radeon driver

I’ve seen some subtle problems when running a randeon-driven pc.

Missrendering of tables and scrolling problems:
This bug affects mach64 cards as well the randeon one (I can’t know if
this problem affects other cards at the moment).

When browsing some text pages under voyager with many tables, or, with
tables that have a very small border the initial rendering is fine
but when scrolling some pixels gets eliminated.

For example:

|---------------|
|--------------|| ← second line of black pixels that becomes white
|| ||

---------------

Here you can see an ugly ascii table :slight_smile:, generated by a code like
like this:

yo

When scrolling, the black border often is miss-scrolled and only
one of the two black pixels mantains the black color (the other
becomes white).

I can grab this issue if can help.


Missrefreshing jpeg images:
Affects the readeon driver only. Actually to reproduce it I do the
following:

I open a web page that contains at least one jpeg, any size,
no smoothing, quality 90%, 1x1x1 subsampling with floating point DCT
method (if can help fixing the problem).

Again, the initial rendering is fine (the image is perfectly
rendered). But if you hide the window and cause an expose event
(for example by switching the console) many pixels of that image
takes the background color of the page.

Again, I can grab it.

Bests.


Wave++

Hello Wave++,

This issue was recently fixed and should be available in the next release.

regards,
Dave B.




Wave++ wrote:

I’ve seen some subtle problems when running a randeon-driven pc.

Missrendering of tables and scrolling problems:
This bug affects mach64 cards as well the randeon one (I can’t know if
this problem affects other cards at the moment).

When browsing some text pages under voyager with many tables, or, with
tables that have a very small border the initial rendering is fine
but when scrolling some pixels gets eliminated.

For example:

|---------------|
|--------------|| ← second line of black pixels that becomes white
|| ||

---------------

Here you can see an ugly ascii table > :slight_smile:> , generated by a code like
like this:

table border=0 cellpadding=0 cellspacing=2 bgcolor=black
tr>yo</tr
/td></tr
/table

When scrolling, the black border often is miss-scrolled and only
one of the two black pixels mantains the black color (the other
becomes white).

I can grab this issue if can help.


Missrefreshing jpeg images:
Affects the readeon driver only. Actually to reproduce it I do the
following:

I open a web page that contains at least one jpeg, any size,
no smoothing, quality 90%, 1x1x1 subsampling with floating point DCT
method (if can help fixing the problem).

Again, the initial rendering is fine (the image is perfectly
rendered). But if you hide the window and cause an expose event
(for example by switching the console) many pixels of that image
takes the background color of the page.

Again, I can grab it.

Bests.

When??? what version?

Greg,


apps <apps@qnx.com> wrote in message news:3C3A1EA8.2010104@qnx.com

Hello Wave++,

This issue was recently fixed and should be available in the next release.

regards,
Dave B.




Wave++ wrote:

I’ve seen some subtle problems when running a randeon-driven pc.

Missrendering of tables and scrolling problems:
This bug affects mach64 cards as well the randeon one (I can’t know if
this problem affects other cards at the moment).

When browsing some text pages under voyager with many tables, or, with
tables that have a very small border the initial rendering is fine
but when scrolling some pixels gets eliminated.

For example:

|---------------|
|--------------|| ← second line of black pixels that becomes white
|| ||

---------------

Here you can see an ugly ascii table > :slight_smile:> , generated by a code like
like this:

table border=0 cellpadding=0 cellspacing=2 bgcolor=black
tr>yo</tr
/td></tr
/table

When scrolling, the black border often is miss-scrolled and only
one of the two black pixels mantains the black color (the other
becomes white).

I can grab this issue if can help.


Missrefreshing jpeg images:
Affects the readeon driver only. Actually to reproduce it I do the
following:

I open a web page that contains at least one jpeg, any size,
no smoothing, quality 90%, 1x1x1 subsampling with floating point DCT
method (if can help fixing the problem).

Again, the initial rendering is fine (the image is perfectly
rendered). But if you hide the window and cause an expose event
(for example by switching the console) many pixels of that image
takes the background color of the page.

Again, I can grab it.

Bests.
\

I believe he is refering to 6.2, which is comming out soon.

E.


Greg Orvets <orvets@lambdatech.com> wrote:

When??? what version?

Greg,



apps <> apps@qnx.com> > wrote in message news:> 3C3A1EA8.2010104@qnx.com> …
Hello Wave++,

This issue was recently fixed and should be available in the next release.

regards,
Dave B.




Wave++ wrote:

I’ve seen some subtle problems when running a randeon-driven pc.

Missrendering of tables and scrolling problems:
This bug affects mach64 cards as well the randeon one (I can’t know if
this problem affects other cards at the moment).

When browsing some text pages under voyager with many tables, or, with
tables that have a very small border the initial rendering is fine
but when scrolling some pixels gets eliminated.

For example:

|---------------|
|--------------|| ← second line of black pixels that becomes white
|| ||

---------------

Here you can see an ugly ascii table > :slight_smile:> , generated by a code like
like this:

table border=0 cellpadding=0 cellspacing=2 bgcolor=black
tr>yo</tr
/td></tr
/table

When scrolling, the black border often is miss-scrolled and only
one of the two black pixels mantains the black color (the other
becomes white).

I can grab this issue if can help.


Missrefreshing jpeg images:
Affects the readeon driver only. Actually to reproduce it I do the
following:

I open a web page that contains at least one jpeg, any size,
no smoothing, quality 90%, 1x1x1 subsampling with floating point DCT
method (if can help fixing the problem).

Again, the initial rendering is fine (the image is perfectly
rendered). But if you hide the window and cause an expose event
(for example by switching the console) many pixels of that image
takes the background color of the page.

Again, I can grab it.

Bests.
\