On July 27, 2003 04:46 pm, Gregory M. Turner wrote:
WARN("FDIIsCabinet failed.\n");
ERR("FDIIsCabinet failed.\n");
Why these changes, they don't seem to make sense. The function handles the problem just fine, no need to spam the user with such messages, we already have too many IMO.
On Sunday 27 July 2003 09:35 pm, Dimitrie O. Paun wrote:
On July 27, 2003 04:46 pm, Gregory M. Turner wrote:
WARN("FDIIsCabinet failed.\n");
ERR("FDIIsCabinet failed.\n");Why these changes, they don't seem to make sense. The function handles the problem just fine, no need to spam the user with such messages, we already have too many IMO.
There is a "legitimate" way for this to occur, which makes it arguably excessive, but my reasoning was that 99% of the time this would instead be the result of some prior error which had occurred silently... Since there was no memory leak FIXME in fdi.c we get a failure in msiexec or whatever and it gives some message like "File not found" but debugmsg +file reveals nothing because no file operation occurs...
I dunno I guess I thought maybe cabinet.dll would just be silent in this case and too easy to miss as a culprit... maybe it's stupid.
On July 27, 2003 11:32 pm, Gregory M. Turner wrote:
I dunno I guess I thought maybe cabinet.dll would just be silent in this case and too easy to miss as a culprit... maybe it's stupid.
I guess if you think they will signal real ERRes most of the time, and your confidence in the code is low, they kinda make sense. I'm just a bit worried about the number of ERRes we produce...
-
Dimitrie O. Paun -
Gregory M. Turner