Parser for includes.

[threonorm]

Depending on how we comment an include :

//`include foo
// `include bar

the script for dependency tracking consider that foo is a dependency, but not bar. When I expect the two to not be dependencies (as they are commented).

I guess connectal used the first form of comment as a feature to signal weird dependies invisible to bluespec but visible to connectal?

The code involved is in bsvdependencies.py:

 m = re.match('//`include "([^\"]+)"', line)
 m1 = re.match('//`include(.*)', line)
 if m:
                iname = m.group(1)
                if iname in abspaths:
                    iname = abspaths[iname]
                else:
                    iname = 'obj/%s' % iname
                includes.append(iname)

I am not sure why this special case. I don't have the full picture.
If connectal really need to do this kind of commented include that should not be seen by BSV, perhaps connectal could use a special form of include like `includeConnectal, that would be in the comment section?

[hanw]

Hi threonorm, Jamey is on vacation with limited email access, so your question may not be answered in time. It sounds like the error is due to some hacky-ness in bsvdependencies.py , would you be able suggest a fix for this? I think you are correct that none of these two cases should generate a dependency.

[threonorm]

Hi hanw, thanks for your answer!

I have a "fix" where I just consider every commented dependency to be not a dependency, this work for my use of the bsvdependencies functions, but I suspect that Jamey uses it in a different way somewhere else where he actually wants to keep this feature. (The way the code is written make me believe that this is really a feature)

For this reason, I don't want to pull request yet, waiting for his opinion is definitely a good idea :).

[jameyhicks]

I guess it would be less Jaclyn if bsvpreprocess returned a list of included files as well as the expanded source code. Do you have time to make that change? Then it could ignore all commented out includes.

[jameyhicks]

autocorrect strikes again :)

[jameyhicks]

In case you did not see the comments I made in one of the chat channels: I think this is an artifact of the interaction between bsvpreprocess and bsvdepend. It's a bit too hackish, now, so probably it would be better if bsvpreprocess returned a list of the included files along with the preprocessed text and then we do not have to treat commented out includes specially. Either that, or we need a much more unique tag for the commented includes. How about a commented out attribute? // (* included_file="filename.bsvi" *) It's too bad the BSV compiler does not support the equivalent of C/C++ #line directive, which is used for just this purpose. If you have a chance to make a pull request along either of these lines I would greatly appreciate it. Happy holidays!

…

On Fri, Dec 2, 2016 at 3:03 PM threonorm ***@***.***> wrote: Hi hanw, thanks for your answer! I have a "fix" where I just consider every commented dependency to be not a dependency, this work for my use of the bsvdependencies functions, but I suspect that Jamey uses it in a different way somewhere else where he actually wants to keep this feature. (The way the code is written make me believe that this is really a feature) For this reason, I don't want to pull request yet, waiting for his opinion is definitely a good idea :). — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#132 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACU3s-ktGy7VqAi3VAL82uMkgZU-4Kluks5rEHmjgaJpZM4LC4WU> .

[threonorm]

Hello Jamey, I am not sure I understand the problem properly : I don't understand what is this special kind of include for and where/what it used for. (What are those include that are commented, but still considered as include?) Though if abstractly speaking you just need a special kind of include, only seen by connectal build system, and added by connectal automatically somehow, I think I agree with your second suggestion, I can modify bsvdependencies to special pick something like //CONNECTALINCLUDE "nameoffile" I can definitely make the one line change in bsvdependencies, but I would break everything by doing so, because I don't know where I should change things as well. (I guess some script add those special include for connectal somewhere : I think those special include are not hand written, are they ? And if they are, it means that I will break every project that is using them.). Thomas Excerpts from Jamey Hicks's message of 2016-12-30 08:10:02 -0500:

…

In case you did not see the comments I made in one of the chat channels: I think this is an artifact of the interaction between bsvpreprocess and bsvdepend. It's a bit too hackish, now, so probably it would be better if bsvpreprocess returned a list of the included files along with the preprocessed text and then we do not have to treat commented out includes specially. Either that, or we need a much more unique tag for the commented includes. How about a commented out attribute? // (* included_file="filename.bsvi" *) It's too bad the BSV compiler does not support the equivalent of C/C++ #line directive, which is used for just this purpose. If you have a chance to make a pull request along either of these lines I would greatly appreciate it. Happy holidays! On Fri, Dec 2, 2016 at 3:03 PM threonorm ***@***.***> wrote: > Hi hanw, thanks for your answer! > > I have a "fix" where I just consider every commented dependency to be not > a dependency, this work for my use of the bsvdependencies functions, but I > suspect that Jamey uses it in a different way somewhere else where he > actually wants to keep this feature. (The way the code is written make me > believe that this is really a feature) > > For this reason, I don't want to pull request yet, waiting for his opinion > is definitely a good idea :). > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#132 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ACU3s-ktGy7VqAi3VAL82uMkgZU-4Kluks5rEHmjgaJpZM4LC4WU> > . >