bibledit-development
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [be] Bibledit for Windows patched source files


From: Dan Dennison
Subject: Re: [be] Bibledit for Windows patched source files
Date: Tue, 3 Mar 2009 06:25:09 -0800

On Mar 3, 2009, at 12:05 AM, Teus Benschop wrote:
It now uses the M4 macro that was provided by the curl guys themselves, and I am not such a M4 / autotools guru so as to be able to do a lot of hacking myself. Probabably the Windows guys of curl can be of help here
(if there's a a special Windows maintainer at all), or else looking at
other windows ports that use libcurl, how these configure their package?

okay. I'll keep looking for how to fix this. In the meantime, it's an annoyance.

Also, is it possible to rename your windows.h ? Because the src/ Makefile has '-I.' appended to the include flags, #include <windows.h> hits the Bibledit one instead of the system one. This makes anything that uses Windows API barf on compilation. Can you either remove the '- I.' or rename windows.h/cpp ?

Must have been a lot of work to get it to a good state!

Debugging SIGSEGVs in gdb :) The sheer joy of it all ;) The fun part now is that you can use Dr. Watson logs on Windows to grab the function in which you crashed :)

Also windows outpost launches in the CMD script instead of from the app.

Ah, so the application sees it running already, and does not need to
launch it, or it assumes it running.

Right now I'm telling it to use 127.0.0.1 as a 'remote' outpost. Quirky, but works. The UI should really get fixed.

That's fine, but why is that? Is Windows refusing the ";"?
I'll try to remember to do that for next time, and if I forget pls.
remind me. I suppose that you've patched them all?

src/scripts.cpp looks complicated and hasn't been fixed yet.

What would really be helpful is to add command line helper functions to tiny_utilities.cpp. Then, instead of building the command lines as strings in every source file, you represent the command line using helper functions, then call it using another helper function. See git- exec.cpp. Otherwise, we're looking at a TON of #ifdefs. It will be ugly, and is already.

Also, git-exec.cpp was done before I realized your use of GwSpawn. Hence, I was calling the Win32 CreateProcess to avoid the blasted popup windows. If we could just ban the use of system(), and use only GwSpawn, both via a tiny_utilities.cpp helper function, that would simplify the porting of these command lines greatly!

As it stands, every use of system() pops a black command prompt window briefly. On slow systems, it might be an annoyance.

Ok, it probably is going to make sense, but I'll be applying the patches
one by one, and test each change under Linux, then if there's any
explosion, I'll call you. You'll see the commit logs going out in due
course.

I was a bit liberal in my inclusion of source files, so some might have no diffs. Let me know the list of identical ones, if any.

Finally, did you include a fix for the 4-9% CPU usage due to the string comparisons? I noticed the Check menu, the httpd search fix, the IPC timeout change, and the libxml2 initialization changes.

dan :)




reply via email to

[Prev in Thread] Current Thread [Next in Thread]