diff options
author | Andrea Diamantini <adjam7@gmail.com> | 2011-01-07 18:53:20 +0100 |
---|---|---|
committer | Andrea Diamantini <adjam7@gmail.com> | 2011-01-07 18:53:20 +0100 |
commit | 39413862a2d2355d3765d986073b29c1c2a1b7d7 (patch) | |
tree | 5be5777239fc6031b0fbc4f92ff668c31ca68f08 /src/webpage.h | |
parent | This GCI task adds a combobox with different settings to choice when hovering... (diff) | |
download | rekonq-39413862a2d2355d3765d986073b29c1c2a1b7d7.tar.xz |
Put ioslaves on hold.
This patch couples those merged in KIO & KDEWebKit in SC 4.6.
I have to say it is probably NOT perfect and ready to merge, but I'd like to
because it is anyway a nice step over actual behavior.
It basically holds and (re)publish slave for apps needing it, letting us fixing
a long standing bug against the infamous TWO connections to do ONE thing (download, move, etc...)
It also gave me the possibility to and/or improve these things:
- better handling filenames, studying content disposition header for "attachment" and "inline" values
- better history handling, when working with kparts (also if this part needs a proper fix, IMHO reimplementing QWebHistory class...)
- a (proper?) fix for POST operations returning content. I admit I am dubious with this part and I need more testing. The nice thing
is that this patch "circumscribes" this problem, letting it easy eventually fix it again
I also copied from kdewebkit a check for exec text files mimetypes (converted to text/plain) and cleaned-up some kDebugs
(and as usual, added a lot..)
Last, sorry for the big commit message and the big/not so big patch here. It is a special case.
commit 29f3021d94e26b39f3e8172a980a6cafcebe5095
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Wed Jan 5 12:17:45 2011 +0100
cleanup comments
commit fc88a37322810fd25c5bfcd16880fc617e7de9f5
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Wed Jan 5 12:06:55 2011 +0100
content-disposition: inline
commit 7c2789a1733e1089b9c66cd6c553185792ec3a72
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Tue Jan 4 12:04:03 2011 +0100
Fix history handling.
For a proper fix here, I think we need to reimplement the tab history, otherwise we cannot manage
tab history browsing on the so called "rekonq" pages...
commit 892c4d8dc78fda43f67566a3e359bd9805ee3282
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Tue Jan 4 11:57:29 2011 +0100
Two cases here:
1) POST operations returning content
2) right file names
commit 3b1d5873b7eb743406f3b8c5abb79ce6728d0103
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Tue Jan 4 03:24:21 2011 +0100
Stupid kDebugs...
commit 73d038a757185816f40795d4dd4a68e11659c971
Author: Andrea Diamantini <adjam7@gmail.com>
Date: Tue Jan 4 02:32:45 2011 +0100
It needs testing from people (just) having last KDE SC 4.6 code on...
Diffstat (limited to 'src/webpage.h')
-rw-r--r-- | src/webpage.h | 9 |
1 files changed, 8 insertions, 1 deletions
diff --git a/src/webpage.h b/src/webpage.h index 41ddd877..7633b26d 100644 --- a/src/webpage.h +++ b/src/webpage.h @@ -59,6 +59,7 @@ public: inline void setIsOnRekonqPage(bool b) { _isOnRekonqPage = b; }; inline KUrl loadingUrl() { return _loadingUrl; }; + inline QString suggestedFileName() { return _suggestedFileName; }; public Q_SLOTS: void downloadAllContentsWithKGet(QPoint); @@ -81,15 +82,21 @@ private Q_SLOTS: void showSSLInfo(QPoint); void updateImage(bool ok); + void copyToTempFileResult(KJob*); + +private: void downloadReply(const QNetworkReply *reply, const QString &suggestedFileName = QString()); private: QString errorPage(QNetworkReply *reply); - QUrl _loadingUrl; + KUrl _loadingUrl; ProtocolHandler _protHandler; WebSslInfo _sslInfo; + QString _mimeType; + QString _suggestedFileName; + bool _networkAnalyzer; bool _isOnRekonqPage; }; |