>"apps that should work in TDE and will reflect poorly on TDE
>if included broken"
Based upon the response rate you've seen in this list, my guess is
nobody gives a hoot about Trinity's reputation.
Darrell
>I understand your frustration Darrell, and no you are not being
>mean. You are
>apparently taking the information I am posting in a way other than
>it is intended to be received.
Yesterday I spent a couple of hours trying to resolve bug report
1842. The Help button code is in place but the button does not
appear.
Not to mention that the default size of the list box window clips
half the text to be unreadable.
After several attempts to patch with my awful C++ skills, I walked
away in frustration. Then I asked myself: Does anybody use this low
quality garbage? Who actually writes such crappy software? Is there
a reason why such apps never were pulled into the main KDE trunk?
Why are we supporting junk? Why do people use such software and
then pretend the app is not half broken?
This is what I mean by "chicken shit" bugs. These kinds of bugs
never should exist in the first place.
Today I asked myself, Why am I spending time on junk? Stop building
the crappy stuff and the problem goes away. Nobody else seems to
care that this grabage is broken, why should I?
The real problem is my standards are too high. I need to learn to
accept that three round wheels and one wheel shaped like an octagon
is "good enough."
Darrell
>What if you want an actual database backend for kmymoney?
>
>In kmymoney's case "works for me" is a qualified "works for me" at
>best. The
>documentation blows up the build as well unless you specifically
>disable the pdf
>documents or install a 3rd party perl package of html2ps:
Does anybody actually use kmymoney? Somebody with knowledge who can
attest whether these additional build options hamper or cripple
usage? Is a database backend critical for usage or does the app use
an alternate method for storing data?
I have no idea because I don't use the app.
Is the emperor wearing clothes?
Darrell
>I tried KFtpgrappeb on a debian mirror, but after establishing a
>connection, it "stops" on the "fetching file lise".
>There is definitely something wrong somewhere.
I am going to be frank here and will sound harsh and mean. We are
eager to add older kde3 apps to the repo but part of me thinks we
do this with a bean-counter mindset --- to impress people how many
apps we have. From my own usage and observation I find many of
those older apps useless fluff.
We spend zero time on quality control and testing. Browse the bug
tracker --- filled with paper cut issues. Or, to be blunt, what I
call "chicken shit" issues. The bug reports all just sit there with
no attention.
David, file Yet Another Bug Report. I don't know what else to do.
Darrell
All,
Picking through the remaining files to build, I ran across tde-construct that
apparently is an autobuild meta-package for TDE that advertises the ability to
automatically download the TDE sources and build TDE for you as easy as:
"cd meta/kde;make install"
By default "Konstruct" installs to ~/trinity.5.7/ which means you don't have
to possess root privileges or risk to damage your system or affect another KDE.
Man, if this thing works -- we've been doing it the hard-way all along ;-) But
seriously, has anyone tested any of it? It seems to contain all of the base
install and many of the applications.
I suspect none of the links point to correct sources and it would take a bit
of CPR to get something to actually build, but the idea is good. Anybody ever
built/tested it?
--
David C. Rankin, J.D.,P.E.
All,
I've testing both ftp clients in applications (kftpgrabber and kasablanca). Of
the two, kasablanca will connect, but does not display any remote contents in
the file areas. I can issue commands e.g. PASV, help, dir, and the log output
shows the remote server sees the command and responds by changing directories
remotely and issues the CWD response showing successful directory change, but no
file listing is ever available. (tried both with/without [x] correct pasv) Here
is the sample log window output:
220-Gesellschaft fuer wissenschaftliche Datenverarbeitung mbH Goettingen
220-====================================================================
220-This is a Linux PC (Dell PE-2650, 2 CPUs P4/2800, 12 GB RAM)
220-running SuSE-Linux-8.2 with SuSE kernel 2.4.20-64GB-SMP.
220-/pub is ext3, a 2 TB Partition within a Transtec T6100F16R1
220-IDE/FC Raid array with 15+1 SATA disks Hitachi 240 GB Raid5
220-
220-For questions and other contacts to the ftp-admin, mailto:emoenke@gwdg.de
220-All transfers are logged. The local time is Tue Feb 18 08:25:44 2014.
220-
220-Limits: Domains gwdg.de, uni-goettingen.de, mpg.de: 200 sessions
220- Other domains: Mon - Fri 8:00 - 18:00: 1000 sessions
220- other times: 2000 sessions
220-Your domain limit: 1000 active sessions: 14
220-
220-Login as "ftp" or "anonymous" and use your email address as password.
220-
220 ftp.gwdg.de FTP server (Version wu-2.4.2-academ[BETA-18-em](1) Fri Apr 4
23:14:50 MEST 2003) ready.
USER anonymous
331 Guest login ok, send your complete e-mail address as password.
PASS *
230-Hello User at mail.3111skyline.com,
230-we have 15 users (max 1000) logged in in your class at the moment.
230-Local time is: Tue Feb 18 08:25:44 2014
230-All transfers are logged. If you don't like this, disconnect now.
230-
230-tar-on-the-fly and gzip-on-the-fly are implemented; to get a whole
230-directory "foo", "get foo.tar" or "get foo.tar.gz" may be used.
230-Please use gzip-on-the-fly only if you need it; most files already
230-are compressed, and I will kill your processes if you waste my
230-ressources.
230-
230-The command "site exec locate pattern" will create a list of all
230-path names containing "pattern".
230-
230 Guest login ok, access restrictions apply.
CWD /pub
250-====================================================================
250-Gesellschaft fuer wissenschaftliche Datenverarbeitung mbH Goettingen
250-====================================================================
250-This is a Linux PC (Dell PE-2650, 2 CPUs P4/2800, 12 GB RAM)
250-running SuSE-Linux-8.2 with SuSE kernel 2.4.20-64GB-SMP.
250-/pub is ext3, a 2 TB Partition within a Transtec T6100F16R1
250-IDE/FC Raid array with 15+1 SATA disks Hitachi 240 GB Raid5
250-
250-For questions or contacts to the ftp-admin, mailto:emoenke at gwdg.de
250-
250 CWD command successful.
PWD
257 "/pub" is current directory.
TYPE A
200 Type set to A.
PASV
227 Entering Passive Mode (134,76,12,3,231,230)
500 '': command not understood.
list
150 Opening ASCII mode data connection for /bin/ls.
CWD /pub/opensuse
226 Transfer complete.
PWD
250 CWD command successful.
TYPE A
257 "/pub/opensuse" is current directory.
PASV
200 Type set to A.
list
227 Entering Passive Mode (134,76,12,3,232,67)
I do not know where the list or transfer went, but it never made it into my
window.
I then tested with opensuse/kde3/kftpgrabber and I was able to connect and
retrieve a directory listing and transfer files. The same with gftp, login to
the same server and it works just fine. This looks like it may be a generic tde
issue with ftp clients.
This only effects applications/<ftp clients>. Konqueror ftp works fine in TDE.
Type: ftp://david@server/home/david into the addressbar, the askpass dialog pops
up and the connection and listing is perfect. So whatever kftpgrabber and
kasablanca do differently is not working, but what konqueror does -- works!
kftpgrabber likewise shows a connection with the remote server being made, but
no file listing is ever displayed. The big issue with kftpgrabber is a race
condition that occurs when the listing/handshake fails or when connection over
ssh is attempted. It effectively locks the side of kftp the connection occurs on
(left/right) and then pegs the CPU at 40-50%. Quitting kills the hung connection.
A sample of the kftpgrabber connection log is:
[01:53:37] KFTPGrabber logger initialized.
[01:55:41] *** Connecting to nirvana:21...
[01:55:41] *** Connected with server, waiting for welcome message...
[01:55:41] 220 Welcome message
[01:55:41] USER david
[01:55:41] 331 Please specify the password.
[01:55:41] PASS (hidden)
[01:55:42] 230 Login successful.
[01:55:42] SYST
[01:55:42] 215 UNIX Type: L8
[01:55:42] FEAT
[01:55:42] 211-Features:
[01:55:42] EPRT
[01:55:42] EPSV
[01:55:42] MDTM
[01:55:42] PASV
[01:55:42] REST STREAM
[01:55:42] SIZE
[01:55:42] TVFS
[01:55:42] UTF8
[01:55:42] 211 End
[01:55:42] PWD
[01:55:42] 257 "/home/david"
[01:55:42] *** Connected.
[01:55:42] *** Fetching directory listing...
[01:55:42] TYPE A
[01:55:42] 200 Switching to ASCII mode.
[01:55:42] EPSV
[01:55:42] 229 Entering Extended Passive Mode (|||40351|)
[01:55:42] *** Establishing data connection with 192.168.6.17:40351...
Both clients show similar behavior - like some part of the ftp-ioslave isn't
communicating correctly resulting in no listing. Can someone test and confirm
this behavior. Thoughts? Suggestions? Small potatoes apps, but curious neither
are not working.
--
David C. Rankin, J.D.,P.E.
> The issue seems to be in this obscure sqlite source included in
>kmymoney:
>
>23011-qt-sqlite3-0.2.tar.gz
>
>that explains why the files were missed in the q->tq renaming
>scheme.
That explains what?
> I'll see
>if I can come up with some sort of patch, but the git tree
>including an external
>.tar.gz source of sqlite3 seems like a really bad idea :)
Yet this is working for everybody else?
If kmymoney does not include an intrnal sqlite engine, like amarok,
then kmymoney should be built against the external installed
version. There probably is a configuration option to build with the
external version or an option to not build with the internal
version, which then reverts to the external version. Probably
should first test that before ripping sources.
Darrell
>> The only header partially matching that description is:
>>
>> /opt/tqt3/include/ntqsqldriverplugin.h
>>
>> Is that the correct header? If so, what's the 'nt' for?
>>
>
>I think it is supposed to be:
>
>/usr/include/tqt/tqsqldriverplugin.h
>
>looks like a q->tq naming issue....
When compiling tqt3 rather than qt3, the tqtinterface files have a
'n' prefix. Therefore ntqsqldriverplugin.h is the correct name. I
have the same file on my systems.
I performed a full rebuild last night and kmymoney built without
error.
Darrell
Darrell, All,
kmymoney fails right off the bat during build with the following:
==> Building - tde-kmymoney...
make all-recursive
make[1]: Entering directory '/build/tde-kmymoney/src/kmymoney'
Making all in qt-sqlite3-0.2
make[2]: Entering directory '/build/tde-kmymoney/src/kmymoney/qt-sqlite3-0.2'
cd sqlite3 && /opt/tqt3/bin/qmake sqlite3.pro "QMAKE=/opt/tqt3/bin/qmake" -o
Makefile
cd sqlite3 && make -f Makefile
make[3]: Entering directory
'/build/tde-kmymoney/src/kmymoney/qt-sqlite3-0.2/sqlite3'
g++ -c -pipe -g -fvisibility=hidden -fvisibility-inlines-hidden -Wall -W
-march=i686 -mtune=generic -O2 -pipe -fstack-protector --param=ssp-buffer-size=4
-D_REENTRANT -fPIC -DQT_NO_DEBUG -DQT_THREAD_SUPPORT -DQT_PLUGIN -DQT_SHARED
-I/opt/tqt3/mkspecs/default -I. -I/opt/tqt3/include -o smain.o smain.cpp
smain.cpp:37:30: fatal error: qsqldriverplugin.h: No such file or directory
#include <qsqldriverplugin.h>
^
compilation terminated.
The only header partially matching that description is:
/opt/tqt3/include/ntqsqldriverplugin.h
Is that the correct header? If so, what's the 'nt' for?
--
David C. Rankin, J.D.,P.E.
>Please, can you assess whether any of these unintended renaming
>should be
>added to the script convert_existing_qt3_app_to_tqt3? The script
>is still
>actively used for convert KDE3 applications to Trinity and for
>preparing
>updates for existing applications.
I am not intimate with the conversion scripts. If I understand
correctly the following additions will help.
CAM_RETQ_/CAM_REQ_
CHARFRETQ_/CHARFREQ_
CHISTQ_DOF/CHISQ_DOF
CPUFRETQ/CPUFREQ
ETQBACK/EQBACK
ETQ_CHANNELS/EQ_CHANNELS
ETQ_MAX/EQ_MAX
ETQ_NOT/EQ_NOT
ETQ_SET/EQ_SET
ETQVIEW/EQVIEW
ETQUALIZER/EQUALIZER
ETQGAINS/EQGAINS
ETQDATA/EQDATA
FRETQUENCY/FREQUENCY
GETETQDATA/GETEQDATA
HTQ_VIDEOCD/HQ_VIDEOCD
INCRRTQ_WAIT/INCRRQ_WAIT
IRTQ/IRQ
IRTQ_/IRQ_
KONTQ/KONQ
KONTQ_/KONQ_
KONTQSBC/KONQSBC
KONTQFEED/KONQFEED
KONTQFILE/KONQFILE
KONTQSIDEBAR/KONQSIDEBAR
KWTQ_/KWQ_
MPEG_SETQ/MPEG_SEQ
PTQ_/PQ_
RETQSEN/REQSEN
RETQBUF/REQBUF
RETQ_/REQ_
RETQIMG/REQIMG
RTQ_IFS/RQ_IFS
RTQ_SIZE/RQ_SIZE
SAMPLE_FRETQ_/SAMPLE_FREQ_
SETQNO/SEQNO
SETQNO_/SEQNO_
SETQ_NO/SEQ_NO
SETQ_MASK/SEQ_MASK
SETQACCESSMGR/SEQACCESSMGR
SETQLINEWIDGET/SETLINEWIDGET
SETQMAN/SEQMAN
SND_SETQ_EVENT/SND_SEQ_EVENT
SUBTQ_/SUBQ_
KTQ_CLASSNAME/KQ_CLASSNAME
TDETQ_CLASSNAME/TDEQ_CLASSNAME
TQUIET/QUIET
etqdialog/eqdialog
Litquids/Liquids
RetqCount/ReqCount
retquis/requis (French)
Retquisiti/Requisiti (Italian)
setqCount/setCount
"setq "/"set "
stquid/squid
Stquiggle/Squiggle
stquiggle/squiggle
stquiggly/squiggly
tquerytable/querytable
Kontqui/Konqui
kontqui/konqui
Of course, none of these unintended renames affected compiling
because the compiler does not "read" variable names the same way as
humans. to compilers, variables are only "tokens." As long as the
unintended renames are 100% consistent the program compiles just
fine. Fixing the unintended renames is important to us humans only,
to improve readability and establish a degree of professionalism.
Darrell