Improve decode block error logging #1353
Conversation
|
CI gfxreconstruct build queued with queue ID 83942. |
|
CI gfxreconstruct build # 3498 running. |
|
This is a great addition! What do you think about adding the "frame %lu, block %lu" to the end of the error messages? I'm not sure if Bob's scripts or any scripts in the wild expect to find these existing error messages by searching for "[gfxrecon] WARNING - Failed" or "[gfxrecon] WARNING - Missing". I also think you could omit the square brackets in that case; parsing out those frame numbers and block numbers should be pretty straightforward. |
|
CI gfxreconstruct build queued with queue ID 83983. |
MarkY-LunarG
force-pushed
the
marky_improve_unrecog_message
branch
from
79d9d8f to
900aac0
Compare
|
CI gfxreconstruct build queued with queue ID 84039. |
I talked to Bob and he said his scripts will handle square brackets or parentheses or whatever. In fact, he would only look at this manually if an issue arose, he has no plans to automate a script on this. Although he says he could simply using regular expressions if he needed to. Therefore, I moved it to the end as you suggested and put them in parentheses instead. Additionally, I lower-cased "frame" and "block" based on his suggestion. |
|
Adding @lunarpapillo just so he can sit in and see his name being thrown around so Cavalierly. |
|
CI gfxreconstruct build # 3503 running. |
|
CI gfxreconstruct build # 3503 passed. |
When a decode block error occurs, indicate which frame and block are responsible. This should help at least diagnose if a block is at the begining or end of a trace.
MarkY-LunarG
force-pushed
the
marky_improve_unrecog_message
branch
from
900aac0 to
8bca77a
Compare
|
CI gfxreconstruct build queued with queue ID 84977. |
|
CI gfxreconstruct build # 3517 running. |
|
CI gfxreconstruct build # 3517 passed. |
|
Oh cool, I didn't actually know about those format specifiers. They do exactly what we want (https://www.phind.com/search?cache=llnuuji1691dcyfheuxw3lvs, https://en.cppreference.com/w/cpp/header/cinttypes). As long as nobody changes the types of these members later, the casts I suggested will never be needed. |
When a decode block error occurs, indicate which frame and block are responsible. This should help at least diagnose if a block is at the begining or end of a trace.