues with the encoding of the degree symbol (°), 
> but this only goes into the GUI, they are not involved on the handling 
> of netlists or vhdl code...
>
> ----
> So, what do you think? Go back or go forward?
>
> Sorry for the mess!
>
> Last but not least, if you can give me another week I can include 
> freehdl and icarus on the Mac OSX package, so it would be more or less 
> like the Windows installer, right?
>
> Best regards,
> Guiherme