Open
Conversation
Adds support for * Path canonicalization * Making paths absolute * Add path padding character to padding signature * Stripping quotes * Backslash escaping * Corrects regex "match" bug * Adds error handling to SFN name processing * Adds file creation to sfn processing so sfns can be generated for "non existent" files * Adds method to validate a path is the proper length * Updates path padding to leverage the above api additions * Corrects a missing return statement * Adds new error types for current and future use Signed-off-by: John Parent <john.parent@kitware.com>
Signed-off-by: John Parent <john.parent@kitware.com>
Ensures child stderr pipes are handled gracefully Corrects an issue where the error pipes were not copied on assign or copy or included in the process start info or closed. Signed-off-by: John Parent <john.parent@kitware.com>
Some packages install themselves as readonly, or with other file permissions that conflict with what we're doing with this compiler wrapper. Thus, we need the ability to override (and restore) those settings. For this compiler wrapper to perform it's duties, it needs to be able to read and write the various binaries artifacts from a build. We thus need to be able to: 1. Obtain and store current permissions 2. Grant ourselves the required permissions 3. Do our reading/writing 4. Restore previous permissions On Windows, file access is handled by the read/write bit, as on other platforms, but also by Access Control Lists (ACLs) which are populated by Discretionary Access Control Lists (DACLs) which are the atomic units responsible for granting rights. This PR adds an interface to inspect, store, update, and restore the read/write bits and DACLs of a given file. This functionality is wrapped up in a RAII based class that essentially provides scoped permissions to a file for scope of the permissions object. Note: this assumes the user creating the permissions is the user driving this wrapper, or this user has admin rights, otherwise this will fail (gracefully, but still a failure) Signed-off-by: John Parent <john.parent@kitware.com>
Previously we were searching a set of prefixes for DLLs during relocation. If we found a dll that matched the dll we were looking for (based on file name) we performed relocation based on that path. This is both dangerous and extraneous. This is dangerous as mutli config layouts may have the same binary with the same name in mutliple different paths for different configs or variations. Since we were previously only checking the filename, this could lead to a false positive detection and bad relocation not detected until runtime. This is extraneous as we should never need to search. We have the dll locations before and after relocation, whether from the stage to install prefix or from buildcache to buildcache, so rather than a filesystem search, we can have a linear time operation where we search through a list of relocation old->new prefix mappings. Spack core will set an environment variable of the structure: old_prefix|new_prefix and the compiler wrapper now composes a map out of that list and then PE files looking to relocate their internal dll references get a constant time lookup. Signed-off-by: John Parent <john.parent@kitware.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Relocation: Change strategy
Previously we were searching a set of prefixes for DLLs during relocation. If we found a dll that matched the dll we were looking for (based on file name) we performed relocation based on that path. This is both dangerous and extraneous.
This is dangerous as mutli config layouts may have the same binary with the same name in mutliple different paths for different configs or variations. Since we were previously only checking the filename, this could lead to a false positive detection and bad relocation not detected until runtime.
This is extraneous as we should never need to search. We have the dll locations before and after relocation, whether from the stage to install prefix or from buildcache to buildcache, so rather than a filesystem search, we can have a linear time operation where we search through a list of relocation old->new prefix mappings.
Spack core will set an environment variable of the structure:
old_prefix|new_prefix and the compiler wrapper now composes a map out of that list and then PE files looking to relocate their internal dll references get a constant time lookup.
Based on #26