Status of main branch, and debugging on Windows

Wed Feb 12 11:22:40 EST 2014

On Wed, Feb 12, 2014 at 3:13 AM, Tomasz Wasilczyk <twasilczyk at pidgin.im>wrote:

> W dniu 12.02.2014 06:19, Jean-Sébastien Guay pisze:
>
>  Using the instructions here:
>> https://developer.pidgin.im/wiki/BuildingWinPidgin/3.0.0
>> and with some small modifications to get everything building, I was able
>> to make an installer and install my own build for Windows. And it runs :-)
>> (aside: as-is the main branch would not build on Windows currently, as
>> pidgin_start() needs to get an extra HINSTANCE parameter if _WIN32 is
>> defined... don't you have an auto-build system that tests that new
>> checkins build on all platforms, at a minimum?)
>> (aside 2: where should I send my patch which fixes the Windows build for
>> discussion? Create a ticket or this mailing list?)
>>
>
> I haven't built win32 version recently, so it may be broken. You may send
> patch (and these "small modifications" you mentioned) here, or directly to
> me, if you wish. We have a CI system [1] deployed just a week ago, but we
> don't test win32 here, as cross-compilation doesn't work for default
> branch. Making it possible is on my TODO list.
>
>
>  However some things don't seem to work, in particular related to smileys
>> (which is a real bummer as that's what I want to work on)
>>
>> For example:
>> (...)
>>
>
> All these smiley-related issues are already known (for me, at least) and
> also are present on my TODO list. Except the MSN pidgin-vs-official client
> difference you mentioned before. You may investigate it on 2.x.y branch, if
> you wish (as I said, 3.0.0 have almost broken smileys).

Yep, smileys are broken because we changed our HTML render from IMHtml to
Webkit in 3.0.0. So there sould be some work to get them working as
expected.

On the MSN side, IIRC, the problem is the change from MSNP15 to MSN18 which
changed all the mechanism used to send data. Here, we have a dilema, invest
time on MSN knowing it is dead and it is just a matter of time for MS to
shut down the servers or fix the issues and hope they don't kill the
servers soon?

You can guess I have been on the "it is dead, it doesn't worth a minute"
side. =P

Anyway, it would be nice to get that fixed but... the time... and MS, and
MSNP is really really undocumented and it is a pain.

 I haven't tested more than that, but I'm afraid more things could be
> broken in the main branch which would make using it harder even if I
> *did* fix the bug I want to fix :-)
>

Indeed, the main branch have a bunch of things broken. But some of us uses
> it on a daily basis, so it should remain usable. Especially the gtk2 builds.
>
>
>
>  I could make a build of the 2.x (I guess 2.10 at this point) branch, but
>> honestly, the prepare-workspace.sh script made things much simpler than
>> to have to download and unpack a bunch of dependencies manually so I
>> thought I'd try 3.0.0-devel first. Kudos to Tomek Wasilczyk who made
>> that script, it's a great way to get going quickly :-) If the general
>> concensus is that it would be easier to use the 2.10 branch and work off
>> that then I'll go through the trouble of setting up the dependencies
>> manually... Or maybe I'll adapt Tomek's script to that branch?
>>
>
> Thanks :). In fact, this script is thought as a step toward making full
> cross compilation working. I'm not sure, if we leave a possibility to build
> directly on win32, when it will be possible.
>
> And currently, 2.x.y branch is only for fixing bugs - all new features
> goes to 3.0.0. But if something doesn't work on 2.x.y as expected, don't
> hesitate to fix it!
>
>
>
>  And last question, on Windows, how would I get started with debugging my
>> own build (with gdb I guess)? I tried using the cygwin-installed gdb,
>> and passing one -s argument for each .dbgsym file that gets installed
>> when you check the "Debugging symbols" checkbox in the installer, but
>> gdb still says when it loads pidgin.exe that it finds no debugging
>> symbols.
>>
>
> Debugging on Windows is broken at the moment, due to ASLR usage. I'm
> afraid, the only debugging method is "printf" at the moment.
>
>
> [1] https://www.guifications.org/jenkins/view/Pidgin/
>
>
> _______________________________________________
> Devel mailing list
> Devel at pidgin.im
> https://pidgin.im/cgi-bin/mailman/listinfo/devel
>

-- 
Masca

A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pidgin.im/pipermail/devel/attachments/20140212/151df0dd/attachment-0001.html>