Re: Conditional Compilation support in INF file

["Gao, Liming" <liming.gao@intel.com>]

Karunakar and Mike:
   Basetools doesn't support Feature Flag Expressions in INF file. It just ignores them. Because there are other solutions, I would like to propose to update INF spec to remove Feature Flag Expressions and avoid the confuse. 

Thanks
Liming
> -----Original Message-----
> From: Kinney, Michael D
> Sent: Saturday, January 12, 2019 2:54 AM
> To: KarunakarPoosapalli@Dell.com; Gao, Liming <liming.gao@intel.com>; lersek@redhat.com; edk2-devel@lists.01.org; Kinney,
> Michael D <michael.d.kinney@intel.com>
> Cc: Sumanth.Vidyadhara@dell.com; Shekar.Babu.S@dell.com; Gao, Liming <liming.gao@intel.com>; Sriramkumar.Raju@dell.com
> Subject: RE: [edk2] Conditional Compilation support in INF file
> 
> Karunakar,
> 
> Feature Flag Expressions is a concept that is defined in the
> specs, but is a feature that is not fully implemented.  Liming
> should be able to provide details on what has been implemented
> and validated in BaseTools.  I think the reason that this feature
> has not been implemented fully is that we have been able to find
> alternate ways to get the equivalent results.  Here are a few
> example techniques:
> 
> 3a: One approach is to use multiple INF files and select the
>     Right INF for different types of platform builds in DSC file.
>     If multiple modules share source files but have different
>     elements produced/consumed/depex, then multiple INF
>     file is an approach that makes the produced/consumed/depex
>     clear and is compatible with UDP Spec.
> 
> 3c: One approach is to use DEBUG_CODE(), DEBUG_CODE_BEGIN(),
>     and DEBUG_CODE_END() macros in a single version of the
>     source files and enable/disable the debug code using BIT2
>     in the following PCD:
> 
>   ## The mask is used to control DebugLib behavior.<BR><BR>
>   #  BIT0 - Enable Debug Assert.<BR>
>   #  BIT1 - Enable Debug Print.<BR>
>   #  BIT2 - Enable Debug Code.<BR>
>   #  BIT3 - Enable Clear Memory.<BR>
>   #  BIT4 - Enable BreakPoint as ASSERT.<BR>
>   #  BIT5 - Enable DeadLoop as ASSERT.<BR>
>   # @Prompt Debug Property.
>   # @Expression  0x80000002 | (gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask & 0xC0) == 0
>   gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask|0|UINT8|0x00000005
> 
> 3c: Another approach is to use multiple INF files and select
>     the one needed for debug/release in DSC file.
> 
> If these techniques, or other techniques that other community
> Member may be using do not work for your use cases, and
> Feature Flag Expressions are the best approach, then Bugzillas
> Can be entered with supporting use cases to justify adding the
> feature to BaseTools.
> 
> Thanks,
> 
> Mike
> 
> > -----Original Message-----
> > From: edk2-devel [mailto:edk2-devel-
> > bounces@lists.01.org] On Behalf Of
> > KarunakarPoosapalli@Dell.com
> > Sent: Thursday, January 10, 2019 10:34 PM
> > To: Kinney, Michael D <michael.d.kinney@intel.com>; Gao,
> > Liming <liming.gao@intel.com>; lersek@redhat.com; edk2-
> > devel@lists.01.org
> > Cc: Sumanth.Vidyadhara@dell.com; Shekar.Babu.S@dell.com;
> > Gao, Liming <liming.gao@intel.com>;
> > Sriramkumar.Raju@dell.com
> > Subject: Re: [edk2] Conditional Compilation support in
> > INF file
> >
> > Hi All,
> >
> > Thank you very much for your valuable thoughts.
> > Below are my concerns/thoughts, Could you please help on
> >
> > 1. Is there any module or INF already using
> > FeatureFlagExpression feature support in current edk2
> > source?
> > 2. If not, could you please help in providing more
> > detailed spec/Doc to verify this support.
> > 3. Below are the few of use cases we're looking for, Did
> > really FeatureFlagExpression support all of this?
> >    		a.	If we've this support we can add different
> > protocols in DEPEX section, like we can put condition
> > check and add rotococol1 for Notebook and Protocol2 for
> > Desktop.
> > 		b.           Help in simplifying build files. We
> > could dispense with the prefix to inf files and more
> > developer convenient.
> > 		c.	Sometimes we would like to add Conditional  .C
> > and .H for file inclusion. Helps in our Debug / release
> > builds as well..
> > 				Example; we do have release and debug
> > build, although Source Section has - IA32,X64, Common,
> > IPF, EBC exclusively.
> > 					  But for the same Source section, we
> > can't have choices based on our build inputs...
> >
> >         						#if Condition1
> >              						  [Sources]
> >                					    Main.c
> >       						#elif Condition2
> >               						 [Sources]
> >                    					    DebugMain.c
> >
> > 4. I don't think INF file support MACRO support, How
> > complex in modifying the BaseTools to support Condition
> > checks and MACRO support?
> >
> > Thanks & Regards,
> > Karunakar
> >
> > -----Original Message-----
> > From: Kinney, Michael D
> > [mailto:michael.d.kinney@intel.com]
> > Sent: Thursday, January 10, 2019 10:23 PM
> > To: Gao, Liming; Laszlo Ersek; Poosapalli, Karunakar;
> > edk2-devel@lists.01.org; Kinney, Michael D
> > Cc: Vidyadhara, Sumanth; Gao, Liming; Raju, SriramKumar
> > Subject: RE: [edk2] Conditional Compilation support in
> > INF file
> >
> >
> > [EXTERNAL EMAIL]
> >
> > There is one additional constraint for adding
> > conditionals to INF files.
> >
> > The UEFI Distribution Packaging Specification defines a
> > format to share package and module content in a standard
> > format and uses an XML schema for the metadata.  We need
> > to be able to convert between INF <--> XML and DEC <-->
> > XML.
> >
> > http://www.uefi.org/sites/default/files/resources/Dist_P
> > ackage_Spec_1_1.pdf
> >
> > If we define extensions to INF or DEC files, we need to
> > make sure these transforms are still supported.  If an
> > extension prevents these transforms, then we either need
> > to change the extension to be compatible or work on an
> > update to the UDP spec to support the extension in the
> > XML.
> >
> > Best regards,
> >
> > Mike
> >
> > > -----Original Message-----
> > > From: edk2-devel [mailto:edk2-devel-
> > > bounces@lists.01.org] On Behalf Of Gao, Liming
> > > Sent: Thursday, January 10, 2019 7:48 AM
> > > To: Laszlo Ersek <lersek@redhat.com>;
> > > KarunakarPoosapalli@Dell.com; edk2-devel@lists.01.org
> > > Cc: Sumanth.Vidyadhara@dell.com; Gao, Liming
> > <liming.gao@intel.com>;
> > > Sriramkumar.Raju@dell.com
> > > Subject: Re: [edk2] Conditional Compilation support in
> > INF file
> > >
> > > I have same question. What flexibility is expected in
> > INF? I see one
> > > request in [Depex] section. So, PCD support in [Depex]
> > is added.
> > >
> > > Edk2 INF is used to describe the source code behavior.
> > > If the source uses Ppi/Protocol/Guid/Pcd, these
> > information are always
> > > required to be described in INF file. The compiler can
> > optimize the
> > > code and remove the unused Ppi/Protocol/Guid/Pcd. It
> > doesn't need
> > > developer specify the conditional statement.
> > >
> > > Thanks
> > > Liming
> > > > -----Original Message-----
> > > > From: Laszlo Ersek [mailto:lersek@redhat.com]
> > > > Sent: Thursday, January 10, 2019 8:54 PM
> > > > To: KarunakarPoosapalli@Dell.com; Gao, Liming
> > > <liming.gao@intel.com>; edk2-devel@lists.01.org
> > > > Cc: Sumanth.Vidyadhara@dell.com;
> > > Sriramkumar.Raju@dell.com
> > > > Subject: Re: [edk2] Conditional Compilation support
> > in
> > > INF file
> > > >
> > > > On 01/10/19 07:03, KarunakarPoosapalli@Dell.com
> > wrote:
> > > > > Hi All,
> > > > >
> > > > > I agree with providing the support like
> > > "FixedAtBuild PCD in INF". And we need to modify or
> > provide support in
> > > BaseTools to support
> > > > this feature.
> > > > >
> > > > > There are more use cases or flexibility to
> > developer
> > > if we support Conditional compilation support in INF.
> > > > > As we're providing support in BaseTools for
> > > FixedAtBuild PCD support in INF, Is there any
> > challenges or drawbacks
> > > in  providing
> > > > conditional compilation support in INF?
> > > >
> > > > This is not for me to say authoritatively, but I'm
> > > unaware of any
> > > > specific use case that cannot be solved without this
> > > feature addition,
> > > > and any further complexity to BaseTools should be
> > > strongly justified.
> > > > "More convenient" is too vague for me, and the
> > > BaseTools code is already
> > > > hard to read and debug.
> > > >
> > > > That's just my opinion, again.
> > > >
> > > > Thanks
> > > > Laszlo
> > > _______________________________________________
> > > edk2-devel mailing list
> > > edk2-devel@lists.01.org
> > > https://lists.01.org/mailman/listinfo/edk2-devel
> > _______________________________________________
> > edk2-devel mailing list
> > edk2-devel@lists.01.org
> > https://lists.01.org/mailman/listinfo/edk2-devel