Vulkan-Docs/checkLinks.py

316 lines
12 KiB
Python
Raw Normal View History

2016-02-16 09:53:44 +00:00
#!/usr/bin/python3
#
Change log for January 5, 2018 Vulkan 1.0.67 spec update: * Bump API patch number and header version number to 67 for this update. * Update copyright dates to 2018 Github Issues: * Fix texture lookup functions in `GL_KHR_vulkan_glsl` specification (public pull request 363). * Clarify the state waited semaphores are left in when a call to flink:vkQueuePresentKHR fails (public issue 572). * Cleanup descriptions of slink:VkObjectTablePushConstantEntryNVX and slink:VkObjectTableDescriptorSetEntryNVX (public issue 583) * Remove redundant flink:vkCmdSetDiscardRectangleEXT valid usage statements (public pull 586). * Make dynamic state array length valid usage statements implicit for flink:vkCmdSetViewportWScalingNV, flink:vkCmdSetDiscardRectangleEXT, and flink:vkCmdSetViewport (public pull 589). * Clarify meaning of window extent (0,0) in slink:VkSwapchainKHR for the Windows and X11 platforms, in their respective extensions (public issue 590). * Allow flink:vkGetPastPresentationTimingGOOGLE to return ename:VK_INCOMPLETE (public issue 604). * Add synchronization valid usage statements to flink:vkAcquireNextImage (public pull 611). * Fix some broken external links and internal xrefs (public pull 613). * Clean up slink:VkViewport valid usage statements in the presence or absence of relevant extensions (public pull 623). * Remove ename:VK_IMAGE_LAYOUT_DEPTH_READ_ONLY_STENCIL_ATTACHMENT_OPTIMAL_KHR token from VK_KHR_maintenance2 from the non-extension VU path for slink:VkGraphicsPipelineCreateInfo (public issue 628). * Miscellaneous minor markup fixes - extension name strings (public pull 631), Notes (pull 633), queue names emitted by generator scripts (pull 634), block formatting in slink:VkDescriptorUpdateTemplateEntryKHR (pull 635), ename:VK_FORMAT_FEATURE_SAMPLED_IMAGE_FILTER_CUBIC_BIT_IMG (pull 641), quotes and apostrophes (pull 643), * Miscellaneous minor grammar fixes (public pull 644). * Fix markup macros so usage like ptext:*Src* works (public pull 647). Internal Issues: * Clarify in the `VK_KHR_surface` and `VK_KHR_swapchain` extensions that parameter combinations which aren't supported for normal images are also unsupported for presentable images, even if the parameter values are individually supported as reported by the surface capability queries (internal issue 1029). * Fixed XML typo in the valid value field of the pname:sType member of slink:VkPhysicalDeviceExternalMemoryHostPropertiesEXT (internal issue 1100). Other Issues: * Add memory semantics validity rules to the <<spirvenv-module-validation, Validation Rules within a Module>> section of the SPIR-V environment appendix, and specify that sequentiality consistency is not supported. This forbids certain cases like "`Load+Release`" that we don't expect to ever be meaningful. * Document mapping of OpenGL Shading Language barriers to SPIR-V scope and semantics in the `GL_KHR_vulkan_glsl` specification. New Extensions: * `VK_EXT_conservative_rasterization`
2018-01-06 01:39:15 +00:00
# Copyright (c) 2015-2018 The Khronos Group Inc.
2016-02-16 09:53:44 +00:00
#
Change log for April 22, 2016 Vulkan 1.0.11 spec update: * Bump API patch number and header version number to 11 for this update. Github Issues: * Clarify the WSI extension language by switching from the fuzzier "ownership" language to more-consistent "acquire" language (public issue 117). * Clarify that memory barriers apply to all commands in the dependency chains in the flink:vkGetRenderAreaGranularity command and the <<synchronization-execution-and-memory-dependencies,Execution And Memory Dependencies>> section (public issue 132). * Clarify that a queue family is a set of queues in the <<fundamentals-execmodel,Execution Model>> section (public issue 166). * Removed requirement from valid usage language that VkPresentInfoKHR::waitSemaphoreCount must be greater than 0 (public issue 171). * Fix broken internal links, describe structures consistently, use consistent style for SPIR-V codewords, and tag normative terms that were missing asciidoc tags (public issue 183 and ancillary markup/normative language fixes). * Fix typos for slink:VkPhysicalDeviceLimits member names in slink:VkImageCreateInfo validity language (public issue 184). Internal Issues: * Document that the requested patch version number specified as part of slink:VkApplicationInfo::pname:apiVersion is ignored when creating an instance (internal issue 176). * Clarify handling of extension structs in the <<fundamentals-validusageValid Usage>> section (internal issue 254). * Update required slink:VkImageFormatProperties::pname:maxMipLevels to be limited to the maximum allowed mipmap pyramid size corresponding to the actual maximum supported size for the format (internal issue 256). * Modify the <<features-extentperimagetype,Allowed Extent Values Based On Image Type>> section so the allowed maximum extent is the maximum image dimension supported for each dimension of the type of texture being queried (internal issue 257). * Clarify in the <<spirvenv-module-validation,Validation Rules within a Module>> section that at least one of the code:LocalSize execution mode or code:WorkgroupSize decoration is required for each compute shader entry point in a shader module (internal issue 279). * Add validity rules for formats in flink:vkCmdClearColorImage and flink:vkCmdClearDepthStencilImage (internal issue 283). * Clarify that slink:VkImageFormatProperties::pname:maxResourceSize is an upper bound, and that it may not be possible to create an image anywhere near that size (internal issue 284). Other Commits: * Fix various minor markup errors reported by validation scripts. * Change copyright from Khronos Free Use License to Apache 2.0 license on relevant script/XML/header files. This does not affect the specification source copyright.
2016-04-21 08:08:38 +00:00
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
2016-02-16 09:53:44 +00:00
#
Change log for April 22, 2016 Vulkan 1.0.11 spec update: * Bump API patch number and header version number to 11 for this update. Github Issues: * Clarify the WSI extension language by switching from the fuzzier "ownership" language to more-consistent "acquire" language (public issue 117). * Clarify that memory barriers apply to all commands in the dependency chains in the flink:vkGetRenderAreaGranularity command and the <<synchronization-execution-and-memory-dependencies,Execution And Memory Dependencies>> section (public issue 132). * Clarify that a queue family is a set of queues in the <<fundamentals-execmodel,Execution Model>> section (public issue 166). * Removed requirement from valid usage language that VkPresentInfoKHR::waitSemaphoreCount must be greater than 0 (public issue 171). * Fix broken internal links, describe structures consistently, use consistent style for SPIR-V codewords, and tag normative terms that were missing asciidoc tags (public issue 183 and ancillary markup/normative language fixes). * Fix typos for slink:VkPhysicalDeviceLimits member names in slink:VkImageCreateInfo validity language (public issue 184). Internal Issues: * Document that the requested patch version number specified as part of slink:VkApplicationInfo::pname:apiVersion is ignored when creating an instance (internal issue 176). * Clarify handling of extension structs in the <<fundamentals-validusageValid Usage>> section (internal issue 254). * Update required slink:VkImageFormatProperties::pname:maxMipLevels to be limited to the maximum allowed mipmap pyramid size corresponding to the actual maximum supported size for the format (internal issue 256). * Modify the <<features-extentperimagetype,Allowed Extent Values Based On Image Type>> section so the allowed maximum extent is the maximum image dimension supported for each dimension of the type of texture being queried (internal issue 257). * Clarify in the <<spirvenv-module-validation,Validation Rules within a Module>> section that at least one of the code:LocalSize execution mode or code:WorkgroupSize decoration is required for each compute shader entry point in a shader module (internal issue 279). * Add validity rules for formats in flink:vkCmdClearColorImage and flink:vkCmdClearDepthStencilImage (internal issue 283). * Clarify that slink:VkImageFormatProperties::pname:maxResourceSize is an upper bound, and that it may not be possible to create an image anywhere near that size (internal issue 284). Other Commits: * Fix various minor markup errors reported by validation scripts. * Change copyright from Khronos Free Use License to Apache 2.0 license on relevant script/XML/header files. This does not affect the specification source copyright.
2016-04-21 08:08:38 +00:00
# http://www.apache.org/licenses/LICENSE-2.0
2016-02-16 09:53:44 +00:00
#
Change log for April 22, 2016 Vulkan 1.0.11 spec update: * Bump API patch number and header version number to 11 for this update. Github Issues: * Clarify the WSI extension language by switching from the fuzzier "ownership" language to more-consistent "acquire" language (public issue 117). * Clarify that memory barriers apply to all commands in the dependency chains in the flink:vkGetRenderAreaGranularity command and the <<synchronization-execution-and-memory-dependencies,Execution And Memory Dependencies>> section (public issue 132). * Clarify that a queue family is a set of queues in the <<fundamentals-execmodel,Execution Model>> section (public issue 166). * Removed requirement from valid usage language that VkPresentInfoKHR::waitSemaphoreCount must be greater than 0 (public issue 171). * Fix broken internal links, describe structures consistently, use consistent style for SPIR-V codewords, and tag normative terms that were missing asciidoc tags (public issue 183 and ancillary markup/normative language fixes). * Fix typos for slink:VkPhysicalDeviceLimits member names in slink:VkImageCreateInfo validity language (public issue 184). Internal Issues: * Document that the requested patch version number specified as part of slink:VkApplicationInfo::pname:apiVersion is ignored when creating an instance (internal issue 176). * Clarify handling of extension structs in the <<fundamentals-validusageValid Usage>> section (internal issue 254). * Update required slink:VkImageFormatProperties::pname:maxMipLevels to be limited to the maximum allowed mipmap pyramid size corresponding to the actual maximum supported size for the format (internal issue 256). * Modify the <<features-extentperimagetype,Allowed Extent Values Based On Image Type>> section so the allowed maximum extent is the maximum image dimension supported for each dimension of the type of texture being queried (internal issue 257). * Clarify in the <<spirvenv-module-validation,Validation Rules within a Module>> section that at least one of the code:LocalSize execution mode or code:WorkgroupSize decoration is required for each compute shader entry point in a shader module (internal issue 279). * Add validity rules for formats in flink:vkCmdClearColorImage and flink:vkCmdClearDepthStencilImage (internal issue 283). * Clarify that slink:VkImageFormatProperties::pname:maxResourceSize is an upper bound, and that it may not be possible to create an image anywhere near that size (internal issue 284). Other Commits: * Fix various minor markup errors reported by validation scripts. * Change copyright from Khronos Free Use License to Apache 2.0 license on relevant script/XML/header files. This does not affect the specification source copyright.
2016-04-21 08:08:38 +00:00
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
2016-02-16 09:53:44 +00:00
# checkLinks.py - validate link/reference API constructs in files
#
# Usage: checkLinks.py files > logfile
#
# Uses vkapi.py, which is a Python representation of relevant parts
# of the Vulkan API.
import copy, os, pdb, re, string, sys
from vkapi import *
global curFile, curLine, sectionDepth
global errCount, warnCount, emittedPrefix, printInfo
curFile = '???'
curLine = -1
sectionDepth = 0
errCount = 0
warnCount = 0
emittedPrefix = {}
printInfo = False
# Called before printing a warning or error. Only prints once prior
# to output for a given file.
def emitPrefix():
global curFile, curLine, emittedPrefix
if (curFile not in emittedPrefix.keys()):
emittedPrefix[curFile] = None
print('Checking file:', curFile)
print('-------------------------------')
def info(*args, **kwargs):
global curFile, curLine, printInfo
if (printInfo):
emitPrefix()
print('INFO: %s line %d:' % (curFile, curLine),
' '.join([str(arg) for arg in args]))
# Print a validation warning found in a file
def warning(*args, **kwargs):
global curFile, curLine, warnCount
warnCount = warnCount + 1
emitPrefix()
print('WARNING: %s line %d:' % (curFile, curLine),
' '.join([str(arg) for arg in args]))
# Print a validation error found in a file
def error(*args, **kwargs):
global curFile, curLine, errCount
errCount = errCount + 1
emitPrefix()
print('ERROR: %s line %d:' % (curFile, curLine),
' '.join([str(arg) for arg in args]))
# See if a tag value exists in the specified dictionary and
# suggest it as an alternative if so.
def checkTag(tag, value, dict, dictName, tagName):
if (value in dict.keys()):
warning(value, 'exists in the API but not as a',
tag + ': .', 'Try using the', tagName + ': tag.')
# Report an error due to an asciidoc tag which doesn't match
# a corresponding API entity.
def foundError(errType, tag, value):
global curFile, curLine
error('no such', errType, tag + ':' + value)
# Try some heuristics to detect likely problems such as missing vk
# prefixes or the wrong tag.
# Look in all the dictionaries in vkapi.py to see if the tag
# is just wrong but the API entity actually exists.
checkTag(tag, value, flags, 'flags', 'elink')
checkTag(tag, value, enums, 'enums', 'elink')
checkTag(tag, value, structs, 'structs', 'slink/sname')
checkTag(tag, value, handles, 'handles', 'slink/sname')
checkTag(tag, value, defines, 'defines', 'slink/sname')
2016-02-16 09:53:44 +00:00
checkTag(tag, value, consts, 'consts', 'ename')
checkTag(tag, value, protos, 'protos', 'flink/fname')
checkTag(tag, value, funcpointers, 'funcpointers', 'tlink/tname')
# Look for missing vk prefixes (quirky since it's case-dependent)
# NOT DONE YET
# Look for param in the list of all parameters of the specified functions
# Returns True if found, False otherwise
def findParam(param, funclist):
for f in funclist:
if (param in protos[f]):
info('parameter:', param, 'found in function:', f)
return True
return False
# Initialize tracking state for checking links/includes
def initChecks():
global curFile, curLine, curFuncs, curStruct, accumFunc, sectionDepth
global errCount, warnCount
global incPat, linkPat, pathPat, sectionPat
# Matches asciidoc single-line section tags
sectionPat = re.compile('^(=+) ')
# Matches any asciidoc include:: directive
pathPat = re.compile('^include::([\w./_]+)\[\]')
# Matches asciidoc include:: directives used in spec/ref pages (and also
Change log for May 20, 2017 Vulkan 1.0.50 spec update: * Bump API patch number and header version number to 50 for this update. Github Issues: * Fix numerous minor issues with the VK_EXT_debug_report extension (public issues 478, 483, 486, 489, 490). Internal Issues: * Update flink:vkAllocateDescriptorSets to specify conditions under which to return ename:VK_ERROR_FRAGMENTED_POOL or ename:VK_ERROR_OUT_OF_POOL_MEMORY instead of out-of-host/out-of-device-memory, and improve the <<fundamentals-errorcodes, description of those errors (internal issue 654). * Add a NOTE documenting that flink:vkAcquireNextImageKHR can only signal a single semaphore, and how to deal with that when multiple physical devices in a logical device need to wait on it (internal issue 730). * Improve description of pname:pNext chains of slink:VkPhysicalDeviceImageFormatInfo2KHR and slink:VkImageFormatProperties2KHR (internal issue 814). * Clean up math markup issues in the <<textures, Image Operations>> chapter (internal issue 818). * Update validusage target to use more robust code for preprocessing, by making direct use of Asciidoctor's preprocessor. Added uniqueItems check to JSON vu schema and add clean_validusage target (internal issue 826). * Update style guide to prohibit writing non-self-contained (on a single bullet point) Valid Usage statements, and modify offending Valid Usage statements in the Specification to match, to assist with automatic extraction for the validation layers (internal issue 828). * Add ename:VK_VALIDATION_CHECK_SHADERS_EXT to elink:VkValidationCheckEXT of the `VK_EXT_validation_flags` extension, to selectively disable shader validation. * Remove duplicate valid usage statement for slink:VkImageMemoryBarrier. * Modify reflow.py script to place VUID tag anchors standalone on a line following their corresponding bullet point, and reflow the spec text accordingly (this had been pending since the initial tag deployment). New Extensions: * `VK_AMD_texture_gather_bias_lod`
2017-05-21 00:00:50 +00:00
# others such as validity). This is specific to the layout of the api/
# includes and allows any path precding 'api/' followed by the category
# (protos, structs, enums, etc.) followed by the name of the proto,
# struct, etc. file.
incPat = re.compile('^.*api/(\w+)/(\w+)\.txt')
2016-02-16 09:53:44 +00:00
# Lists of current /protos/ (functions) and /structs/ includes. There
# can be several protos contiguously for different forms of a command
curFuncs = []
curStruct = None
# Tag if we should accumulate funcs or start a new list. Any intervening
# pname: tags or struct includes will restart the list.
accumFunc = False
# Matches all link names in the current spec/man pages. Assumes these
# macro names are not trailing subsets of other macros. Used to
# precede the regexp with [^A-Za-z], but this didn't catch macros
# at start of line.
linkPat = re.compile('([efpst](name|link)):(\w*)')
# Total error/warning counters
errCount = 0
warnCount = 0
# Validate asciidoc internal links in specified file.
# infile - filename to validate
# follow - if True, recursively follow include:: directives
# included - if True, function was called recursively
# Links checked are:
# fname:vkBlah - Vulkan command name (generates internal link)
# flink:vkBlah - Vulkan command name
# sname:VkBlah - Vulkan struct name (generates internal link)
# slink:VkBlah - Vulkan struct name
# elink:VkEnumName - Vulkan enumeration ('enum') type name
# ename:VK_BLAH - Vulkan enumerant token name
# pname:name - parameter name to a command or a struct member
# tlink:name - Other Vulkan type name (generates internal link)
# tname:name - Other Vulkan type name
def checkLinks(infile, follow = False, included = False):
global curFile, curLine, curFuncs, curStruct, accumFunc, sectionDepth
global errCount, warnCount
global incPat, linkPat, pathPat, sectionPat
# Global state which gets saved and restored by this function
oldCurFile = curFile
oldCurLine = curLine
curFile = infile
curLine = 0
# N.b. dirname() returns an empty string for a path with no directories,
# unlike the shell dirname(1).
if (not os.path.exists(curFile)):
error('No such file', curFile, '- skipping check')
# Restore global state before exiting the function
curFile = oldCurFile
curLine = oldCurLine
return
inPath = os.path.dirname(curFile)
Change log for November 25, 2016 Vulkan 1.0.35 spec update: * Bump API patch number and header version number to 35 for this update. Github Issues: * Document in the <<memory-device-hostaccess,Host Access>> section that mapping and unmapping does not invalidate or flush the mapped memory (public issues 27, 126). * Redefine the entire <<synchronization>> chapter in terms of consistent and well defined terminology, that's called out at the start of the chapter. This terminology is applied equally to all synchronization types, including subpass dependencies, submissions, and much of the implicit ordering stuff dotted around the spec. Key terms are laid out in the <<synchronization-dependencies,Execution and Memory Dependencies>> section at the top of the rewritten chapter (public issues 128, 131, 132, 217, 299, 300, 302, 306, 322, 346, 347, 371, 407). * Specify order of submission for batches in the <<vkQueueSubmit,vkQueueSubmit>> and <<vkQueueBindSparse,vkQueueBindSparse>> commands (public issue 371). * Add valid usage statements to each of the WSI extension sections indicating that the WSI-specific structure parameters must be valid, and remove automatically generated valid usage statements now covered by the manual sections (public issue 383). * Clarify render pass compatibility for flink:vkCmdExecuteCommands (public issue 390). Internal Issues: * Update +vk.xml+ to make previously explicit valid usage statements for <<vkDebugReportMessageEXT,vkDebugReportMessageEXT>> implicit instead (internal issue 553). * Add valid usage statement for slink:VkCreateImageInfo preventing creation of 1D sparse images (internal issue 573). * Fix Python scripts to always read/write files in utf-8 encoding, and a logic error in reflib.py which could cause a fatal error for malstructured asciidoc (internal issues 578, 586).
2016-11-26 10:33:44 +00:00
fp = open(curFile, 'r', encoding='utf-8')
2016-02-16 09:53:44 +00:00
for line in fp:
curLine = curLine + 1
# Track changes up and down section headers, and forget
# the current functions/structure when popping up a level
match = sectionPat.search(line)
if (match):
Change log for May 20, 2017 Vulkan 1.0.50 spec update: * Bump API patch number and header version number to 50 for this update. Github Issues: * Fix numerous minor issues with the VK_EXT_debug_report extension (public issues 478, 483, 486, 489, 490). Internal Issues: * Update flink:vkAllocateDescriptorSets to specify conditions under which to return ename:VK_ERROR_FRAGMENTED_POOL or ename:VK_ERROR_OUT_OF_POOL_MEMORY instead of out-of-host/out-of-device-memory, and improve the <<fundamentals-errorcodes, description of those errors (internal issue 654). * Add a NOTE documenting that flink:vkAcquireNextImageKHR can only signal a single semaphore, and how to deal with that when multiple physical devices in a logical device need to wait on it (internal issue 730). * Improve description of pname:pNext chains of slink:VkPhysicalDeviceImageFormatInfo2KHR and slink:VkImageFormatProperties2KHR (internal issue 814). * Clean up math markup issues in the <<textures, Image Operations>> chapter (internal issue 818). * Update validusage target to use more robust code for preprocessing, by making direct use of Asciidoctor's preprocessor. Added uniqueItems check to JSON vu schema and add clean_validusage target (internal issue 826). * Update style guide to prohibit writing non-self-contained (on a single bullet point) Valid Usage statements, and modify offending Valid Usage statements in the Specification to match, to assist with automatic extraction for the validation layers (internal issue 828). * Add ename:VK_VALIDATION_CHECK_SHADERS_EXT to elink:VkValidationCheckEXT of the `VK_EXT_validation_flags` extension, to selectively disable shader validation. * Remove duplicate valid usage statement for slink:VkImageMemoryBarrier. * Modify reflow.py script to place VUID tag anchors standalone on a line following their corresponding bullet point, and reflow the spec text accordingly (this had been pending since the initial tag deployment). New Extensions: * `VK_AMD_texture_gather_bias_lod`
2017-05-21 00:00:50 +00:00
info('Match sectionPat for line:', line)
2016-02-16 09:53:44 +00:00
depth = len(match.group(1))
if (depth < sectionDepth):
info('Resetting current function/structure for section:', line)
curFuncs = []
curStruct = None
sectionDepth = depth
match = pathPat.search(line)
if (match):
incpath = match.group(1)
Change log for May 20, 2017 Vulkan 1.0.50 spec update: * Bump API patch number and header version number to 50 for this update. Github Issues: * Fix numerous minor issues with the VK_EXT_debug_report extension (public issues 478, 483, 486, 489, 490). Internal Issues: * Update flink:vkAllocateDescriptorSets to specify conditions under which to return ename:VK_ERROR_FRAGMENTED_POOL or ename:VK_ERROR_OUT_OF_POOL_MEMORY instead of out-of-host/out-of-device-memory, and improve the <<fundamentals-errorcodes, description of those errors (internal issue 654). * Add a NOTE documenting that flink:vkAcquireNextImageKHR can only signal a single semaphore, and how to deal with that when multiple physical devices in a logical device need to wait on it (internal issue 730). * Improve description of pname:pNext chains of slink:VkPhysicalDeviceImageFormatInfo2KHR and slink:VkImageFormatProperties2KHR (internal issue 814). * Clean up math markup issues in the <<textures, Image Operations>> chapter (internal issue 818). * Update validusage target to use more robust code for preprocessing, by making direct use of Asciidoctor's preprocessor. Added uniqueItems check to JSON vu schema and add clean_validusage target (internal issue 826). * Update style guide to prohibit writing non-self-contained (on a single bullet point) Valid Usage statements, and modify offending Valid Usage statements in the Specification to match, to assist with automatic extraction for the validation layers (internal issue 828). * Add ename:VK_VALIDATION_CHECK_SHADERS_EXT to elink:VkValidationCheckEXT of the `VK_EXT_validation_flags` extension, to selectively disable shader validation. * Remove duplicate valid usage statement for slink:VkImageMemoryBarrier. * Modify reflow.py script to place VUID tag anchors standalone on a line following their corresponding bullet point, and reflow the spec text accordingly (this had been pending since the initial tag deployment). New Extensions: * `VK_AMD_texture_gather_bias_lod`
2017-05-21 00:00:50 +00:00
info('Match pathPat for line:', line)
info(' incpath =', incpath)
2016-02-16 09:53:44 +00:00
# An include:: directive. First check if it looks like a
# function or struct include file, and modify the corresponding
# current function or struct state accordingly.
match = incPat.search(incpath)
if (match):
Change log for May 20, 2017 Vulkan 1.0.50 spec update: * Bump API patch number and header version number to 50 for this update. Github Issues: * Fix numerous minor issues with the VK_EXT_debug_report extension (public issues 478, 483, 486, 489, 490). Internal Issues: * Update flink:vkAllocateDescriptorSets to specify conditions under which to return ename:VK_ERROR_FRAGMENTED_POOL or ename:VK_ERROR_OUT_OF_POOL_MEMORY instead of out-of-host/out-of-device-memory, and improve the <<fundamentals-errorcodes, description of those errors (internal issue 654). * Add a NOTE documenting that flink:vkAcquireNextImageKHR can only signal a single semaphore, and how to deal with that when multiple physical devices in a logical device need to wait on it (internal issue 730). * Improve description of pname:pNext chains of slink:VkPhysicalDeviceImageFormatInfo2KHR and slink:VkImageFormatProperties2KHR (internal issue 814). * Clean up math markup issues in the <<textures, Image Operations>> chapter (internal issue 818). * Update validusage target to use more robust code for preprocessing, by making direct use of Asciidoctor's preprocessor. Added uniqueItems check to JSON vu schema and add clean_validusage target (internal issue 826). * Update style guide to prohibit writing non-self-contained (on a single bullet point) Valid Usage statements, and modify offending Valid Usage statements in the Specification to match, to assist with automatic extraction for the validation layers (internal issue 828). * Add ename:VK_VALIDATION_CHECK_SHADERS_EXT to elink:VkValidationCheckEXT of the `VK_EXT_validation_flags` extension, to selectively disable shader validation. * Remove duplicate valid usage statement for slink:VkImageMemoryBarrier. * Modify reflow.py script to place VUID tag anchors standalone on a line following their corresponding bullet point, and reflow the spec text accordingly (this had been pending since the initial tag deployment). New Extensions: * `VK_AMD_texture_gather_bias_lod`
2017-05-21 00:00:50 +00:00
info('Match incPat for line:', line)
2016-02-16 09:53:44 +00:00
# For prototypes, if it is preceded by
# another include:: directive with no intervening link: tags,
# add to the current function list. Otherwise start a new list.
# There is only one current structure.
category = match.group(1)
tag = match.group(2)
# @ Validate tag!
# @ Arguably, any intervening text should shift to accumFuncs = False,
# e.g. only back-to-back includes separated by blank lines would be
# accumulated.
if (category == 'protos'):
if (tag in protos.keys()):
if (accumFunc):
curFuncs.append(tag)
else:
curFuncs = [ tag ]
# Restart accumulating functions
accumFunc = True
info('curFuncs =', curFuncs, 'accumFunc =', accumFunc)
else:
error('include of nonexistent function', tag)
elif (category == 'structs'):
if (tag in structs.keys()):
curStruct = tag
# Any /structs/ include means to stop accumulating /protos/
accumFunc = False
info('curStruct =', curStruct)
else:
error('include of nonexistent struct', tag)
if (follow):
# Actually process the included file now, recursively
newpath = os.path.normpath(os.path.join(inPath, incpath))
info(curFile, ': including file:', newpath)
checkLinks(newpath, follow, included=True)
matches = linkPat.findall(line)
for match in matches:
# Start actual validation work. Depending on what the
# asciidoc tag name is, look up the value in the corresponding
# dictionary.
tag = match[0]
value = match[2]
if (tag == 'fname' or tag == 'flink'):
if (value not in protos.keys()):
foundError('function', tag, value)
elif (tag == 'sname' or tag == 'slink'):
if (value not in structs.keys() and
value not in handles.keys()):
foundError('aggregate/scalar/handle/define type', tag, value)
2016-02-16 09:53:44 +00:00
elif (tag == 'ename'):
if (value not in consts.keys() and value not in defines.keys()):
2016-02-16 09:53:44 +00:00
foundError('enumerant/constant', tag, value)
elif (tag == 'elink'):
if (value not in enums.keys() and value not in flags.keys()):
foundError('enum/bitflag type', tag, value)
elif (tag == 'tlink' or tag == 'tname'):
if (value not in funcpointers.keys()):
foundError('function pointer/other type', tag, value)
elif (tag == 'pname'):
# Any pname: tag means to stop accumulating /protos/
accumFunc = False
# See if this parameter is in the current proto(s) and struct
foundParam = False
if (curStruct and value in structs[curStruct]):
info('parameter', value, 'found in struct', curStruct)
elif (curFuncs and findParam(value, curFuncs)):
True
else:
warning('parameter', value, 'not found. curStruct =',
curStruct, 'curFuncs =', curFuncs)
else:
# This is a logic error
error('unknown tag', tag + ':' + value)
fp.close()
if (errCount > 0 or warnCount > 0):
if (not included):
print('Errors found:', errCount, 'Warnings found:', warnCount)
print('')
if (included):
info('----- returning from:', infile, 'to parent file', '-----')
# Don't generate any output for files without errors
# else:
# print(curFile + ': No errors found')
# Restore global state before exiting the function
curFile = oldCurFile
curLine = oldCurLine
if __name__ == '__main__':
follow = False
if (len(sys.argv) > 1):
for file in sys.argv[1:]:
if (file == '-follow'):
follow = True
elif (file == '-info'):
printInfo = True
else:
initChecks()
checkLinks(file, follow)
else:
print('Need arguments: [-follow] [-info] infile [infile...]', file=sys.stderr)