Be the first to post a text review of PyMetrics. Rate and review a project by clicking thumbs up or thumbs down in the right column.

View all reviews

• Tracker artifact added

  Anonymous created the update keywords for McCabe cyclomatic complexity artifact

  posted by nobody 337 days ago

• Tracker artifact added

  Anonymous created the unable to unzip on windows artifact

  posted by nobody 374 days ago

• Tracker artifact added

  Anonymous created the unable to install artifact

  posted by nobody 410 days ago

• Tracker artifact added

  Anonymous created the Windows XP has problems with pymetrics folder name artifact

  posted by nobody 444 days ago

• File released: /pymetrics/pymetrics-0.8.1/pymetrics-0.8.1.tar.gz

  posted 472 days ago

• File released: /pymetrics/pymetrics-0.8.0/pymetrics-0.8.0.tar.gz

  posted 472 days ago

• pymetrics pymetrics-0.8.1 file released: pymetrics-0.8.1.tar.gz

  Change Log from 0.8.0 to 0.8.1 ============================== 1) Bug fixes: a) Version 0.8.0 was Python version sensitive. It behaved differently when run under versions before 2.5 and those at or after 2.5 due to changes in the way __import__ worked. b) BUGLIST was not being update which each new release. Change Log from 0.7.9 to 0.8.0 ============================== Converted distribution to use the standard Python setup.py method. Thanks to Carlos Garcia Campos (carlosgc-at-gnome.org). Change Log from 0.7.6 to 0.7.9 ============================== 1) Bug fixes: a) ProcessRun was not executed under all circumstances. b) Code blocks in __main__ were ignored. c) Missing input files raised an exception and terminate after printing help. d) Class doc strings were counted as function doc strings. e) Metric module with missing metric class raised a 'key' exception. f) File with trailing whitespace on last indented line caused exception. 2) Output sent to stderr also sent to stdout. This allows user to see error in content since the output from stderr and stdout are not in sync. 3) Changed default metric class name so that for metric module wxYz, the default metric class name is WxYzMetric (note capitalized name) when the metric class name is omitted in the --includeMetrics parameter. 4) Added Decorators to newFeatures24 metric. Change Log from 0.7.1 to 0.7.6 ============================== 1) Added Features a) New option -B to suppress Basic Metrics output b) Added the new -e option. When the option is used, the SQL command file does not try to recreate the SQL tables. In other words, only INSERT commands are generated. Use this option if you are sure the SQL data and token tables already exist. This option implies the -N option (it deletes the old SQL command file, if it exists). c) Added computation of COCOMO 2's Source Lines Of Code (SLOC) metric. In COCOMO 2, SLOC is defined as: * Only Source lines that are delivered as part of the product are included. * Test drivers and other support software is excluded * Source lines are created by the project staff. Code created by applications generators is excluded * One SLOC is one logical line of code * Declarations are counted as SLOC * Comments are not counted as SLOC * Blank and empty lines are not counted as SLOC The COCOMO 1 model's SLOC is now called DSI (Delivered Source Instructions) in COCOMO 2. For example, in COCOMO 2, a Python "if-else" statement is counted as one SLOC, but might be counted as several DSIs. Given all that, PyMetrics will count test code if requested. I believe that this is correct to count test code because it is the basis of good Extreme Programming methodology. This SLOC metric also introduced raw text metrics. This has lead to a new member function of MetricBase: processSrcLines( self, srcLines, *a, **kw) where srcLines is a string containing the raw text of the current file. This member function is called once per file. The SLOC metric is computed by default. The following definitions are used to compute SLOC: 1) All literals are treated as occupying one logical line, regardless of the number of physical lines that are spanned. 2) All parenthesized lists, expressions, and argument lists are treated as occupying one logical line, regardless of the number of physical lines spanned. 3) All documentation strings are ignored. 2) Bug Fixes a) In Help, the default value displayed for --quiet was reversed. It now reads more logically. 3) Miscellaneous a) Now clearly identify generated code with a comment in the first line starting with "-- Automatically generated ..." 4) Added testSLOC.py to sample/ subdirectory. It tests most of weird cases that may occur in a Python source file due to continuations, literals of various types, and code spanning multiple lines. The source code shows what is counted with comments on the end of each counted source line. As of thie release, the SLOC metric for this file is 28. Change Log from 0.6.0 to 0.7.1 ============================== Quite a lot has changed from the last version. The changes fall into two categories: 1) Added Features a) More command line options to add flexibility to what is computed and displayed. PLEASE RUN PyMetrics TO DISPLAY ALL THE OPTIONS. b) Easier creation and inclusion of your own metrics c) Creation of a SQL table, called metricData, in the the metrics database that contains all the metrics that have been calculated. This table also tracks metrics over time, so you can compare metrics from one run to the next. d) Added library name field to SQL output. This is an user-defined name to identify a collection of modules. For example, all the Zope modules might have a 'zope' library name while all your modules might have your initials as a library name. In this way, you could compare the metrics in your library against the metrics in the Zope library. You could also use this feature to define a standard set of metrics with a library name of 'std' and then you could measure all other modules against the 'std' modules. 2) Refactoring and Reworking a) Due to the major enhancements introduced in this version, a number of modules have been refactored to simplify them and make them more testable. b) Unit tests have been written for a number of modules. The tests are contained in the "test" subdirectory. The "test" subdirectory has been added to this distribution. 3) Added missing LICENSE and "examples" subdirectory. 4) Fixed various bugs, none critical. Change Log from 0.5.3 to 0.6 ============================ 1) Modified PyMetrics to handle IOErrors more gracefully and continue executing, if possible. 2) Check for lines ending in '\r' and convert character to blank since Python normally converts '\r' to '\n' on Windows and Mac systems. 3) Changed metric format so each metric number appears before its name. 4) Added the -f/--files option to allow the user to specify a file containing a list of modules to analyse. 5) Fixed Halstead metric code to handle degenerate programs better. 6) Changed SQL and CSV output to have symbolic names for types of names in fully-qualified function and class names. 7) Changed semantic type of 'self' from VARNAME to KEYWORD. 8) Now count any token of semantic significance, like keywords, function names, etc. 9) Fixed issue where tarball dir was not created at a higher level directory with the name of the version in the tarball. 10) Counts of Token Types now listed in alphabetic order. 11) Added -K option to suppress display of token types. 12) Introduced MetricBase class for use in designing your own metrics. Member functions include: processToken( self, fcnName, className, tok, *args, **kwds ) processStmt( self, fcnName, className, stmt, *args, **kwds ) processBlock( self, fcnName, className, block, *args, **kwds ) processFunction( self, fcnName, className, fcn, *args, **kwds ) processClass( self, fcnName, className, cls, *args, **kwds ) processModule( self, moduleName, module, *args, **kwds ) processRun( self, run, *args, **kwds ) where: context is context in which last token occurred. metrics is metrics up to the point where function called tok is last completed token stmt is last completed statement block is last completed block fcn is last completed function cls is last completed class mod is last completed module run is last completed run (i.e., everything) Each of the stmt, block, fcn, cls, mod, and run is a list of tokens. The context is a dictionary related to the last token processed before invoking one of the processX member function. The context contains: inFile current file name fcnDepth current nested function depth classDepth current nested class depth parenDepth current parenthesis depth bracketDepth current depth of brackets ([]) braceDepth current depth of braces ({}) blockNum current unique block number blockDepth current nested block depth 13) By default, the Halstead metric is not displayed. It is still available, as explained by the Help. This makes the -H option redundant at this point. I have removed this metric because I have not found it to be generally useful. You may differ and include it in any of you runs. 14) By default, the token and keyword counts are not displayed. They are still available using the -k option. The -K option is still present, but redundant at this point. Again, I did not find this metric of particular use, so I made it optional. 15) By far, the most time spend in running PyMetrics is in writing out the SQL and CSV files. Suppressing one or both, if you do not need them, will speed up processing. However, I did speed up the processing of these files by 20-50%. Change Log from 0.5.2 to 0.5.3 ============================== 1) PyMetrics is now available under CVS. 2) Fixed a bug where the tokenize library module allowed the position of a comment on a line by itself to affect the indentation of subsequent blocks. This would cause a crash if the affected code included a function return statement. 3) Fixed bug where the McCabe metric routine was being called twice for most tokens. 4) Changed tokenize module to yield 5 parameters, as it originally did. At an early stage, I erroneously thought that I needed to pass back an extra value. 5) Sped up processing slightly by skipping redundant tests. 6) Changed subdirectory name from tests to examples. Change Log from 0.5.1 to 0.5.2 ============================== This is a bug fix release. 1) Fixed bug where program failed if first line of file started with #!. Change Log from 0.5.0 to 0.5.1 ============================== This is mainly a bug fix release. The changes made also include: 1) Fixed several divide by zero bugs that could occur for degenerate Python programs. 2) Corrected version.py to reflect the correct version and name of the program. 3) Removed %DocStrings metric since it was meaningless. 4) Fixed bug where nested classes and nested functions were not correctly reported. 5) Changed format of report on classes/functions having or missing Doc strings. 6) Report on classes/functions now uses fully qualified names. 7) Refactored some code to common up handling of certain types of tokens. 8) Added the -z/--zero command line option. Normally, zero and empty metrics are not shown. This option forces the values to be displayed. Change Log from 0.4.0 to 0.5.0 ============================== 1) Simplified processing of command line arguents and parameter handling by objectifying parameter parsing. 2) Eliminated need for global variables when handling command line parameters and options. 3) Rearranged basic metrics to appear before computed metrics. 4) Moved code out of the main module, PyMetrics.py, into separate files. 5) Clarified some options in the Help output. Also added the -S option to suppress output of the SQL command file. 6) Defined and implemented endOfToken, endOfStmt, endOfFcn, endOfClass, endOfModule, and endOfRun functions for main program. 7) Implemented a very simple subscribe feature to allow various metric classes to be registered for the events in point 6).

  posted 472 days ago

• pymetrics pymetrics-0.8.0 file released: pymetrics-0.8.0.tar.gz

  Change Log from 0.7.9 to 0.8.0 ============================== Converted distribution to use the standard Python setup.py method. Thanks to Carlos Garcia Campos (carlosgc-at-gnome.org). Change Log from 0.7.6 to 0.7.9 ============================== 1) Bug fixes: a) ProcessRun was not executed under all circumstances. b) Code blocks in __main__ were ignored. c) Missing input files raised an exception and terminate after printing help. d) Class doc strings were counted as function doc strings. e) Metric module with missing metric class raised a 'key' exception. f) File with trailing whitespace on last indented line caused exception. 2) Output sent to stderr also sent to stdout. This allows user to see error in content since the output from stderr and stdout are not in sync. 3) Changed default metric class name so that for metric module wxYz, the default metric class name is WxYzMetric (note capitalized name) when the metric class name is omitted in the --includeMetrics parameter. 4) Added Decorators to newFeatures24 metric. Change Log from 0.7.1 to 0.7.6 ============================== 1) Added Features a) New option -B to suppress Basic Metrics output b) Added the new -e option. When the option is used, the SQL command file does not try to recreate the SQL tables. In other words, only INSERT commands are generated. Use this option if you are sure the SQL data and token tables already exist. This option implies the -N option (it deletes the old SQL command file, if it exists). c) Added computation of COCOMO 2's Source Lines Of Code (SLOC) metric. In COCOMO 2, SLOC is defined as: * Only Source lines that are delivered as part of the product are included. * Test drivers and other support software is excluded * Source lines are created by the project staff. Code created by applications generators is excluded * One SLOC is one logical line of code * Declarations are counted as SLOC * Comments are not counted as SLOC * Blank and empty lines are not counted as SLOC The COCOMO 1 model's SLOC is now called DSI (Delivered Source Instructions) in COCOMO 2. For example, in COCOMO 2, a Python "if-else" statement is counted as one SLOC, but might be counted as several DSIs. Given all that, PyMetrics will count test code if requested. I believe that this is correct to count test code because it is the basis of good Extreme Programming methodology. This SLOC metric also introduced raw text metrics. This has lead to a new member function of MetricBase: processSrcLines( self, srcLines, *a, **kw) where srcLines is a string containing the raw text of the current file. This member function is called once per file. The SLOC metric is computed by default. The following definitions are used to compute SLOC: 1) All literals are treated as occupying one logical line, regardless of the number of physical lines that are spanned. 2) All parenthesized lists, expressions, and argument lists are treated as occupying one logical line, regardless of the number of physical lines spanned. 3) All documentation strings are ignored. 2) Bug Fixes a) In Help, the default value displayed for --quiet was reversed. It now reads more logically. 3) Miscellaneous a) Now clearly identify generated code with a comment in the first line starting with "-- Automatically generated ..." 4) Added testSLOC.py to sample/ subdirectory. It tests most of weird cases that may occur in a Python source file due to continuations, literals of various types, and code spanning multiple lines. The source code shows what is counted with comments on the end of each counted source line. As of thie release, the SLOC metric for this file is 28. Change Log from 0.6.0 to 0.7.1 ============================== Quite a lot has changed from the last version. The changes fall into two categories: 1) Added Features a) More command line options to add flexibility to what is computed and displayed. PLEASE RUN PyMetrics TO DISPLAY ALL THE OPTIONS. b) Easier creation and inclusion of your own metrics c) Creation of a SQL table, called metricData, in the the metrics database that contains all the metrics that have been calculated. This table also tracks metrics over time, so you can compare metrics from one run to the next. d) Added library name field to SQL output. This is an user-defined name to identify a collection of modules. For example, all the Zope modules might have a 'zope' library name while all your modules might have your initials as a library name. In this way, you could compare the metrics in your library against the metrics in the Zope library. You could also use this feature to define a standard set of metrics with a library name of 'std' and then you could measure all other modules against the 'std' modules. 2) Refactoring and Reworking a) Due to the major enhancements introduced in this version, a number of modules have been refactored to simplify them and make them more testable. b) Unit tests have been written for a number of modules. The tests are contained in the "test" subdirectory. The "test" subdirectory has been added to this distribution. 3) Added missing LICENSE and "examples" subdirectory. 4) Fixed various bugs, none critical. Change Log from 0.5.3 to 0.6 ============================ 1) Modified PyMetrics to handle IOErrors more gracefully and continue executing, if possible. 2) Check for lines ending in '\r' and convert character to blank since Python normally converts '\r' to '\n' on Windows and Mac systems. 3) Changed metric format so each metric number appears before its name. 4) Added the -f/--files option to allow the user to specify a file containing a list of modules to analyse. 5) Fixed Halstead metric code to handle degenerate programs better. 6) Changed SQL and CSV output to have symbolic names for types of names in fully-qualified function and class names. 7) Changed semantic type of 'self' from VARNAME to KEYWORD. 8) Now count any token of semantic significance, like keywords, function names, etc. 9) Fixed issue where tarball dir was not created at a higher level directory with the name of the version in the tarball. 10) Counts of Token Types now listed in alphabetic order. 11) Added -K option to suppress display of token types. 12) Introduced MetricBase class for use in designing your own metrics. Member functions include: processToken( self, fcnName, className, tok, *args, **kwds ) processStmt( self, fcnName, className, stmt, *args, **kwds ) processBlock( self, fcnName, className, block, *args, **kwds ) processFunction( self, fcnName, className, fcn, *args, **kwds ) processClass( self, fcnName, className, cls, *args, **kwds ) processModule( self, moduleName, module, *args, **kwds ) processRun( self, run, *args, **kwds ) where: context is context in which last token occurred. metrics is metrics up to the point where function called tok is last completed token stmt is last completed statement block is last completed block fcn is last completed function cls is last completed class mod is last completed module run is last completed run (i.e., everything) Each of the stmt, block, fcn, cls, mod, and run is a list of tokens. The context is a dictionary related to the last token processed before invoking one of the processX member function. The context contains: inFile current file name fcnDepth current nested function depth classDepth current nested class depth parenDepth current parenthesis depth bracketDepth current depth of brackets ([]) braceDepth current depth of braces ({}) blockNum current unique block number blockDepth current nested block depth 13) By default, the Halstead metric is not displayed. It is still available, as explained by the Help. This makes the -H option redundant at this point. I have removed this metric because I have not found it to be generally useful. You may differ and include it in any of you runs. 14) By default, the token and keyword counts are not displayed. They are still available using the -k option. The -K option is still present, but redundant at this point. Again, I did not find this metric of particular use, so I made it optional. 15) By far, the most time spend in running PyMetrics is in writing out the SQL and CSV files. Suppressing one or both, if you do not need them, will speed up processing. However, I did speed up the processing of these files by 20-50%. Change Log from 0.5.2 to 0.5.3 ============================== 1) PyMetrics is now available under CVS. 2) Fixed a bug where the tokenize library module allowed the position of a comment on a line by itself to affect the indentation of subsequent blocks. This would cause a crash if the affected code included a function return statement. 3) Fixed bug where the McCabe metric routine was being called twice for most tokens. 4) Changed tokenize module to yield 5 parameters, as it originally did. At an early stage, I erroneously thought that I needed to pass back an extra value. 5) Sped up processing slightly by skipping redundant tests. 6) Changed subdirectory name from tests to examples. Change Log from 0.5.1 to 0.5.2 ============================== This is a bug fix release. 1) Fixed bug where program failed if first line of file started with #!. Change Log from 0.5.0 to 0.5.1 ============================== This is mainly a bug fix release. The changes made also include: 1) Fixed several divide by zero bugs that could occur for degenerate Python programs. 2) Corrected version.py to reflect the correct version and name of the program. 3) Removed %DocStrings metric since it was meaningless. 4) Fixed bug where nested classes and nested functions were not correctly reported. 5) Changed format of report on classes/functions having or missing Doc strings. 6) Report on classes/functions now uses fully qualified names. 7) Refactored some code to common up handling of certain types of tokens. 8) Added the -z/--zero command line option. Normally, zero and empty metrics are not shown. This option forces the values to be displayed. Change Log from 0.4.0 to 0.5.0 ============================== 1) Simplified processing of command line arguents and parameter handling by objectifying parameter parsing. 2) Eliminated need for global variables when handling command line parameters and options. 3) Rearranged basic metrics to appear before computed metrics. 4) Moved code out of the main module, PyMetrics.py, into separate files. 5) Clarified some options in the Help output. Also added the -S option to suppress output of the SQL command file. 6) Defined and implemented endOfToken, endOfStmt, endOfFcn, endOfClass, endOfModule, and endOfRun functions for main program. 7) Implemented a very simple subscribe feature to allow various metric classes to be registered for the events in point 6).

  posted 473 days ago

• Tracker comment added

  sf-robot commented on the setup.py file to install PyMetrics artifact

  posted by sf-robot 782 days ago

• pymetrics 0.7.9 file released: pymetrics-0.7.9.tar.gz

  Change Log from 0.7.6 to 0.7.9 ============================== 1) Bug fixes: a) ProcessRun was not executed under all circumstances. b) Code blocks in __main__ were ignored. c) Missing input files raised an exception and terminate after printing help. d) Class doc strings were counted as function doc strings. e) Metric module with missing metric class raised a 'key' exception. f) File with trailing whitespace on last indented line caused exception. 2) Output sent to stderr also sent to stdout. This allows user to see error in content since the output from stderr and stdout are not in sync. 3) Changed default metric class name so that for metric module wxYz, the default metric class name is WxYzMetric (note capitalized name) when the metric class name is omitted in the --includeMetrics parameter. 4) Added Decorators to newFeatures24 metric. Change Log from 0.7.1 to 0.7.6 ============================== 1) Added Features a) New option -B to suppress Basic Metrics output b) Added the new -e option. When the option is used, the SQL command file does not try to recreate the SQL tables. In other words, only INSERT commands are generated. Use this option if you are sure the SQL data and token tables already exist. This option implies the -N option (it deletes the old SQL command file, if it exists). c) Added computation of COCOMO 2's Source Lines Of Code (SLOC) metric. In COCOMO 2, SLOC is defined as: * Only Source lines that are delivered as part of the product are included. * Test drivers and other support software is excluded * Source lines are created by the project staff. Code created by applications generators is excluded * One SLOC is one logical line of code * Declarations are counted as SLOC * Comments are not counted as SLOC * Blank and empty lines are not counted as SLOC The COCOMO 1 model's SLOC is now called DSI (Delivered Source Instructions) in COCOMO 2. For example, in COCOMO 2, a Python "if-else" statement is counted as one SLOC, but might be counted as several DSIs. Given all that, PyMetrics will count test code if requested. I believe that this is correct to count test code because it is the basis of good Extreme Programming methodology. This SLOC metric also introduced raw text metrics. This has lead to a new member function of MetricBase: processSrcLines( self, srcLines, *a, **kw) where srcLines is a string containing the raw text of the current file. This member function is called once per file. The SLOC metric is computed by default. The following definitions are used to compute SLOC: 1) All literals are treated as occupying one logical line, regardless of the number of physical lines that are spanned. 2) All parenthesized lists, expressions, and argument lists are treated as occupying one logical line, regardless of the number of physical lines spanned. 3) All documentation strings are ignored. 2) Bug Fixes a) In Help, the default value displayed for --quiet was reversed. It now reads more logically. 3) Miscellaneous a) Now clearly identify generated code with a comment in the first line starting with "-- Automatically generated ..." 4) Added testSLOC.py to sample/ subdirectory. It tests most of weird cases that may occur in a Python source file due to continuations, literals of various types, and code spanning multiple lines. The source code shows what is counted with comments on the end of each counted source line. As of thie release, the SLOC metric for this file is 28. Change Log from 0.6.0 to 0.7.1 ============================== Quite a lot has changed from the last version. The changes fall into two categories: 1) Added Features a) More command line options to add flexibility to what is computed and displayed. PLEASE RUN PyMetrics TO DISPLAY ALL THE OPTIONS. b) Easier creation and inclusion of your own metrics c) Creation of a SQL table, called metricData, in the the metrics database that contains all the metrics that have been calculated. This table also tracks metrics over time, so you can compare metrics from one run to the next. d) Added library name field to SQL output. This is an user-defined name to identify a collection of modules. For example, all the Zope modules might have a 'zope' library name while all your modules might have your initials as a library name. In this way, you could compare the metrics in your library against the metrics in the Zope library. You could also use this feature to define a standard set of metrics with a library name of 'std' and then you could measure all other modules against the 'std' modules. 2) Refactoring and Reworking a) Due to the major enhancements introduced in this version, a number of modules have been refactored to simplify them and make them more testable. b) Unit tests have been written for a number of modules. The tests are contained in the "test" subdirectory. The "test" subdirectory has been added to this distribution. 3) Added missing LICENSE and "examples" subdirectory. 4) Fixed various bugs, none critical. Change Log from 0.5.3 to 0.6 ============================ 1) Modified PyMetrics to handle IOErrors more gracefully and continue executing, if possible. 2) Check for lines ending in '\r' and convert character to blank since Python normally converts '\r' to '\n' on Windows and Mac systems. 3) Changed metric format so each metric number appears before its name. 4) Added the -f/--files option to allow the user to specify a file containing a list of modules to analyse. 5) Fixed Halstead metric code to handle degenerate programs better. 6) Changed SQL and CSV output to have symbolic names for types of names in fully-qualified function and class names. 7) Changed semantic type of 'self' from VARNAME to KEYWORD. 8) Now count any token of semantic significance, like keywords, function names, etc. 9) Fixed issue where tarball dir was not created at a higher level directory with the name of the version in the tarball. 10) Counts of Token Types now listed in alphabetic order. 11) Added -K option to suppress display of token types. 12) Introduced MetricBase class for use in designing your own metrics. Member functions include: processToken( self, fcnName, className, tok, *args, **kwds ) processStmt( self, fcnName, className, stmt, *args, **kwds ) processBlock( self, fcnName, className, block, *args, **kwds ) processFunction( self, fcnName, className, fcn, *args, **kwds ) processClass( self, fcnName, className, cls, *args, **kwds ) processModule( self, moduleName, module, *args, **kwds ) processRun( self, run, *args, **kwds ) where: context is context in which last token occurred. metrics is metrics up to the point where function called tok is last completed token stmt is last completed statement block is last completed block fcn is last completed function cls is last completed class mod is last completed module run is last completed run (i.e., everything) Each of the stmt, block, fcn, cls, mod, and run is a list of tokens. The context is a dictionary related to the last token processed before invoking one of the processX member function. The context contains: inFile current file name fcnDepth current nested function depth classDepth current nested class depth parenDepth current parenthesis depth bracketDepth current depth of brackets ([]) braceDepth current depth of braces ({}) blockNum current unique block number blockDepth current nested block depth 13) By default, the Halstead metric is not displayed. It is still available, as explained by the Help. This makes the -H option redundant at this point. I have removed this metric because I have not found it to be generally useful. You may differ and include it in any of you runs. 14) By default, the token and keyword counts are not displayed. They are still available using the -k option. The -K option is still present, but redundant at this point. Again, I did not find this metric of particular use, so I made it optional. 15) By far, the most time spend in running PyMetrics is in writing out the SQL and CSV files. Suppressing one or both, if you do not need them, will speed up processing. However, I did speed up the processing of these files by 20-50%. Change Log from 0.5.2 to 0.5.3 ============================== 1) PyMetrics is now available under CVS. 2) Fixed a bug where the tokenize library module allowed the position of a comment on a line by itself to affect the indentation of subsequent blocks. This would cause a crash if the affected code included a function return statement. 3) Fixed bug where the McCabe metric routine was being called twice for most tokens. 4) Changed tokenize module to yield 5 parameters, as it originally did. At an early stage, I erroneously thought that I needed to pass back an extra value. 5) Sped up processing slightly by skipping redundant tests. 6) Changed subdirectory name from tests to examples. Change Log from 0.5.1 to 0.5.2 ============================== This is a bug fix release. 1) Fixed bug where program failed if first line of file started with #!. Change Log from 0.5.0 to 0.5.1 ============================== This is mainly a bug fix release. The changes made also include: 1) Fixed several divide by zero bugs that could occur for degenerate Python programs. 2) Corrected version.py to reflect the correct version and name of the program. 3) Removed %DocStrings metric since it was meaningless. 4) Fixed bug where nested classes and nested functions were not correctly reported. 5) Changed format of report on classes/functions having or missing Doc strings. 6) Report on classes/functions now uses fully qualified names. 7) Refactored some code to common up handling of certain types of tokens. 8) Added the -z/--zero command line option. Normally, zero and empty metrics are not shown. This option forces the values to be displayed. Change Log from 0.4.0 to 0.5.0 ============================== 1) Simplified processing of command line arguents and parameter handling by objectifying parameter parsing. 2) Eliminated need for global variables when handling command line parameters and options. 3) Rearranged basic metrics to appear before computed metrics. 4) Moved code out of the main module, PyMetrics.py, into separate files. 5) Clarified some options in the Help output. Also added the -S option to suppress output of the SQL command file. 6) Defined and implemented endOfToken, endOfStmt, endOfFcn, endOfClass, endOfModule, and endOfRun functions for main program. 7) Implemented a very simple subscribe feature to allow various metric classes to be registered for the events in point 6).

  posted 801 days ago