Version: 1.0.0 (using KDE 3.3.1, compiled sources) Compiler: gcc version 3.3.4 OS: Linux (i686) release 2.6.9 It seems to me it would be fairly simple to guess whether the file in question is a indirect file or not, ie. is it a smallish ASCII file that contains a valid file/path name.
On Tuesday 30 November 2004 22:40, D.V.Wiebe wrote: > It seems to me it would be fairly simple to guess whether the file in > question is a indirect file or not, ie. is it a smallish ASCII file that > contains a valid file/path name. This is by design presently.
Created attachment 8500 [details] Proposed fix... Get rid of check for .cur (as it checks other aspects unique to an indirect file as well)
Created attachment 8501 [details] Perhaps a better fix Perhaps this is a better way...
On Tuesday 30 November 2004 23:52, Matthew Truch wrote: > Perhaps this is a better way... that should be endsWith(), not !endsWith() I think. Or switch the values for ret. Anyway, I guess it can't hurt. Anyone object to me applying this style of change? I might fudge with the numbers a bit more based on what other data sources (especially ASCII) return, but in principle I don't think it hurts anything.
The approach seems fine... On December 1, 2004 12:19 am, George Staikos wrote: > ------- You are receiving this mail because: ------- > You are the assignee for the bug, or are watching the assignee. > > http://bugs.kde.org/show_bug.cgi?id=94197 > > > > > ------- Additional Comments From staikos kde org 2004-12-01 06:19 ------- > > On Tuesday 30 November 2004 23:52, Matthew Truch wrote: > > Perhaps this is a better way... > > that should be endsWith(), not !endsWith() I think. Or switch the > values for ret. Anyway, I guess it can't hurt. Anyone object to me > applying this style of change? I might fudge with the numbers a bit more > based on what other data sources (especially ASCII) return, but in > principle I don't think it hurts anything. > _______________________________________________ > Kst mailing list > Kst@kde.org > https://mail.kde.org/mailman/listinfo/kst
> that should be endsWith(), not !endsWith() I think. Or switch the values > for ret. Boy, do I feel sheepish. Correct.
CVS commit by staikos: indirect files no longer have to end in .cur, adjust ascii to return a weaker match percentage if it also looks like an indirect file. FEATURE: 94197 M +1 -1 ascii/ascii.cpp 1.28 M +5 -3 indirect/indirect.cpp 1.12 --- kdeextragear-2/kst/kst/datasources/ascii/ascii.cpp #1.27:1.28 @@ -364,5 +364,5 @@ int understands_ascii(KConfig *cfg, cons } else if (QRegExp("^[\\s]*(([Nn][Aa][Nn]|(\\-\\+)?[Ii][Nn][Ff]|[0-9\\+\\-\\.eE]+)[\\s]*)+").exactMatch(s)) { // a number - this may be an ascii file - assume that it is - return 75; + return QFile::exists(s.stripWhiteSpace()) ? 49 : 75; } else { return 20; --- kdeextragear-2/kst/kst/datasources/indirect/indirect.cpp #1.11:1.12 @@ -131,6 +131,8 @@ QStringList provides_indirect() { int understands_indirect(KConfig*, const QString& filename) { - if (!filename.endsWith(".cur")) { // Do we really have to do this? - return 0; + int percent = 50; + + if (filename.endsWith(".cur")) { // Do we really have to do this? + percent = 100; } @@ -145,5 +147,5 @@ int understands_indirect(KConfig*, const } - return QFile::exists(ifn.stripWhiteSpace()) ? 100 : 0; + return QFile::exists(ifn.stripWhiteSpace()) ? percent : 0; }