Reviewed-by: Maciej Rabeda <maciej.rab...@linux.intel.com>
On 06-Dec-21 02:17, Michael D Kinney wrote:
Hello EDK II Maintainers, A detailed evaluation of the DEBUG_CODE() formatting issue has been completed. The reason DEBUG_CODE() is a challenge is that this looks like a macro from a C parsing perspective, but the EDK II usage places C statements or blocks of C code as the parameter to this macro. There are actually 2 methods available to mark a statement or a block of code to be included when the DEBUG_PROPERTY_DEBUG_CODE_ENABLED bit is enabled in PcdDebugPropertyMask. One is DEBUG_CODE() and the other is to mark the beginning and end of a code block with DEBUG_CODE_BEGIN() and DEBUG_CODE_END(). In fact, DEBUG_CODE() is implemented using DEBUG_CODE_BEGIN() and DEBUG_CODE_END() macros. #define DEBUG_CODE(Expression) \ DEBUG_CODE_BEGIN (); \ Expression \ DEBUG_CODE_END () A complete review for the use of these DEBUG_CODE macros was performed on the edk2 repo. Uncrustify performs good formatting for code blocks between DEBUG_CODE_BEGIN() and DEBUG_CODE_END(). This is because these look like simple macros calls with no parameters and the lines of C code between these 2 macros is formatted correctly. The uncrustify formatting issues are only present with the use of DEBUG_CODE(). Simple use cases of DEBUG_CODE(Expression) where Expression is a single C statement also look correct. A medium complexity use case where Expression is a code block of simple statements or even some local variables and simple statements also look correct. It is only complex code blocks that use C statements such as if/for/while/case that include the use of braces {} does uncrustify perform incorrect formatting. The recommended solution to this issue is to convert the use of DEBUG_CODE() to DEBUG_CODE_BEGIN() / DEBUG_CODE_END() for cases where the Expression passed to DEBUG_CODE() is the complex use case that contains statements that use braces {}. There are 57 instances of this pattern across 40 files in the edk2 repo. I have posted a branch with these additional patches: https://github.com/mdkinney/edk2/tree/Bug_3737_3760_3739_ApplyUncrustifyChanges_V7 I have performed CompareBuild tests with this revised patch series with the DEBUG_CODE changes. It passes 100% showing no binary differences. https://github.com/mdkinney/edk2/actions/runs/1542454606 I have opened a PR to run this patch series through EDK II CI. It also passes 100%. https://github.com/tianocore/edk2/pull/2236 The summary of changes made since the V6 review are: 1) Change uncrustify configuration assignment alignment threshold to 0 align_assign_thresh = 0 2) Replace "<param>, OPTIONAL" with "<param> OPTIONAL," 3) Replace DEBUG_CODE(Expression) with DEBUG_CODE_BEGIN(); Expression DEBUG_CODE_END() if Expression is complex and contains braces {}. 4) No changes to uncrustify tool required. Please review the differences between the following 2 branches and provide feedback or a Series Reviewed-by if you agree with these additional changes. https://github.com/mdkinney/edk2/tree/Bug_3737_3760_3739_ApplyUncrustifyChanges_V6 https://github.com/mdkinney/edk2/tree/Bug_3737_3760_3739_ApplyUncrustifyChanges_V7 The goal is to complete the review and get the uncrustify change committed tomorrow so the extended hard freeze can be lifted. Thanks, Mike-----Original Message----- From: Kinney, Michael D <michael.d.kin...@intel.com> Sent: Thursday, December 2, 2021 6:23 PM To: Michael Kubacki <mikub...@linux.microsoft.com>; devel@edk2.groups.io; maciej.rab...@linux.intel.com; Michael Kubacki <michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm <l...@nuviainc.com>; Kinney, Michael D <michael.d.kin...@intel.com> Subject: RE: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 Hello EDK II Maintainers, I have entered BZ 3760 to make the use of the OPTIONAL keyword style consistent for all of edk2 repo and to be compatible with uncrustify. I have posted the following V6 branch that does the EFI_D_* to DEBUG_* changes, the OPTIONAL keyword style changes, and the uncrustify changes with the one configuration change for assignment alignment. https://github.com/mdkinney/edk2/tree/Bug_3737_3760_3739_ApplyUncrustifyChanges_V6 Please provide feedback on the code style in this branch with the known DEBUG_CODE() issue still present. If we are able to quickly update uncrustify to handle DEBUG_CODE(), I will generate a V7. Thanks, Mike-----Original Message----- From: Kinney, Michael D <michael.d.kin...@intel.com> Sent: Thursday, December 2, 2021 4:53 PM To: Michael Kubacki <mikub...@linux.microsoft.com>; devel@edk2.groups.io; maciej.rab...@linux.intel.com; Michael Kubacki <michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm <l...@nuviainc.com>; Kinney, Michael D <michael.d.kin...@intel.com> Subject: RE: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 Michael, Yes. Please update the patch series that adds UncrustifyCheck with those changes. Thanks, Mike-----Original Message----- From: Michael Kubacki <mikub...@linux.microsoft.com> Sent: Thursday, December 2, 2021 4:31 PM To: Kinney, Michael D <michael.d.kin...@intel.com>; devel@edk2.groups.io; maciej.rab...@linux.intel.com; MichaelKubacki<michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm <l...@nuviainc.com> Subject: Re: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 Hi Mike, Thank you for the detailed analysis and recommendations. I agree with the recommendations and I will try to have an Uncrustify tool update by tomorrow for (4). That will require an update in uncrustify_ext_dep.yaml to pull in the new version. I'm assuming I should also update uncrustify.cfg to set align_assign_thresh = 0 in that new patch series? Regards, Michael On 12/2/2021 7:18 PM, Kinney, Michael D wrote:Hi Michael, CORRECTION: set align_assign_threshold to 0. Reponses inline below. I would like to summarize the 4 issues raised in the past day along with the recommendations. 1) Exclusion feature for UncrustifyCheck. There are 2 directories with 8 files total that maintainers have noted they would like to see not go through uncrustify formatting. Today the only content that is skipped is BaseTools and submodules. Adding a general purpose exclusion feature would then require all developers to make sure their method of using uncrustify also excludes those same areas. This requires extra steps for all developers and maintainers. If we do not add the exclusion feature, then the 8 files will require an extra step to sync with the original source of those files. The rate of changes of these 8 files is very low today. RECOMMENDATION: Do not add exclusion feature at this time. Revisit if the extra work to maintain the files that would be candidates for exclusions increases significantly. 2) Alignment of assignments. The threshold of 4 characters appears to be too low and causes source files that are already aligned to become unaligned. RECOMMENDATION: Change threshold to the default value of 0 which means no limit. align_assign_thresh= 0 3) Alignment of parameters in function declaration not correct. The root cause of this is the use of the OPTIONAL keyword. If the OPTIONAL keyword is removed, then the alignment is correct. The alignment is also correct if the OPTIONAL keyword appears before the ','. If the OPTIONAL keyword appears after the ',', then the format is not correct. The OPTIONAL keyword indicates that the parameter in the function is not required and may be passed in as NULL or 0 or some other default value defined by the API. It makes more sense for this OPTIONAL keyword that follows the parameter names to appear before the ',' so it is scoped to the parameter on that line. If it appears after the ',', then C parsers thinks it is a prefix (IN, OUT, CONST, volatile, static) for the next parameter in the function. RECOMMENDATION: Update patch series with a global search and replace so OPTIONAL keyword always appears before the ',' on the same line. RegEx search string: ',( *)OPTIONAL( *)' RegEx replace string: ' $1OPTIONAL,$2' 4) Format issues with complex blocks in DEBUG_CODE(), or between DEBUG_CODE_BEGIN() and DEBUG_CODE_END(). Uncrustify treats these as macros and is not aware that the parameter passed into the macro call is a block of C code that needs to be formatted. Complex blocks with if/while/for/case statements are impacted the most. RECOMMENDATION: Update the uncrustify with an edk2 specific extension to treat these macros as a block of code as if they were surrounded by an extra set of braces {}. I have posted a branch for testing purposes that implements (2) and (3). Branch: https://github.com/mdkinney/edk2/tree/TestOnly_Bug_3737_3739_ApplyUncrustifyChanges_V6_OPTIONAL_Keyword_Fix PR: https://github.com/tianocore/edk2/pull/2233 Status: PASS CompareBuild: https://github.com/mdkinney/edk2/actions/runs/1532855914 Status: PASS You can see what changed by fetching and comparing the following 2 branches: https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5 https://github.com/mdkinney/edk2/tree/TestOnly_Bug_3737_3739_ApplyUncrustifyChanges_V6_OPTIONAL_Keyword_Fix Please provide feedback on the RECOMMENDATIONS above. I will go ahead and prepare of V6 version of the patch series now that that test results are all PASS. Best regards, Mike-----Original Message----- From: Kinney, Michael D <michael.d.kin...@intel.com> Sent: Thursday, December 2, 2021 4:15 PM To: Michael Kubacki <mikub...@linux.microsoft.com>; devel@edk2.groups.io; maciej.rab...@linux.intel.com; MichaelKubacki<michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm<l...@nuviainc.com>;Kinney, Michael D <michael.d.kin...@intel.com> Subject: RE: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 Hi Michael, Reponses inline below. I would like to summarize the 4 issues raised in the past day along with the recommendations. 1) Exclusion feature for UncrustifyCheck. There are 2 directories with 8 files total that maintainers have noted they would like to see not go through uncrustify formatting. Today the only content that is skipped is BaseTools and submodules. Adding a general purpose exclusion feature would then require all developers to make sure their method of using uncrustify also excludes those same areas. This requires extra steps for all developers and maintainers. If we do not add the exclusion feature, then the 8 files will require an extra step to sync with the original source of those files. The rate of changes of these 8 files is very low today. RECOMMENDATION: Do not add exclusion feature at this time. Revisit if the extra work to maintain the files that would be candidates for exclusions increases significantly. 2) Alignment of assignments. The threshold of 4 characters appears to be too low and causes source files that are already aligned to become unaligned. RECOMMENDATION: Change threshold to the default value of 0 which means no limit. align_assign_thresh= 4 3) Alignment of parameters in function declaration not correct. The root cause of this is the use of the OPTIONAL keyword. If the OPTIONAL keyword is removed, then the alignment is correct. The alignment is also correct if the OPTIONAL keyword appears before the ','. If the OPTIONAL keyword appears after the ',', then the format is not correct. The OPTIONAL keyword indicates that the parameter in the function is not required and may be passed in as NULL or 0 or some other default value defined by the API. It makes more sense for this OPTIONAL keyword that follows the parameter names to appear before the ',' so it is scoped to the parameter on that line. If it appears after the ',', then C parsers thinks it is a prefix (IN, OUT, CONST, volatile, static) for the next parameter in the function. RECOMMENDATION: Update patch series with a global search and replace so OPTIONAL keyword always appears before the ',' on the same line. RegEx search string: ',( *)OPTIONAL( *)' RegEx replace string: ' $1OPTIONAL,$2' 4) Format issues with complex blocks in DEBUG_CODE(), or between DEBUG_CODE_BEGIN() and DEBUG_CODE_END(). Uncrustify treats these as macros and is not aware that the parameter passed into the macro call is a block of C code that needs to be formatted. Complex blocks with if/while/for/case statements are impacted the most. RECOMMENDATION: Update the uncrustify with an edk2 specific extension to treat these macros as a block of code as if they were surrounded by an extra set of braces {}. I have posted a branch for testing purposes that implements (2) and (3). Branch: https://github.com/mdkinney/edk2/tree/TestOnly_Bug_3737_3739_ApplyUncrustifyChanges_V6_OPTIONAL_Keyword_Fix PR: https://github.com/tianocore/edk2/pull/2233 Status: PASS CompareBuild: https://github.com/mdkinney/edk2/actions/runs/1532855914 Status: PASS You can see what changed by fetching and comparing the following 2 branches: https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5 https://github.com/mdkinney/edk2/tree/TestOnly_Bug_3737_3739_ApplyUncrustifyChanges_V6_OPTIONAL_Keyword_Fix Please provide feedback on the RECOMMENDATIONS above. I will go ahead and prepare of V6 version of the patch series now that that test results are all PASS. Best regards, Mike-----Original Message----- From: Michael Kubacki <mikub...@linux.microsoft.com> Sent: Thursday, December 2, 2021 1:57 PM To: devel@edk2.groups.io; Kinney, Michael D <michael.d.kin...@intel.com>; maciej.rab...@linux.intel.com; MichaelKubacki<michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm<l...@nuviainc.com>Subject: Re: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 My reply is inline. Regards, Michael On 12/2/2021 2:45 PM, Michael D Kinney wrote:Hi Maciej, Thanks for the feedback. * Example #1.This appears to be caused by the following uncrustify settings: # The threshold for aligning on '=' in assignments. # Use a negative number for absolute thresholds. # # 0: No limit (default). align_assign_thresh= 0# number The edk2 setting for this is: align_assign_thresh= 4 This means blocks of assignments that are different than more than 4 spaces will be considered a new block. ‘HwAddreLen’ and ‘Xid’ are more than 4 characters in length different.Same for ‘Xid’ and ‘Reserved’.So uncrustify treats these as 3 different assignment blocks. If we change to the default value of 0: No limit, this example is treated as a single block and all ‘=’ are aligned. Is there a reason ‘4’ was selected?Is there is any harm in using the default of 0?We can certainly change the threshold. '4' was derived by experimentation. This is an area that is somewhat subjective and every case is difficult to cover well. Please feel free to suggest any changes.I recommend we use the default value of 0. That way, any code that choose to align assignments will still be aligned. If a developer does not like short assignments and long assignments in the same code block to use the long assignment column, they can always break the block up into multiple blocks by adding a carriage return.For the benefit of others, this file can be a useful refernce understanding spans, gaps, and thresholds. https://github.com/uncrustify/uncrustify/blob/master/documentation/htdocs/configuration.txt* Example #2: Uncruistfy is confused by the DEBUG_CODE() macro.This is not a traditional macro because the contents of the macro is a block of C code.I do not know how to convince uncrustify that the contents of a macro like function call to be treated as a code block from an indent perspective.I believe this would have to be overridden in the Uncrustify fork since DEBUG_CODE() is being treated as a macro function call and code blocks are not expected there. In addition, some special treatment might be needed for alignment in between DEBUG_CODE_BEGIN()/DEBUG_CODE_END(). I'm happy to look at this. However, regression validation can take a while so I'd like to make sure we need this now. In cases that do not have additional code blocks, it seems to format fairly well. Is this prevalent and impactful enough it must be fixed now? Or, could we revisit it with a follow up patch?How long does regression testing take? I see about 20 instances of DEBUG_CODE() that are producing bad formatting. If the content inside DEBUG_CODE() is a single line of a single block of statements without any if/while/for/case statements that require further indent, then the format looks ok. DEBUG_CODE_BEGIN() and DEBUG_CODE_END() would look better if the contents between are also indented one level.* Example #3/#4: Uncrustify is confused by the OPTIONAL keyword.The edk2 config declares it as a QUALIFIER like IN and OUT.However, OPTIONAL only appears at the end of the line that declares a parameter to a function.There are 3 forms. One with comma after OPTIONAL.One with comma before OPTIONAL, and one with no comma if the parameter is the last parameter in the function declaration. TYPE ParamName OPTIONAL, TYPEParamName, OPTIONAL TYPEParamName OPTIONAL OPTIONAL is defined to nothing in edk2 builds.From a uncrustify perspective, we really want is to be ignored or more correctly treated it as a token that is attached to ParamName and the combination of ParamName and OPTIONAL treated as one unit to determine indents."TYPE ParamName, OPTIONAL" seems especially problematic and presents an inconsistency with the other formats. Mike, can you please let me know if you have the same observation and your thoughts on a consistent pattern?Thanks, Mike *From:*devel@edk2.groups.io <devel@edk2.groups.io> *On Behalf Of *Maciej Rabeda *Sent:* Thursday, December 2, 2021 10:27 AM *To:* devel@edk2.groups.io; Kinney, Michael D <michael.d.kin...@intel.com>; Michael Kubacki <michael.kuba...@microsoft.com>; Andrew Fish (af...@apple.com) <af...@apple.com>; Leif Lindholm <l...@nuviainc.com> *Subject:* Re: [edk2-devel] Uncrustify Conversion Detailed Plan and Extended Hard Freeze Update #4 Hey Mike, While most of the changes related to fixing coding style violations, there are a couple of changes to NetworkPkg in that PR that make the code less readable. Examples below. Example 1: Example 2: Example 3: Example 4: On 30-Nov-21 23:34, Michael D Kinney wrote: Hello, Thank you for your patience during this extended hard freeze. Just one more step to go.There has been a delay in the review of the patch series with the uncrustify source changes.PR(6).This patch series was not sent out as patch review email because of its very large size.It only contains source style changes and the CompareBuild tool and GitHub action has shown there are no binary differences introduced with these source style changes. If you are a package maintainer, then please review the following branch/PR for your package contents and review the EDK II CI results and BuildCompare results.I do not expect a line by line review because we already had time to provide feedback on the source style performed by uncrustify.Instead, a Reviewed-by for your package indicates that you have reviewed the EDK II CI results and CompareBuild tool functionality and results and you accept the source style changes to your package. *https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5<https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5>*https://github.com/tianocore/edk2/pull/2229 <https://github.com/tianocore/edk2/pull/2229> *https://github.com/mdkinney/edk2/actions/runs/1521618836<https://github.com/mdkinney/edk2/actions/runs/1521618836>Additional details on this update below. Thank you, Mike Changes from Update #3 ---------------------------------------------------------------------------- * Pushed PR (5) * Added link to PR(6). EDK II CI Status is PASS. Build Compare PASS. * Waiting for review of PR (6) * Review of PR (7) completed and waiting for review of PR (6) ---------------------------------------------------------------------------- Changes from Update #2 ---------------------------------------------------------------------------- * Changed order of PRs swapping (4) and (5).The PR that activates increases the max CI agent job time is independent of all the other PRs and its review is complete, so it can be committed now. * Pushed PRs (1), (2), (3), (4). * Waiting for review to complete for PRs (5) and (6) * Reviews complete for PR (7) * Identifies steps using git filter-branch to apply uncrustify changes to a code review patch series that was generated before the uncrustify changes avoiding manual merge. * Identified steps using git filter-repo to generate an alternate history of the edk2 repo with uncrustify changes applied on every commit.This may be useful when evaluating changes to files using tools like git blame without the large uncrustify patch series. --------------------------------------------------------------------------- Changes from Update #1 ---------------------------------------------------------------------------- * Changed order of PRs swapping (6) and (7).The PR that activates EDK II CI check UncrustifyCheck has to be last because it unconditionally checks all C/H files in all packages.Not just files that have been modified like some of the other checkers. * Updated link to the branch with the UncrustifyCheck plugin that has been updated with a one line change and Reviewed-by and Tested-by tags. https://github.com/mdkinney/edk2/tree/Bug_3748_add_uncrustify_ci_plugin_v6<https://github.com/mdkinney/edk2/tree/Bug_3748_add_uncrustify_ci_plugin_v6>* Reviews complete for (1), (2), (3), (5), and (7) --------------------------------------------------------------------------- Michael Kubacki and I have prepared the patches required to apply the uncrustify changes and enable EDK II CI to check all submitted patches have been run through uncrustify. We have verified through the CompareBuild GitHub Action that the format changes performed by uncrustify have no functional changes. All of the OBJ, LIB, DLL, EFI, FFS, FV, and FD files match 100% across 70 VS2019/GCC5 builds of all package/platform DSC files in the edk2 repo. The hard freeze will be extended after the edk2-stable202111 tag until all uncrustify related changes are committed.We do not expect this to take more than a few days.Do not push any PRs until the hard freeze is lifted. The changes are broken up into 7 patch series/PRs.The PRs are ordered so they can be submitted using the normal submission process and EDK II CI will pass for each one.Details are listed below. Uncrustify 73.0.3 for EDK II ============================= * Sources:https://dev.azure.com/projectmu/_git/Uncrustify <https://dev.azure.com/projectmu/_git/Uncrustify> * Documentation:https://dev.azure.com/projectmu/Uncrustify/_wiki/wikis/Uncrustify.wiki/1/Project-Mu-(EDK-II)-Fork-Readme <https://dev.azure.com/projectmu/Uncrustify/_wiki/wikis/Uncrustify.wiki/1/Project-Mu-(EDK-II)-Fork-Readme>* Download:https://dev.azure.com/projectmu/Uncrustify/_packaging?_a=package&feed=mu_uncrustify&package=mu-uncrustify-release&protocolType=NuGet&version=73.0.3 <https://dev.azure.com/projectmu/Uncrustify/_packaging?_a=package&feed=mu_uncrustify&package=mu-uncrustify- release&protocolType=NuGet&version=73.0.3>Installing Uncrustify ====================== The Uncrustify tool is installed automatically when the Pytools environment is used and the stuart* commands are run to complete the environment setup.Please see: https://github.com/tianocore/edk2/tree/master/.pytool#running-ci-locally<https://github.com/tianocore/edk2/tree/master/.pytool#running-ci-locally>Uncrustify can also be installed from the download page listed above or built from sources from the source link above. The Documentation link provides instruction on how to run uncrustify from the command line or install as a Visual Studio Code plugin.The main uncrustify documentation also describes how to integrate with a few other editors. We have also discussed a client side githook.That effort has not started. Let us know if that is a feature you would find useful. Developer impact for new code reviews ====================================== Once the uncrustify checker is active in EDK II CI, developers must make sure their patches are run through the uncrustify tool before sending the patches for review. Developers must install and run uncrustify against changes files before sending patch review emails or submitting PR for EDK II CI.If EDK II CI detects and differences in source formatting, then EDK II CI will fail and the developer must run uncrustify and resubmit the patches. Developer impact to patch series/PRs reviewed during edk2-stable201121 soft/hard freeze ======================================================================================= Developers must rebase their changes after the uncrustify source changes are committed.The branch with a preview of the uncrustify changes can be used to start this rebase work. https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5<https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5>The following steps can be used to update an existing branch with the required uncrustify format.This is the Windows version.I will add the Linux version soon. 1) Fetch and checkout and rebase to latest edk2/master git fetch origin git checkout master git rebase origin/master 2) Make a backup copy of plugin UncrustifyCheck outside WORKSPACE. (e.g. C:\Temp\UncrustifyCheck) so the uncrustify tool executable and EDK II specific uncrustify configuration file available when working with a branch that does not have those tools in its scope. xcopy .pytool\Plugin\UncrustifyCheck C:\Temp\UncrustifyCheck 3) Check out the patch series branch (e.g. MyBranch) git checkout MyBranch 4) Rebase patch series against edk2-stable202111 git rebase edk2-stable202111 5) Create new branch for the uncrustifed version (e.g. MyBranch_Uncrustified) git checkout -b MyBranch_Uncrustified 6) Use git filter-branch to uncrustify all the commits in the series between the rebase target from (2) and HEAD of the branch.A filter can be used to scope the uncrustify operations to only the C/H files in the specific package the patch series is against. (e.g. DynamicTablesPkg). BaseTools should always be excluded.If the package scoped filter is not used, it will still work, but will take longer to run because uncrustify will rescan every C/H files in the whole repo. git filter-branch --tree-filter "git ls-files DynamicTablesPkg*.c DynamicTablesPkg*.h :!BaseTools/* |c:\\Temp\\UncrustifyCheck\\mu-uncrustify-release_extdep\\Windows-x86\\uncrustify.exe -c c:\\Temp\\UncrustifyCheck\\uncrustify.cfg -F - --replace --no-backup --if-changed" edk2-stable202111..HEAD7) Now that all the individual patches in the branch are uncrustified, rebase against latest edk2/master that is already uncrustified. git rebase master 8) Verify the patches in this new branch. Impacts to tracing history across the uncrusity changes ======================================================= Tools the view file and line history do work with the large uncrustify patch series.One impact is that the operations can be very slow due to the large uncrustify patches. One option to provide a faster experience is to provide an alternate version of the edk2 repository as "documentation" that has the entire history re-written with uncrustify run on every commit. The tool called git-filter-repo can be used to perform this transformation and runs in a reasonable period of time (a few hours) https://github.com/newren/git-filter-repo <https://github.com/newren/git-filter-repo> https://github.com/newren/git-filter-repo/blob/main/contrib/filter-repo-demos/lint-history<https://github.com/newren/git-filter-repo/blob/main/contrib/filter-repo-demos/lint-history>The following steps can be used to perform this transformation. This is the Windows version. I will add the Linux version soon. ** WARNING **This operation modifies(rewrites) all the commits in the local copy of the repo.Do not perform these steps on a local repo you are using for active development. 1) Clone edk2 into a new directory (see **WARNING**) git clonehttps://github.com/tianocore/edk2.git <https://github.com/tianocore/edk2.git> edk2-uncrustified cd edk2-uncrustified 2) Setup python virtual env, install pytools, and run stuart commands to setup build environment which includes installing uncrustify tools. https://github.com/tianocore/edk2/tree/master/.pytool#running-ci-locally<https://github.com/tianocore/edk2/tree/master/.pytool#running-ci-locally>3) Make a backup copy of plugin UncrustifyCheck outside WORKSPACE. (e.g. C:\Temp\UncrustifyCheck) so the uncrustify tool executable and EDK II specific uncrustify configuration file available when working with a branch that does not have those tools in its scope. xcopy .pytool\Plugin\UncrustifyCheck C:\Temp\UncrustifyCheck 4) Use lint-history.py from git-filter-repo examples https://github.com/newren/git-filter-repo <https://github.com/newren/git-filter-repo> https://github.com/newren/git-filter-repo/blob/main/contrib/filter-repo-demos/lint-history<https://github.com/newren/git-filter-repo/blob/main/contrib/filter-repo-demos/lint-history>Line #127 - Add try except around subprocess.check_call() with except being pass.This is required because there are a few commits of C files in the edk2 repo that have incorrect C syntax and do not build with a C compiler and break the uncrustify parser.Skip reformat of C files that can not be parsed by uncrustify.These rare instances are addressed in the commit that fixes the C syntax error. Run this slightly modified version of lint-history.Include only C/H files and exclude directories that start with 'Tools' or 'BaseTools'. This step took about 2.2 hours on a laptop. lint-history.py --relevant "return (not filename.startswith(b'Tools') and not filename.startswith(b'BaseTools') and(filename.endswith(b'.c') or filename.endswith(b'.h')))"c:\\work\\GitHub\\tianocore\\foo\\UncrustifyCheck\\mu-uncrustify-release_extdep\\Windows-x86\\uncrustify.exe-cc:\\work\\GitHub\\tianocore\\foo\\UncrustifyCheck\\uncrustify.cfg --replace --no-backup --if-changedOrder of PRs to apply during extended hard freeze ================================================== 1) Update EmulatorPkg Win Host [BuildOptions] MSFT CC_FLAGS to not force debug information *https://bugzilla.tianocore.org/show_bug.cgi?id=3747 <https://bugzilla.tianocore.org/show_bug.cgi?id=3747> *https://github.com/mdkinney/edk2/tree/Bug_3747_EmulatorPkg_WinHost_ReproducibleBuild<https://github.com/mdkinney/edk2/tree/Bug_3747_EmulatorPkg_WinHost_ReproducibleBuild>*https://github.com/tianocore/edk2/pull/2215 <https://github.com/tianocore/edk2/pull/2215> * Required for EmulatorPkg to pass CompareBuild for VS2019 IA32/X64 builds. * Status: Review complete.PR pushed. 2) EccCheck should not revert staged and local changes *https://bugzilla.tianocore.org/show_bug.cgi?id=2986 <https://bugzilla.tianocore.org/show_bug.cgi?id=2986> *https://github.com/mdkinney/edk2/tree/Bug_2986_EccCheckRemoveGitRevert_V2<https://github.com/mdkinney/edk2/tree/Bug_2986_EccCheckRemoveGitRevert_V2>*https://github.com/tianocore/edk2/pull/2216 <https://github.com/tianocore/edk2/pull/2216> * Required for EDK II CI to complete in a reasonable period of time when processing the 4000+ source file style changes made by uncrustify. * Also fixes critical bugs that can potentially corrupt git state when EccCheck is run locally. * Status: Review complete.PR pushed. 3) Update pytool LicenseCheck plugin to use temp directory for diff output file *https://bugzilla.tianocore.org/show_bug.cgi?id=3746 <https://bugzilla.tianocore.org/show_bug.cgi?id=3746> *https://github.com/mdkinney/edk2/tree/Bug_3746_LicenseCheckUseDiffOutputFile_V2<https://github.com/mdkinney/edk2/tree/Bug_3746_LicenseCheckUseDiffOutputFile_V2>*https://github.com/tianocore/edk2/pull/2217 <https://github.com/tianocore/edk2/pull/2217> * Required to reduce EDK II CI build times. * Status: Review complete.PR pushed. 4) Update max job time from 60 min to 120 minutes in .azurepipelines/templates *https://bugzilla.tianocore.org/show_bug.cgi?id=3750 <https://bugzilla.tianocore.org/show_bug.cgi?id=3750> *https://github.com/mdkinney/edk2/tree/Bug_3750_IncreaseAzurePipelinesTimeout<https://github.com/mdkinney/edk2/tree/Bug_3750_IncreaseAzurePipelinesTimeout>*https://github.com/tianocore/edk2/pull/2219 <https://github.com/tianocore/edk2/pull/2219> * Required to allow EccCheck of uncrustify changes to complete on Azure Pipelines CI agents without timing out. * Status: Review complete.PR pushed. 5) Update Package YAML to ignore specific ECC files/errors *https://bugzilla.tianocore.org/show_bug.cgi?id=3749 <https://bugzilla.tianocore.org/show_bug.cgi?id=3749> *https://github.com/mdkinney/edk2/tree/Bug_3749_EccCheckIgnoreFilesErrors<https://github.com/mdkinney/edk2/tree/Bug_3749_EccCheckIgnoreFilesErrors>*https://github.com/tianocore/edk2/pull/2218 <https://github.com/tianocore/edk2/pull/2218> * Required to pass EccCheck * Status: Review complete. PR pushed 6) Uncrustify Source Changes *https://bugzilla.tianocore.org/show_bug.cgi?id=3737 <https://bugzilla.tianocore.org/show_bug.cgi?id=3737> *https://bugzilla.tianocore.org/show_bug.cgi?id=3739 <https://bugzilla.tianocore.org/show_bug.cgi?id=3739> *https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5<https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5>*https://github.com/tianocore/edk2/pull/2229 <https://github.com/tianocore/edk2/pull/2229> * Build comparison result PASS:https://github.com/mdkinney/edk2/actions/runs/1521618836<https://github.com/mdkinney/edk2/actions/runs/1521618836>* EFI_D_ -> DEBUG changes required to pass PatchCheck * Uncrustify format changes required to pass UncrustifyCheck * Status: Waiting for review 7) UncrustifyCheck EDK II CI Plugin *https://bugzilla.tianocore.org/show_bug.cgi?id=3748 <https://bugzilla.tianocore.org/show_bug.cgi?id=3748> *https://github.com/mdkinney/edk2/tree/Bug_3748_add_uncrustify_ci_plugin_v6<https://github.com/mdkinney/edk2/tree/Bug_3748_add_uncrustify_ci_plugin_v6>* Required to enforce all PRs submitted to EDK II CI match uncrustify format. * Unconditionally checks all packages.Can not be committed until all C/H source files have been updated. * Status: Review complete Combined Branch/PR for Review/Test ================================== * Build Comparison results must pass 100% across the full set of PRs before the individual PRs can be pushed in the order listed above. * Branch:https://github.com/mdkinney/edk2/tree/TestOnly_Uncrustify_PR_Series<https://github.com/mdkinney/edk2/tree/TestOnly_Uncrustify_PR_Series>* PR:https://github.com/tianocore/edk2/pull/2229 <https://github.com/tianocore/edk2/pull/2229> Status = PASS * CompareBuild: Branch:https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5<https://github.com/mdkinney/edk2/tree/Bug_3737_3739_ApplyUncrustifyChanges_V5>--ref1:ef9a059cdb15844fe52a49af2bf7d86b9dd3e9bf --ref2:Bug_3737_3739_ApplyUncrustifyChanges_V5 Extra Options: -n 4 --quiet Results:https://github.com/mdkinney/edk2/actions/runs/1521618836<https://github.com/mdkinney/edk2/actions/runs/1521618836>30 VS2019 build comparisons PASS 40 GCC5 build comparisons PASS 100% PASS The following git log shows the set of patches from --ref1 to --ref 2across which there are no differences in any of the OBJ/LIB/DLL/EFI/FFS/FV/FD files. --ref2 b7d4bf0675b7 (HEAD -> Bug_3737_3739_ApplyUncrustifyChanges_V5) UnitTestFrameworkPkg: Apply uncrusitifychanges7f03d25f60e7 UefiPayloadPkg: Apply uncrusitify changes 0bfd8d9b5ac9 UefiCpuPkg: Apply uncrusitify changes e1cd9bfb9dea StandaloneMmPkg: Apply uncrusitify changes 5da2f65be378 SourceLevelDebugPkg: Apply uncrusitify changes 95b86de07e5d SignedCapsulePkg: Apply uncrusitify changes fe71d97246c4 ShellPkg: Apply uncrusitify changes 54c21c952992 SecurityPkg: Apply uncrusitify changes 187a3785f12b RedfishPkg: Apply uncrusitify changes 810100002a46 PcAtChipsetPkg: Apply uncrusitify changes 276a695c0cf2 OvmfPkg: Apply uncrusitify changes 303c0a91ab07 NetworkPkg: Apply uncrusitify changes bc80792cd1b1 MdePkg: Apply uncrusitify changes 3ea86be17a2a MdeModulePkg: Apply uncrusitify changes c70ef11ed0cd IntelFsp2WrapperPkg: Apply uncrusitify changes c0291221f252 IntelFsp2Pkg: Apply uncrusitify changes 6a479952a690 FmpDevicePkg: Apply uncrusitify changes 3a7c05b7070d FatPkg: Apply uncrusitify changes b789f98c8959 EmulatorPkg: Apply uncrusitify changes 952d7a1c9220 EmbeddedPkg: Apply uncrusitify changes a1cc9881bab6 DynamicTablesPkg: Apply uncrusitify changes 50654dfe5785 CryptoPkg: Apply uncrusitify changes ed965a02dfa1 ArmVirtPkg: Apply uncrusitify changes 9744023fbc46 ArmPlatformPkg: Apply uncrusitify changes 7a1cde5f5bba ArmPkg: Apply uncrusitify changes 19d17e0913e8 UefiCpuPkg: Change use of EFI_D_* to DEBUG_* ffa718b4f994 SourceLevelDebugPkg: Change use of EFI_D_* to DEBUG_* b86cb3c5e5b4 ShellPkg: Change use of EFI_D_* to DEBUG_* c7c42204dc07 SecurityPkg: Change use of EFI_D_* to DEBUG_* 16b8e6f958e4 PcAtChipsetPkg: Change use of EFI_D_* to DEBUG_* 0ac3f8b2dac5 OvmfPkg: Change use of EFI_D_* to DEBUG_* bc5004b8d294 NetworkPkg: Change use of EFI_D_* to DEBUG_* 6f671a8e2377 MdePkg: Change use of EFI_D_* to DEBUG_* a10c610ff9a3 MdeModulePkg: Change use of EFI_D_* to DEBUG_* 09a3bddba390 FatPkg: Change use of EFI_D_* to DEBUG_* 59c61318246a EmulatorPkg: Change use of EFI_D_* to DEBUG_* 3a80367dda3b EmbeddedPkg: Change use of EFI_D_* to DEBUG_* 23eb1aaf80ca ArmVirtPkg: Change use of EFI_D_* to DEBUG_* 875914b45c54 ArmPlatformPkg: Change use of EFI_D_* to DEBUG_* eb2eca82b451 ArmPkg: Change use of EFI_D_* to DEBUG_* f0f3f5aae7c4 (origin/master, origin/HEAD, master) UnitTestFrameworkPkg: Update YAML to ignore specific ECCfiles/errorsc05734797790 UefiPayloadPkg: Update YAML to ignore specific ECC files/errors c30c40d6c63d StandaloneMmPkg: Update YAML to ignore specific ECC files/errors 9944508e85f1 ShellPkg: Update YAML to ignore specific ECC files/errors 60fa40be458d SecurityPkg: Update YAML to ignore specific ECC files/errors df790cd6b37e MdePkg: Update YAML to ignore specific ECC files/errors 9deb9370766e MdeModulePkg: Update YAML to ignore specific ECC files/errors d7d30e8f219f EmulatorPkg: Update YAML to ignore specific ECC files/errors d5744ecba813 CryptoPkg: Update YAML to ignore specific ECC files/errors c97fee87f0f9 ArmVirtPkg: Update YAML to ignore specific ECC files/errors 1939fc9569f2 ArmPlatformPkg: Update YAML to ignore specific ECC files/errors 365dced2c37a ArmPkg: Update YAML to ignore specific ECC files/errors 76a1ce4d5fec .azurepipelines/templates: Update max pipeline job time to 2 hours 99f84ff47390 .pytools/Plugin/LicenseCheck: Use temp directory for git diff output 3019f1bbabf1 .pytool/Plugin/EccCheck: Add performance optimizations 854462bd3479 .pytool/Plugin/EccCheck: Remove temp directory on exception 69877614fdee .pytool/Plugin/EccCheck: Remove RevertCode() --ref1 ef9a059cdb15 EmulatorPkg/Win/Host: Update CC_FLAGS bb1bba3d7767 (tag: edk2-stable202111) NetworkPkg: Fix invalid pointer for DNS response token on error Best regards, Mike
-=-=-=-=-=-=-=-=-=-=-=- Groups.io Links: You receive all messages sent to this group. View/Reply Online (#84423): https://edk2.groups.io/g/devel/message/84423 Mute This Topic: https://groups.io/mt/87414953/21656 Group Owner: devel+ow...@edk2.groups.io Unsubscribe: https://edk2.groups.io/g/devel/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
