Yesterday I patched many modules with various amounts of inadvertent "TQ" conversions. Most of them were harmless because the compiler doesn't know how to spell and the inadvertent changes were consistently mispelled.
Thus the packages compiled normally. The sources merely looked weird when reading the code. For example, ETQUAL vs. EQUAL, RETQUIRED vs. REQUIRED, UNITQUE vs. UNIQUE, etc.
The main reason for this cleanup effort is to restore the correct "human readability" of the source code.
However, a handful of the inadvertent conversions might have played a role in various bug reports. Might have.
If you have a favorite pet peeve bug then please rebuild from scratch in GIT and test again. No promises.
I suspect very few bugs will be resolved by this effort, but who knows? I now can see SVG images in the file preview tooltip in Konqueror as well as through the embedded viewer. I would appreciate confirmation from others with this bug (bug report 615).
Darrell
On 04/13/2012 12:55 PM, Darrell Anderson wrote:
However, a handful of the inadvertent conversions might have played a role in various bug reports. Might have.
If you have a favorite pet peeve bug then please rebuild from scratch in GIT and test again. No promises.
I suspect very few bugs will be resolved by this effort, but who knows? I now can see SVG images in the file preview tooltip in Konqueror as well as through the embedded viewer. I would appreciate confirmation from others with this bug (bug report 615).
Darrell
New York City.... Git a rope....
On Sat, Apr 14, 2012 at 12:46 AM, David C. Rankin drankinatty@suddenlinkmail.com wrote:
On 04/13/2012 12:55 PM, Darrell Anderson wrote:
However, a handful of the inadvertent conversions might have played a role in various bug reports. Might have.
If you have a favorite pet peeve bug then please rebuild from scratch in GIT and test again. No promises.
I suspect very few bugs will be resolved by this effort, but who knows? I now can see SVG images in the file preview tooltip in Konqueror as well as through the embedded viewer. I would appreciate confirmation from others with this bug (bug report 615).
Darrell
New York City.... Git a rope....
This TQ circus have taken a year.
On 04/14/2012 01:27 AM, Aleksey Midenkov wrote:
New York City.... Git a rope....
This TQ circus have taken a year.
But good progress has been made and many code cleanups have occurred along the way. Many directions were taken only to learn of a better direction to take.
Large projects take a large amount of coordinated manpower. The larger and more coordinated your manpower, the shorter the time it takes to put the code in the pristine shape required for the next release. The lesser the manpower and the less coordinated the team, then the greater the time required to get everything in shape.
It is the 'fundamental rule of development' - on the same level as gravity or magnetism or the weak or strong forces at the quantum level. Maxims. Whether dealing with the four fundamental forces physics or with the fundamental rule of development - they are all unwaivering...
Work to get others who are competent and share the same interest involved with TDE. The will allow us to use the fundamental rule to our benefit :)
On Sat, Apr 14, 2012 at 10:48 AM, David C. Rankin drankinatty@suddenlinkmail.com wrote:
On 04/14/2012 01:27 AM, Aleksey Midenkov wrote:
New York City.... Git a rope....
This TQ circus have taken a year.
But good progress has been made and many code cleanups have occurred along the way. Many directions were taken only to learn of a better direction to take.
Large projects take a large amount of coordinated manpower. The larger and more coordinated your manpower, the shorter the time it takes to put the code in the pristine shape required for the next release. The lesser the manpower and the less coordinated the team, then the greater the time required to get everything in shape.
In case of small and not well coordinated manpower the project should think hardly of its priorities. In this project the goal and essential priority is Usability. This is about numerous UI bugs and blunders, subquality work that should be completed. This makes clear direction. Not mere wandering in woods of code. That will put project in wrong direction.
Timothy, please, remove TQ -> Q proxying. This direction is not good and sooner or later will be reverted. Better sooner, less double work.
On Mon, Apr 16, 2012 at 10:41 AM, Aleksey Midenkov midenok@gmail.com wrote:
In case of small and not well coordinated manpower the project should think hardly of its priorities. ...
... think *hard* ...
In case of small and not well coordinated manpower the project should think hardly of its priorities. In this project the goal and essential priority is Usability. This is about numerous UI bugs and blunders, subquality work that should be completed. This makes clear direction. Not mere wandering in woods of code. That will put project in wrong direction.
Timothy, please, remove TQ -> Q proxying. This direction is not good and sooner or later will be reverted. Better sooner, less double work.
I admit I have my frustrations with the TQ layer. We have yet to see any direct benefit that I am aware. But I don't have the energy or knowledge to do anything about that. I want to believe there is a long term benefit and I'm willing to let the issue go at that. For now.
The issue could be water under the bridge if we addressed the bug tracker. When usability is restored to at least 3.5.10 days then I believe users will ignore the TQ debate. (Developers might not but users will.)
With all of the "Inadvertent TQ" patches I pushed we now can say that inadvertent TQ conversions are unlikely to be the cause of a bug. Although, as David and I discovered this weekend, merely restoring the inadvertent changes can still leave a hidden bug --- something that was patched to overcome the original inadvertent changes.
Some of the usability bugs that are irritating people:
956 Kdcop: Fails to run 922 When logging out with unsaved file, trinity does not ask to save it 897 sftp kioslave fails to parse urls in konqueror address bar 859 Kaffeine does not block screensaver 760 TDE exits too slow 714 kio_man does not work under RHEL 657 Can't start KWord or KPresenter from T-Menu 615 KSVG shows only black previews 586 3.5.13 regression: kio-locate-trinity doesn't show any results in Konqueror 407 [Regression] KMail text disappears 258 ksmserver: Logout confirmation fadeaway is too slow with older hardware
Possibly fixed by reverting previous patches:
927 Kate: Can't concurrently run kate with tdesu (revert patch from bug report 394) 756 Konqueror: Does not update file pane with file changes (7 year old patch!) 692 Kate: focus is broken when using the --use parameter (Needs focus patch and special application settings) 406 [Regression] KMail always checks mail after sending 394 Problems with DCOP After Using kdesu
The SAK issues:
925 [kdesktop] SAK driven secure dialog is not available for use 906 SAK realization is mostly buggy for KDM 898 tsak process taking 90-100% of CPU
Then there are issues such as:
640 .kthemestylerc.lock is created in $TDEDIR/share/config 301 Kompare does not maintain a correct mru list 293 Kde-config Incorrectly Creates a Profile Folder in the System root directory 259 Kompare does not always terminate properly, leaving the process running
We can fix these bugs, which would put the TQ issue behind us.
Darrell
On Mon, Apr 16, 2012 at 11:14 AM, Darrell Anderson humanreadable@yahoo.com wrote:
In case of small and not well coordinated manpower the project should think hardly of its priorities. In this project the goal and essential priority is Usability. This is about numerous UI bugs and blunders, subquality work that should be completed. This makes clear direction. Not mere wandering in woods of code. That will put project in wrong direction.
Timothy, please, remove TQ -> Q proxying. This direction is not good and sooner or later will be reverted. Better sooner, less double work.
I admit I have my frustrations with the TQ layer. We have yet to see any direct benefit that I am aware.
There is no real benefits. I believe that any benefits may be done without TQ layer.
But I don't have the energy or knowledge to do anything about that. I want to believe there is a long term benefit and I'm willing to let the issue go at that. For now.
The issue could be water under the bridge if we addressed the bug tracker. When usability is restored to at least 3.5.10 days then I believe users will ignore the TQ debate. (Developers might not but users will.)
With all of the "Inadvertent TQ" patches I pushed we now can say that inadvertent TQ conversions are unlikely to be the cause of a bug. Although, as David and I discovered this weekend, merely restoring the inadvertent changes can still leave a hidden bug --- something that was patched to overcome the original inadvertent changes.
Some of the usability bugs that are irritating people:
...
I can add a lot of usability bugs from my list. The most irritable (though not a bug) is that cannot upgrade python to latest version due to kde-guidance package limitation. This holds a lot of software from being installed/upgraded or requires efforts for keeping several snapshots of system. The important bugs that are on my mind:
kpowersave segfaults (no battery indicator); gwenview segfaults; rudimental kbluetooth (no real world functionality); network-manager-kde frontend does not support latest network-manager API
There are lots of bugs and minor flaws that I keep recorded and intend to fix like:
stupid message dialog popups in iteration loops; wrong Shift selection in Konqueror, etc.; a lot of kmail bugs including hangs, UI freezes, network and protocol incompleteness; kpowersave incompatibility with TuxOnIce; kdevelop UI incompleteness, etc.; incomplete handling of keyboard/mouse settings which are not restored on eject/insert; and many, many more!
We can fix these bugs, which would put the TQ issue behind us.
The problem is that this TQ issue caused (and will cause in future) a lot of trouble. It is a blocker along with rushed build system development that hold me from development. The year that have spent on that could be much more productive. This makes me even doubt in project future because users go away when they see no improvements in such a long time. And python conflict is the one of good reasons for that.
I can add a lot of usability bugs from my list. The most irritable (though not a bug) is that cannot upgrade python to latest version due to kde-guidance package limitation. This holds a lot of software from being installed/upgraded or requires efforts for keeping several snapshots of system. The important bugs that are on my mind:
kpowersave segfaults (no battery indicator); gwenview segfaults; rudimental kbluetooth (no real world functionality); network-manager-kde frontend does not support latest network-manager API
There are lots of bugs and minor flaws that I keep recorded and intend to fix like:
stupid message dialog popups in iteration loops; wrong Shift selection in Konqueror, etc.; a lot of kmail bugs including hangs, UI freezes, network and protocol incompleteness; kpowersave incompatibility with TuxOnIce; kdevelop UI incompleteness, etc.; incomplete handling of keyboard/mouse settings which are not restored on eject/insert; and many, many more!
Have these bugs/concerns been filed?
I am not trying to sound tripe or pompous! I follow the bug tracker but I never heard of some of these issues. Some of them sound irritating indeed.
Darrell
<snip>
There are lots of bugs and minor flaws that I keep recorded and intend to fix like:
stupid message dialog popups in iteration loops; wrong Shift selection in Konqueror, etc.; a lot of kmail bugs including hangs, UI freezes, network and protocol incompleteness; kpowersave incompatibility with TuxOnIce; kdevelop UI incompleteness, etc.; incomplete handling of keyboard/mouse settings which are not restored on eject/insert; and many, many more!
Many of those bugs can now be addressed since TDE has just gained a fledgling hardware management library in GIT. Can you please file bug reports on them, especially the kpowersave issues and hardware insert/eject problems?
Thanks!
Tim
-
Aleksey Midenkov -
Darrell Anderson -
David C. Rankin -
Timothy Pearson