Hello,
load_tm() function uses a mixture of find_export and direct assignment to export the API for TM. I would like to update to the last one, since will cleanup the module exports structure and avoid casting problems when functions prototypes change.
Then I see that tm uses some defines for the names of the several functions exported to config, e.g., T_FORWARD_NONACK, ..., which will have no other usage, I would go for using directly the names in module exports - makes life easier when browsing the code and will become coherent.
Anyone against?
Thanks, Daniel
Hey,
Daniel-Constantin Mierla wrote:
load_tm() function uses a mixture of find_export and direct assignment to export the API for TM. I would like to update to the last one, since will cleanup the module exports structure and avoid casting problems when functions prototypes change.
Anyone against?
No, just an additional note: For the sake of uniformity, we should do the same with the dialog module as it uses the same kind of mixture.
Not sure about other modules' APIs though -- what about sl and rr?
Cheers,
--Timo
Hello,
On 4/17/10 2:21 PM, Timo Reimann wrote:
Hey,
Daniel-Constantin Mierla wrote:
load_tm() function uses a mixture of find_export and direct assignment to export the API for TM. I would like to update to the last one, since will cleanup the module exports structure and avoid casting problems when functions prototypes change.
Anyone against?
No, just an additional note: For the sake of uniformity, we should do the same with the dialog module as it uses the same kind of mixture.
Not sure about other modules' APIs though -- what about sl and rr?
where is in dialog? I checked and couldn't spot quickly?
in k we should have updated to use load_xyz() instead of find_export for each XYZ API function, because of issued caused by changes in prototypes not signaled in other modules since they were casted. Maybe some were forgotten, if you stop some, report them.
Thanks, Daniel
On 4/17/10 2:29 PM, Daniel-Constantin Mierla wrote:
Hello,
On 4/17/10 2:21 PM, Timo Reimann wrote:
Hey,
Daniel-Constantin Mierla wrote:
load_tm() function uses a mixture of find_export and direct assignment to export the API for TM. I would like to update to the last one, since will cleanup the module exports structure and avoid casting problems when functions prototypes change.
Anyone against?
No, just an additional note: For the sake of uniformity, we should do the same with the dialog module as it uses the same kind of mixture.
Not sure about other modules' APIs though -- what about sl and rr?
where is in dialog? I checked and couldn't spot quickly?
in k we should have updated to use load_xyz() instead of find_export for each XYZ API function, because of issued caused by changes in prototypes not signaled in other modules since they were casted. Maybe some were forgotten, if you stop
well, spot instead of stop ... same letters different meaning :-)
Daniel
some, report them.
Thanks, Daniel
Daniel-Constantin Mierla wrote:
Daniel-Constantin Mierla wrote:
load_tm() function uses a mixture of find_export and direct assignment to export the API for TM. I would like to update to the last one, since will cleanup the module exports structure and avoid casting problems when functions prototypes change.
Anyone against?
No, just an additional note: For the sake of uniformity, we should do the same with the dialog module as it uses the same kind of mixture.
Not sure about other modules' APIs though -- what about sl and rr?
where is in dialog? I checked and couldn't spot quickly?
in k we should have updated to use load_xyz() instead of find_export for each XYZ API function, because of issued caused by changes in prototypes not signaled in other modules since they were casted. Maybe some were forgotten, if you stop
I just now took a look at the tm module and didn't realize before that each and every API function (and not just the API entry point) is loaded by means of find_export.
Sorry about the fuss -- you're right that the dialog module already is cleaned up in this regards, so...
well, spot instead of stop ... same letters different meaning :-)
there's nothing more to spot which makes me stop. :)
Cheers,
--Timo
-
Daniel-Constantin Mierla -
Timo Reimann