I once tried to send something to the developer list and it never appeared there
presumably because you have to be subscribed for stuff to be accepted. I would
rather not get email every time someone sends something to the list when I can
instead look at the mailing list archive. Also the list is sometimes spammed which
is something I definitely do not want. Since I keep my email address fairly private,
I currently get ZERO spam and want to keep it that way.
Maybe MonkeeSage will post those to the list then. I'm not trying to force people to use the list but I know most of the developers don't follow all of the forum traffic and I can't say I see everything myself. I'll try and keep up with these changes. If you file them to the individual bugs, those will be seen by the developers.
Go ahead and e-mail me directly. I'll double-check those against the code and then pass along diffs so that we can check those into CVS. I'll probably go ahead and pass those along to Mozilla too for their MFCEmbed app.
Andrew, please notice the crash bug I reported in the Mozi-Meleon 1.3b-thread with the changes to save "web page, complete". Other than that the save-changes are really nice (and worth the crash ;-) ).
You're saying the crash only occurs when using web page complete to save LOCAL
files ?? Does this happen with the released K-Meleon ? I would like to know if it
happens with the new code using Mozilla 1.2.1, I don't trust Mozilla 1.3b.
yes, *only* *local* pages, only with "web page, complete" and only with
MonkeeSages executable (installed with moz 1.3b exactly like described
in install.txt [1]).
The only thing that is saved is an empty directory _files. Repro-steps can
be found in the other thread. BTW: KM crashes only if I start (or go on)
browsing *internet* sites afterwards, if I browse local files, KM doesn't
crash. Can anyone repro the crash? MonkeeSage, does your KM crash?
I don't have a 1.2.1-based K-M here, but the service pack release (using
moz 1.2b-code) is working fine. Page is saved, images are there, no crash
afterwards, everything fine.
BTW: OS is Win XP Pro SP1.
[1] Well, okay, "nsIQTScriptablePlugin.xpt" is missing here, but that
shouldn't make a difference, should it?
Maybe MonkeeSage will make a Mozilla 1.2.1 build with the updated code, it could
be something different about the Mozilla 1.3 API that's causing the problem.
Out of curiosity, what made you try save web page complete on local files ?
I'm building 1.2.1 at this moment. There is also a cookie issue with my build, and I'm curious if it is a 1.3 bug (I am betting it is). No cookies.txt is written with the 1.3 build, it reads them fine, but it is not writting them...I am thinking it is something in the 1.3 cookies.dll / .xpt in components.
As for the save complete, it works fine for me on local files with 1.3 and no crashes, so I'm not sure what the deal is. I thought I had changed the baracketing on this section slightly:
Careful, that final } matches if(cf.DoModal() == IDOK) { which is far above.
If there's a problem then perhaps it in the 1.3 SaveDocument function call
considering SaveURI is different between 1.2 and 1.3.
If you would, so I don't have to go back through the forums and get the pieces, will you please send the document with your code revisions to: my email ?
A big fat BTW needs to be attached to pulling K-M off CVS using cygwin
It sets the .dsw and .dsp files to UNIX file-type, instead of DOS, so you have to set them back manually (e.g., with UltraEdit32), before VC++ will recognize them, otherwise it shows no projects in the main .dsw!!
set MOZCONFIG=C:\fully\qualified\path\to\mozconfig
Make sure to export / set this environment variable prior to building (like in your env.sh script), or Mozilla won't know about your configuration and will build using the default.
Any ideas how to handle a reload in an individual frame? I culled through the MFCEmbed code as well as the Mozilla code and I didn't see anything for us to hook into that would accomplish that. In Mozilla, it's done with Javascript which isn't really what I was looking for.
Mabye to "refresh" a frame, you can grabe the frameURI, and add a new frame exactly where the old frame was in the layout, and simply pass the original frameURI as the URI to the new frame?
> Maybe MonkeeSage will make a Mozilla 1.2.1 build with the updated code,
Using the version from this afternoon.
> it could be something different about the Mozilla 1.3 API that's causing the
> problem.
Yeah, with 1.2.1 it works like a charm. :-)
> Out of curiosity, what made you try save web page complete on local files ?
I'm always trying to save the same page with *every* km build (I'm using km since
christophe released .1 back then), because there were *often* bugs in the save as-
function... It's quite a simple page, that lists a lot of searchengines with their
logos, that I linked from my (other) local startpage.
Well, yeah, I can save *one* page and that's it. Then I have to completely close km,
open it again and can save another page and so on. So the fix / workaround (dialog-
box...) for the other bug is causing trouble as noted in another thread before.
That's "Version 0.7.1 Build 1 Compiled Tue Feb 27 09:02:53 2003" on Moz 1.2.1.
Ah, well, just got the latest version and with that saving pages is broken again...
Guess I'll have to wait for MonkeeSage to make a new one...
Don't know if it makes much sense to report that now, but anyhow, there was
another bug with the save as-function in the older version:
The file extension won't be added if you choose to save a page that has a different
extension than your chosen save as-option.
Sounds difficult? ;-) Example: Save a php-page, let's call it "filename.php" as
"Web Page, HTML only (*.HTM;*.HTML)". Then you will get "filename.php"
instead of (what one would be exptecting) filename.htm or filename.php.htm.
Another one: If I save home.htm as "Text File (*.TXT)" the saved file will be
"home.htm".
It works when the shown extension is manually removed or changed to an
*unknown* file type.
And the cookie confirmation dialog is broken, too (checkbox doesn't work)...
So, I'm back to the 1.2b-code with KM.7SP1 for now. :-)